PROMISe Intervention Study

PROMISe Intervention Study: Final Report

PROMISe

Intervention Study

Final Report to the Pharmacy

Guild of Australia

(RFT 2003-2, Evaluation of Clinical

Interventions in Community Pharmacies)

This research was funded by the Australian Government Department of Health and Ageing through the Third Community Pharmacy

Agreement Research and Development Program


Page 2 of 361

Table of Contents

1 Executive Summary..................................................................................................... 9

1.1 Introduction and Aims: ........................................................................................ 9

1.2 Methods: ............................................................................................................. 10

1.2.1 Data Collection ............................................................................................. 10

1.2.2 Determination of Value of Interventions ........................................................ 11

1.3 Results: ............................................................................................................... 11

1.4 Conclusions:....................................................................................................... 13

2 Authors and Acknowledgements.............................................................................. 14

2.1 Lead Investigator................................................................................................ 14

2.2 Principal Investigators ....................................................................................... 14

2.3 Acknowledgements............................................................................................ 15

3 Introduction and Objectives...................................................................................... 16

4 Methods...................................................................................................................... 17

4.1 Modifications as a Result of PROMISe Pilot Study .......................................... 17

4.1.1 Development and Modifications to the DOCUMENT categorisation system.. 17

4.1.1.1 Type of Drug Related Problem (Category and Subcategory)................. 18

4.1.1.2 Actions.................................................................................................. 22

4.1.1.3 Recommendations ................................................................................ 23

4.1.1.4 Outcome ............................................................................................... 24

4.1.1.5 Clinical Significance .............................................................................. 25

4.1.1.6 Identification of the Problem (Proactiveness) ........................................ 26

4.1.2 Modifications to Intervention Recording Software ......................................... 26

4.1.2.1 Data Management (Local Storage, Communication and Repository Storage) ............................................................................................................. 27

4.1.2.2 Communication: External Data Transfer................................................ 28

4.1.3 Documenting Interventions in WiniFRED Dispense ...................................... 30

4.1.3.1 Preliminary Steps.................................................................................. 31

4.1.3.2 Recording Category and Subcategory .................................................. 33

4.1.3.3 Recording Action(s)............................................................................... 34

4.1.3.4 Recording Recommendation(s) and Outcome....................................... 35

4.1.3.5 Recording Clinical Significance and Time Taken................................... 37

4.1.3.6 Local Intervention Record: Patient history and Summary ...................... 38

4.1.3.7 Local Intervention Record: Reports ....................................................... 40

4.2 PROMISe Intervention Study Design ................................................................ 41

4.2.1 Sample Size ................................................................................................. 41

4.2.2 Overall Design of PROMISe Intervention Study............................................ 42

4.3 Selection and Randomisation of Pharmacies .................................................. 44


Page 3 of 361

4.3.1 Project Promotion ......................................................................................... 44

4.3.2 Enrolment ..................................................................................................... 46

4.3.3 PROMISe Pharmacy and Pharmacist Questionnaires .................................. 46

4.3.3.1 Pharmacy Owner or Manager Questionnaire ........................................ 47

4.3.3.2 Pharmacist demographics..................................................................... 50

4.3.3.3 Staff workload ....................................................................................... 53

4.3.4 Study Arm Allocation .................................................................................... 53

4.3.4.1 Remuneration Arms .............................................................................. 54

4.3.4.2 Aspirin Intervention Prompt Arm............................................................ 54

4.3.4.3 Observation Allocation .......................................................................... 58

4.4 Induction and preliminary training.................................................................... 60

4.4.1 On-Line Training for the PROMISe project.................................................... 60

4.4.1.1 Clinical Problem Solving Skills .............................................................. 63

4.4.1.2 DOCUMENT Training Scenarios........................................................... 64

4.4.2 WiniFRED Interface Training ........................................................................ 73

4.5 PROMISe Data Sources and Data Processing.................................................. 73

4.5.1 Accumulation of Recorded Interventions....................................................... 74

4.5.1.1 Feedback to Pharmacists and Pharmacies During the Trial .................. 75

4.5.2 Post Trial Information Collection ................................................................... 78

4.5.2.1 Pharmacist opinions of the study........................................................... 78

4.5.2.2 Focus group sessions ........................................................................... 78

4.5.2.3 In-Depth Interviews and Discussion ...................................................... 79

4.5.2.4 Further Exploration of Barriers and Facilitators to Community Pharmacy Interventions............................................................................................................. 80

4.5.3 Clinical Panel Assessment of Interventions................................................... 80

4.5.3.1 Development of Clinical Panel Assessment Methods............................ 81

4.5.3.2 Preparation of Information for Assessment by Clinical Panels............... 85

4.5.3.3 Clinical Panel Composition and Panel Access to Intervention Information 86

4.5.3.4 Clinical Panel Assessment Process ...................................................... 87

4.5.4 Sampling of interventions for assessment..................................................... 90

4.6 Economic Analysis............................................................................................. 91

4.6.1 Detailed Economic Analysis Methods ........................................................... 93

5 Results and Discussion Part 1: Nature and Frequency of Interventions and Factors Affecting Intervention Rate ................................................................................. 94

5.1 Data Collection ................................................................................................... 94

5.2 PROMISe Pharmacy and Pharmacist Recruitment .......................................... 94

5.2.1 PROMISe Pharmacist Training ..................................................................... 95

5.3 Pharmacy Demographics................................................................................... 96


Page 4 of 361

5.3.1 Characteristics of the PROMISe Pharmacies................................................ 96

5.3.1.1 Location and Size of the Pharmacy....................................................... 96

5.3.1.2 Workload of the pharmacy .................................................................... 97

5.3.1.3 Staff mix of the pharmacy ..................................................................... 99

5.3.1.4 Ownership of PROMISe Pharmacies .................................................. 101

5.3.1.5 Information Resources and Clinical Services Provided by the Pharmacy.. ........................................................................................................... 103

5.3.1.6 Quality Care Adopter Status of the Pharmacies .................................. 105

5.3.1.7 Entrepreneurial Orientation (EO) of the pharmacy .............................. 110

5.3.2 Non- PROMISe Pharmacy Characteristics.................................................. 112

5.3.2.1 Location and Size of Non –Participant Pharmacies............................. 112

5.3.2.2 Workload of the pharmacy (non-participant)........................................ 113

5.3.2.3 Staff Mix Of Non-PROMISe Pharmacies ............................................. 114

5.3.2.4 Ownership of Non-Participant Pharmacies.......................................... 116

5.3.2.5 Quality Care Adopter Status of Non-Participant Pharmacies............... 117

5.3.2.6 Entrepreneurial Orientation of Non-Participant Pharmacies ................ 117

5.3.3 Comparison of Characteristics between PROMISe and Non-Participant Pharmacies................................................................................................................ 119

5.3.3.1 Areas of similarity between the pharmacies ........................................ 119

5.3.3.2 Areas of Difference ............................................................................. 119

5.4 Pharmacist Demographics .............................................................................. 130

5.4.1 Age, Year of Graduation and Gender.......................................................... 130

5.4.2 Continuing Education and Qualifications..................................................... 132

5.4.3 Practice Profile ........................................................................................... 133

5.4.4 Role and Duration of Employment of Community Pharmacists ................... 134

5.4.5 Self-Reported Workload of Community Pharmacists .................................. 136

5.4.6 Clinical Skill Assessment ............................................................................ 138

5.4.7 Personal Views of Pharmacy ...................................................................... 143

5.5 DOCUMENT Classification System Training .................................................. 149

5.5.1 Pharmacists’ Competency with the DOCUMENT Classification System ..... 149

5.5.1.1 Categories and subcategories recorded for the 20 scenarios.............. 150

5.5.1.2 Actions recorded for the 20 scenarios ................................................. 153

5.5.1.3 Recommendations .............................................................................. 154

5.5.1.4 Clinical Significance ............................................................................ 156

5.5.1.5 Proactive or Reactive Situations ......................................................... 159

5.6 Issues with Use of Intervention Recording Software..................................... 161

5.6.1 Time Taken to Enter an Intervention........................................................... 161

5.7 Non-Clinical Intervention Data......................................................................... 163

5.7.1 Frequency and Rate of brand substitution .................................................. 163


Page 5 of 361

5.7.2 Drugs Involved in brand substitution ........................................................... 168

5.7.2.1 Grouping of the Drugs involved in Brand substitutions ........................ 171

5.8 Clinical Intervention Data................................................................................. 175

5.8.1 Frequency and Rate ................................................................................... 175

5.8.2 Categories and Subcategories of Interventions........................................... 181

5.8.3 Actions........................................................................................................ 183

5.8.4 Recommendations...................................................................................... 184

5.8.5 Outcomes ................................................................................................... 187

5.8.6 Clinical Significance.................................................................................... 188

5.8.7 Proactive vs Reactive Clinical Interventions................................................ 190

5.8.8 Drugs Involved............................................................................................ 192

5.8.8.1 Number of Clinical Interventions ......................................................... 193

5.8.8.2 Rate of Clinical Interventions............................................................... 199

5.8.8.3 Nature of Clinical Interventions for Specific Groups of Drugs .............. 204

5.8.9 Effect of Remuneration (univariate analysis)............................................... 213

5.8.10 Effect of Aspirin Intervention Prompt on Overall Clinical Intervention Rate (univariate analysis) ................................................................................................... 214

5.8.11 Effect of Observation (univariate analysis) .................................................. 217

5.8.12 Combined Effects of Remuneration, Intervention Prompt and Observation (multivariate analysis) ................................................................................................ 220

5.8.13 Pharmacy Specific Information ................................................................... 224

5.8.13.1 Entrepreneurial Orientation ................................................................. 227

5.8.13.2 QCPP Adopter Status ......................................................................... 227

5.8.14 Pharmacist Specific Factors ....................................................................... 229

5.8.14.1 Pharmacist Factors That May Affect Clinical Intervention Rate ........... 231

5.9 Overall Impact of Automated Intervention Prompt ........................................ 233

5.9.1 Cumulative interventions as a result of the Aspirin Intervention Alert .......... 234

5.9.2 Time taken to complete an aspirin intervention ........................................... 235

5.9.3 Aspirin interventions by pharmacy .............................................................. 235

5.9.4 Potential contraindications to aspirin and subsequent recommendation by the pharmacist ................................................................................................................. 238

6 Results and Discussion Part 2: Actual and Potential Value of Interventions ............ ....................................................................................................................... 242

6.1 Key Methodological Issues.............................................................................. 242

6.1.1 How do we define the value of pharmacist activity? .................................... 249

6.1.2 Defining the Factual and the Counterfactual States and the Effect of the Recommendation Made. ............................................................................................ 251

6.1.3 Accuracy of Specification of Factual and Counterfactual States.................. 252

6.1.3.1 The Narrative Provided to the Assessors ............................................ 253

6.1.3.2 The Number of Possible Consequences of the Intervention ................ 253


Page 6 of 361

6.1.3.3 The Range of Outcomes from the Intervention.................................... 254

6.1.3.4 Compliance with the Suggestions ....................................................... 254

6.1.4 Consideration of Separate Consequences and Outcomes for One Individual and for the Population. ............................................................................................... 254

6.1.5 Comparison of Different Methodologies ...................................................... 256

6.1.5.1 Rupp Method ...................................................................................... 257

6.1.5.2 Hawksworth Method............................................................................ 259

6.1.5.3 Benrimoj Method................................................................................. 260

6.1.5.4 Nesbit Method..................................................................................... 261

6.1.5.5 Dooley Method.................................................................................... 262

6.1.5.6 Buurma Method .................................................................................. 263

6.1.5.7 PROMISe Method............................................................................... 264

6.2 Overview of Economic Results ....................................................................... 265

6.3 Summary of method, data and results............................................................ 273

6.3.1 Step 1: Opportunity for Intervention ............................................................ 274

6.3.1.1 Results of Opportunity for Intervention Step........................................ 275

6.3.2 Step 2: Rate of intervention in current practice ........................................... 276

6.3.2.1 Results of Current Rate of Intervention Step ....................................... 276

6.3.3 Step 3: Value of an average intervention in current practice ....................... 277

6.3.3.1 Results of Determining Average Value Step ....................................... 279

6.3.4 Step 4: Value of Pharmacist Interventions Extrapolated to National Situation ... ................................................................................................................... 280

6.3.4.1 Results of National Extrapolation ........................................................ 280

6.3.5 Step 5: Value of improved rate of intervention. ........................................... 281

6.3.5.1 Results of Increasing Intervention Rates ............................................. 281

6.4 Data set for the economic analysis................................................................. 282

6.4.1 The Key Descriptors ................................................................................... 282

6.4.2 Tables of Pharmacy Activity........................................................................ 285

6.4.3 General indicators....................................................................................... 291

6.4.4 Value of interventions ................................................................................. 295

6.5 Economic analysis ........................................................................................... 305

6.5.1 Step 1: Opportunity for interventions........................................................... 306

6.5.2 Step 2: Current rate of intervention ............................................................. 307

6.5.3 Step 3: Average value of interventions........................................................ 308

6.5.4 Step 4: National value of interventions....................................................... 310

6.5.5 Step 5 Value of improved interventions....................................................... 310

6.6 Results of Economic analyses ........................................................................ 312

6.6.1 Estimate of current value of pharmacist activity .......................................... 319

6.6.2 Estimate of improved rate of activity – aspirin popup .................................. 320


Page 7 of 361

6.6.3 Estimate of improvements resulting from reduced pharmacist workload ..... 323

6.6.4 Estimate of improvement resulting from maximum possible rate of interventions............................................................................................................... 325

7 Results and Discussion Part 3: Barriers and Facilitators to Performing and Recording Clinical Interventions .................................................................................... 328

7.1 Barriers to performing clinical interventions ................................................. 328

7.1.1 Definition and Identification of Clinical Interventions ................................... 328

7.1.2 Business Culture and the Individuals Who Operate in These Systems ....... 329

7.1.3 Time and workload ..................................................................................... 330

7.1.4 Clinical knowledge and continuing education.............................................. 331

7.1.5 Other barriers to performing clinical interventions ....................................... 333

7.2 Facilitators to performing clinical interventions ............................................ 333

7.2.1 Professional satisfaction ............................................................................. 334

7.2.2 Recognition for providing clinical services................................................... 334

7.2.3 Information Continuity................................................................................. 335

7.2.4 Continuing education .................................................................................. 335

7.2.5 Work environment....................................................................................... 335

7.3 National Survey of Clinical Intervention Documentation............................... 336

7.4 Opinions Regarding Remuneration for Interventions.................................... 347

7.4.1 Preferred Payment Models ......................................................................... 347

7.4.2 Remuneration Models Assessed in the Post Study Questionnaire.............. 348

7.4.3 Remuneration Models Explored During the I-view Telephone Survey......... 350

7.4.4 Preferred Rates of Remuneration ............................................................... 351

7.4.4.1 Preferred Rate if a Payment to Each Pharmacy for Each Intervention is Made ........................................................................................................... 351

7.4.4.2 Preferred Rate if a Payment to Each Pharmacist For Each Intervention is Made ........................................................................................................... 351

7.4.4.3 Preferred Rate if a Payment to Each Pharmacy for Selected Interventions (Severe and Moderate only) is Made...................................................................... 352

7.4.4.4 Preferred Rate if a Payment to Each Pharmacist for Selected Interventions (Severe or Moderate only) is Made ................................................... 353

7.4.4.5 Preferred Rate if an Increase in the Dispensing Fee for Prescriptions for High-risk Drugs Occurred....................................................................................... 354

8 Potential Improvements to the Intervention Recording System........................... 356

8.1 Software and IT Changes Suggested.............................................................. 356

8.1.1 System Requirements ................................................................................ 356

8.1.2 Improvement of Software Installation and Testing....................................... 356

8.1.3 Incorporate More Educational Alerts ........................................................... 357

8.1.4 Adjustments to the Interface ....................................................................... 357

8.2 Changes to the PROMISe System Overall ...................................................... 358

8.2.1 Refinement of Classification ....................................................................... 358


Page 8 of 361

8.2.2 Mentor program .......................................................................................... 358

8.2.3 Educational alerts in the dispensing software ............................................. 358

8.2.4 Online training ............................................................................................ 358

8.2.5 Continuing Education.................................................................................. 358

8.2.6 Workshop sessions..................................................................................... 359

9 Conclusions and Recommendations...................................................................... 360

APPENDICES: Refer to Volume Two


Page 9 of 361

1 Executive Summary

1.1 Introduction and Aims:

At present, little is known about community pharmacists’ role in improving consumers’ health by

identifying and resolving medication related issues. The frequency with which such activities occur and

their potential value to the Australian community is of particular interest. The Australian Government

Department of Health and Ageing through the Third Community Pharmacy Agreement Research and

Development Program has funded two related projects to address the area of documentation and

value of pharmacists’ clinical activities.

The first of these projects (Project ID 2003-504) resulted in the development of an innovative

documentation and electronic communication system for medication incidents and pharmacists’

professional interventions (Pharmacy Recording of Medication Incidents and Services Electronically or

PROMISe). The system interfaces seamlessly with two pharmacy dispensing systems, Rex and

WiniFRED, and sends encrypted, HL7-compliant messages to a secure server for collation and

analysis (final report for this first project available at

http://beta.guild.org.au/research/project_display.asp?id=269 ).

This second phase of the research (Project ID 2003-519) encompassed further refinement of the

electronic documentation and communications system and implementation of the PROMISe system

into a sample of pharmacies in order to:

1. determine the frequency with which Australian community pharmacists resolve or prevent drug

related problems

2. evaluate the nature of the problems and the factors influencing the frequency and severity of

the problems

3. estimate the potential value of these clinical interventions in health and economic terms

4. explore barriers and facilitators to the performance and documentation of clinical interventions,

including

a. pharmacist and pharmacy characteristics (e.g. workload)

b. remuneration,

c. observation and

d. the presence of an intervention prompting mechanism


Page 10 of 361

Methods:

1.1.1 Data Collection

Fifty-two community pharmacies in Melbourne, using the WiniFRED dispensing system, participated in

the study. Based on a number of parameters, including location and annual turnover, these

pharmacies seemed broadly representative of Victorian and pharmacies nationally. Full training, onsite

and online support, and remuneration were provided for the participating pharmacists. During the

phases of data collection, between 21st April to 17

th June 2005, pharmacies were randomised to

receive payment for clinical interventions ($15 for each intervention submitted) and allocated to

receive the intervention prompt (see Figure 0-1).

52 Pharmacies

Enrolled

23 Paid in Phase 1

29 Unpaid in Phase 1

11 Observation Pharmacist

12 No observation

Pharmacist

18 No observation

Pharmacist

11Observation Pharmacist

9 No Intervention

Prompt

9 Intervention

Prompt

6 No Intervention

Prompt

5 Intervention

Prompt

2 No Intervention

Prompt

9 Intervention

Prompt

5 No Intervention

Prompt

7 Intervention

Prompt

52 Pharmacies

Paid in Phase 3

Phase 1(2 weeks)

Phase 2(2 weeks)

Phase 3(4 weeks)


29 Paid in Phase 2


12 No observation

Pharmacist

18 No observation

Pharmacist


9 No Intervention

Prompt

9 Intervention

Prompt

6 No Intervention

Prompt

5 Intervention

Prompt

2 No Intervention

Prompt

9 Intervention

Prompt

5 No Intervention

Prompt

7 Intervention

Prompt

Observation Pharmacist Present (3 weeks)

Intervention Prompt Active (4 weeks)

Figure 0-1: Randomisation Schema for PROMISe Intervention Study

The prompt was developed to test the hypothesis that a computer-based reminder could increase

intervention rates associated with a particular type of intervention. The educational alert related to the

use of low-dose aspirin (or other antiplatelet agent) for cardiovascular and cerebral vascular event

prevention in high-risk diabetic patients. The alert was automatically triggered when any oral

antidiabetic agent was selected for dispensing.


Page 11 of 361

1.1.2 Determination of Value of Interventions

The Project Team developed a unique web-based clinical assessment method which takes into

account the probability of a consequence occurring (with the intervention and also without the

intervention), the likely outcomes from the drug-related problem and their potential severity, and also

the “attributability” of the intervention to the pharmacist (i.e. the likelihood that no other health

professional would have detected and resolved the drug-related problem). Four clinical panels,

containing physicians, general practitioners and pharmacists, independently reviewed a total of 291

interventions from the dataset. A comprehensive economic analysis was performed to estimate:

• the economic value to the pharmacy of the pharmacists’ intervention (essentially the opportunity

cost of the pharmacists’ time),

• the economic value to the health care system of the clinical interventions, and

• the economic value of changing the rate of clinical interventions.

The economic evaluation was primarily conducted from the perspective of the health care payer.

Included were costs associated with additional time spent by pharmacists, communication costs

associated with contacting health professionals or patients/carers etc., costs of hospitalisation, and

general practitioner and specialist consultations. The evaluation was intentionally conservative in its

approach, and if the reader of the report has access to the electronic version of the report and the

spread sheet model, they can easily change the assumptions within the spread sheet.

1.2 Results:

Overall, there were 2,396 clinical interventions recorded during the PROMISe study. During this

period, 435,520 prescriptions were dispensed (a rate of approximately 0.55 clinical interventions per

100 prescriptions or approximately one intervention for every 200 prescriptions) for 258,979 patients

(an intervention rate of 0.92 interventions per 100 patients). There was, however, a number of

pharmacies whose intervention rates were significantly higher than this (range 0 to 6.99 clinical

interventions per 100 prescriptions).

Despite a range of information technology issues, the software was reasonably well received and

those pharmacists who used it regularly rapidly became proficient at recording interventions. Half of

the pharmacists were able to record their interventions in 1 minute or less during the third week of the

trial (approximately twice as fast as earlier in the trial). The presence of the educational intervention

prompt and remuneration were each associated with significantly higher rates of recorded clinical

interventions in the early phase of the study. The daily rates of recorded clinical interventions,

however, gradually declined as the study continued.

Eighty percent of the interventions were considered to be proactive i.e. were initiated by the

pharmacist and were not necessary to be undertaken in order to dispense the medication. The

majority of clinical interventions were one of three categories: drug selection problems (22.7%),

dosage problems (19.4%) or education or information problems (17.4%). Drug groups commonly

associated with clinical interventions were antibiotics, drugs for diabetes, cardiovascular drugs and


Page 12 of 361

drugs for respiratory disorders. Almost one-third of the clinical interventions were classified as either of

moderate or severe level of clinical significance by the recording pharmacist. In almost 90% of cases,

the pharmacist investigated the drug-related problem by discussing the issue with the patient or the

carer. In one-third of cases, the pharmacist contacted the prescriber in order to clarify the problem.

Multiple actions were frequent, and the average number of actions per intervention was 1.87. Over

80% of the recommendations made by the pharmacists were indicated as being accepted.

Without the action of the pharmacist, in an average of 72% of cases there would have been no other

health professional who would have performed the intervention. The economic value of the

consequences of the pharmacist intervention was reduced to account for this. The clinical and

economic analysis suggests that the value of Australian community pharmacist interventions related to

prescription medication, in terms of financial costs to the health system prevented, is in the order of

$350M each year, or $17.50 per capita. In addition, around 262,000 hospital bed-days are avoided

(1.3 days per 1000 population) and 53.1M days of adverse health impact are avoided (2.7 days per

capita) per annum. Our estimate is that 0.7 hours of a pharmacist’s time is spent undertaking 6.9

interventions for every 1000 prescriptions (extrapolates to 154,000 hours undertaking 1.61M

interventions nationally each year). For every hour a pharmacist works, their interventions prevent

$17.60 in medical and hospital costs, and for every 100 hours worked, their interventions prevent 1.3

days in hospital.

As a result of each clinical intervention by a pharmacist, there is a mean reduction of:

• 44 days in a lowered health status (5 days of level 3, 21 days of level 2 and 18 days of level 3),

• 0.22 days in hospital at a cost of $174

• 1.0 GP consultations and 0.23 specialist consultations at a cost of $59 to MBS,

• further investigations at a cost of $57 to MBS, and

• $290 in total costs (MBS and hospital combined).

We believe that our estimates of the current value are conservative due to the techniques used for the

estimates. The upper and lower bounds of the annual estimates are shown in Table 1.2-1 below:

Annual Estimate of Current Value of Interventions Value Indicator

Minimum Base (Conservative) Maximum

Total Direct Costs Prevented $182M $349M $623M

Hospital Admission Days Avoided

136,000 262,000 468,000

GP or Specialist Consultations Avoided

0.77M 1.48M 2.64M

Days of Adverse Health Impact Avoided

27.6M 53.1M 94.8M

Table 1.2-1: Annual Estimate of Current Value of Clinical Interventions


Page 13 of 361

Our results also indicate that pharmacists may identify and act upon only a third to a half of all possible

opportunities for clinical intervention, and that the costs of these omissions is likely to be greater than

the benefits of the current rate of intervention.

Our estimate of the additional annual value of increasing pharmacists’ intervention rates to those

achieved in many of the pharmacies in the PROMISe project is:

• $606M in medical and hospital costs,

• 749,000 hospital admission days avoided,

• 2.26 GP or specialist consultations avoided, and

• 91.8M days of adverse health impact avoided.

Again, these estimates are conservative, and the upper and lower bounds are shown in Table 1.2-2

below.

Annual Estimates of Additional Value of Increased Interventions Value Indicator


Total Direct Costs Prevented

$563M $606M $740M


652,000 749,000 850,000

GP or Specialist Consultations Avoided

2.10M 2.26M 2.76M

Days of Adverse Health Impact Avoided

85.4M 91.8M 112M

Table 1.2-2: Annual Estimate of Additional Value of Increased Rate of Interventions

1.3 Conclusions:

The current value of Australian community pharmacists’ interventions in both health and financial

terms is high. However, there is considerable scope for increasing this impact; it is likely that both the

existing rate and the financial value of pharmacists’ interventions could be increased three-fold.

Automated educational alerts within computerised dispensing systems possess significant potential for

increasing pharmacists’ intervention rates, and thereby improving the quality use of medicines. Other

methods that may increase the rate of intervention and are worthy of examination in more detail

include reducing pharmacist workloads (increasing staff levels); introducing payments for meeting

selected intervention targets; and practice-based payments for improved intervention rates.


Page 14 of 361

2 Authors and Acknowledgements

2.1 Lead Investigator

Professor Gregory Peterson

Unit for Medication Outcomes Research and Education

University of Tasmania

Private Bag 83

Hobart TAS 7001

(03) 62262197

[email protected]

2.2 Principal Investigators

Mr Peter Tenni

Senior Research Fellow



Private Bag 83

Hobart TAS 7001

(03) 62261005

[email protected]

Ms Helen Kruup

Project Manager



Private Bag 83

Hobart TAS 7001

(03) 62267526

[email protected]

Dr Omar Hasan

Research Manager



Private Bag 83

Hobart TAS 7001

(03) 62262191

omar,[email protected]


Page 15 of 361

Ms Brita Pekarsky

Senior Research Fellow

Centre for Regulations and Market Analysis

University of South Australia

City West Campus

Adelaide SA 8000

(08) 83020979

[email protected]

Mr James Reeve

Manager, Pharmaceutical Decision Support

National Prescribing Service Ltd

Level 7, 418A Elizabeth St

Surry Hills NSW 2010

(02) 8217 8700

[email protected]

2.3 Acknowledgements

The following persons have provided valuable assistance in the development, testing, analysis and

evaluation of various aspects of the project:

• Mr Rod Unmack, PCA/NU Systems

• Mr Keith Gordjin, Developer, PCA/NU Systems

• Mr Brett O’Halloran, Developer, PCA/NU Systems

• Mr George Pavlidis, PCA/NU Systems

• Mr Ian DeBoos, DeBoos Associates

• Mr Patrick Banks, Phoenix Computer Systems

• Mr Michael Ryan, Michael Ryan and Associates


Page 16 of 361

3 Introduction and Objectives

At present, little is known about community pharmacists’ role in improving consumers’ health by

identifying and resolving medication related issues. The frequency with which such activities occur and

their potential value to the Australian community is of particular interest. The Australian Government

Department of Health and Ageing through the Third Community Pharmacy Agreement Research and

Development Program has funded two related projects to address the area of documentation and

value of pharmacists’ clinical activities.

The first of these projects (Project ID 2003-504) resulted in the development of an innovative

documentation and electronic communication system for medication incidents and pharmacists’

professional interventions (Pharmacy Recording of Medication Incidents and Services Electronically or

PROMISe). The system interfaces seamlessly with two pharmacy dispensing systems, Rex and

WiniFRED, and sends encrypted, HL7-compliant messages to a secure server for collation and

analysis (final report for this first project available at

http://beta.guild.org.au/research/project_display.asp?id=269 ).

This second phase of the research (Project ID 2003-519) encompasses further refinement of the

electronic documentation and communications system and implementation of the PROMISe system

into a sample of pharmacies in order to:

• determine the frequency with which Australian community pharmacists resolve or prevent drug

related problems

• evaluate the nature of the problems and the factors influencing the frequency and severity of

the problems

• estimate the potential value of these clinical interventions in health and economic terms

• explore barriers and facilitators to the performance and documentation of clinical interventions,

including

• pharmacist and pharmacy characteristics (e.g. workload)

• remuneration,

• observation and

• the presence of an intervention prompting mechanism


Page 17 of 361

4 Methods

4.1 Modifications as a Result of PROMISe Pilot Study

The research team had previously developed and pilot tested a system for the electronic recording,

collation and management of medication incidents in community pharmacies.1 The result was an

electronic communications system (Pharmacy Recording of Medication Incidents and Services or

PROMISe) that interfaces seamlessly with two dispensing systems (Rex and WiniFRED) and sends

encrypted, HL7 compliant messages to a secure server. The system was subsequently installed in

seven pharmacies in Tasmania:

• to evaluate aspects of the recording software and usability of the categorisation system,

• to test a number of different questionnaires and activity surveys, and

• to provide some preliminary data for examination.

The methods for this second project, the PROMISe Intervention Study, were based in part on

experience gained during the first PROMISe project. Sections 4.1.1 to 4.1.3 outline changes to the

documentation and recording software that were made as a result of the pilot studies.

4.1.1 Development and Modifications to the DOCUMENT categorisation system

A classification system for drug-related problems (interventions) and their resolution was developed by

the Project Team in the first phase of the PROMISe project.1. This system is termed the DOCUMENT

classification system and consists of a flexible, hierarchical set of classification codes.

The classification system is one of the few in the world that incorporates a description of the type of

problem, the investigations and actions undertaken by the pharmacist to clarify the problem and the

recommendations made to resolve the problem. In addition, there are classifications for the clinical

significance of the problem and the acceptance of the recommendation(s) made.2

The version of DOCUMENT used in the pilot study is outlined in Table 4.1-1 and the detailed complete

version (with scope notes), is shown in Appendix 1.

1 Community Pharmacy Medication Incident Reporting and Management Systems (CPMIRMS) also

known as PROMISe: Pharmacy Recording of Medication Incidents and Services electronic

documentation system: Final Report 2004 available at

http://beta.guild.org.au/research/project_display.asp?id=269

2 Peterson G, Tenni P. Identifying, prioritising and documenting drug-related problems. Australian

Pharmacist 2004: 23(10): 706-9.


Page 18 of 361

• Category of Problem

– seven major categories, each of which has multiple subcategories

• Action(s) to Investigate Problem

– Seven types of investigation and enquiry into the problem

• Significance of Problem

– Five levels of severity from nil to likely hospitalisation

• Recommendation(s) to Resolve Problem

– 17 recommendations grouped in four main areas

• Outcome

– acceptance of recommendation(s)

Table 4.1-1: Outline of DOCUMENT Drug Related Problem Classification System

The system was tested during a data collection period of two weeks in seven pharmacies in southern

Tasmania during 2004. During the two week pilot study, over 500 interventions were recorded by 20

pharmacists. Of these interventions 352 were of a clinical nature. The results of the pilot study were

used to guide the modification of the classification system and other aspects of the PROMISe

intervention study including refinement of questionnaires and the recording software (see section

4.1.2). The following sections outline the different facets of the classification system and how the

DOCUMENT drug related problem classification system was modified for use in the PROMISe

intervention Study.

4.1.1.1 Type of Drug Related Problem (Category and Subcategory)

The first step in classifying an intervention is determining the type of drug related problem. In the

DOCUMENT classification sytem, there are a range of main categories and subcategories to select

from. The frequency of recording of the categories and subcategories during the pilot study is shown in

Table 4.1-2.


Page 19 of 361

Category Subcategory No % OverallTotal for

Category

% within

Category

Duplication 8 2.3% 18.6%

Drug interaction 22 6.3% 51.2%

Wrong dosage form 4 1.1% 9.3%

Previous ADR/allergy 2 0.6% 4.7%

Other drug selection problem 7 2.0% 16.3%

Dose too high 9 2.6% 26.5%

Dose too low 11 3.1% 32.4%

Wrong frequency 3 0.9% 8.8%

Other Dose Problem 11 3.1% 32.4%

Potential drug abuse 4 1.1% 16.0%

Taking too little 8 2.3% 32.0%

Taking too much 5 1.4% 20.0%

Difficulty using dosage form 2 0.6% 8.0%

Other Compliance Problem 6 1.7% 24.0%

Condition not adequately treated 17 4.8% 70.8%

Preventive therapy required 2 0.6% 8.3%

Other Untreated indication Problem 5 1.4% 20.8%

Monitoring Drug Levels 0 0.0% 0.0%

Laboratory Monitoring 15 4.3% 60.0%

Non-Laboratory monitoring 7 2.0% 28.0%

Other Monitoring Problem 3 0.9% 12.0%

Patient drug information request 45 12.8% 40.9%

Confusion about therapy or condition 14 4.0% 12.7%

Demonstration of device 10 2.8% 9.1%

Disease management or advice 26 7.4% 23.6%

Other Education or Information

Problem15 4.3% 13.6%

Non-clinical Not sub-classified 76 21.6% 76 100.0%

Caused by dose too high 2 0.6% 13.3%

Caused by drug interaction 2 0.6% 13.3%

Other Toxicity/Adverse Effect

problem11 3.1% 73.3%

352 100.0% 352

25

34

43

15

110

24

25

Education or

Information

Toxicity or

Adverse

reaction

Total

Monitoring

Drug selection

Over or

underdose

Prescribed

Compliance

Untreated

indications

Table 4.1-2: Frequency of DOCUMENT Categories in PROMISe Pilot Study

A number of sub-categories were used infrequently in the PROMISe pilot study, and if these were

deemed unlikely to be used in a larger study, these subcategories were revised. For example it was

found that “Monitoring - Drug Levels” was not used and it was felt that problems that were of the drug

level monitoring type could be included in “Monitoring – Laboratory”. A review of each of the “Other”

subcategories was also undertaken. From this, problem types that occurred commonly were allocated

new codes, titles and definitions. One new subcategory created was “Toxicity evident”, and an

example of an intervention that falls into this class from the PROMISe data is shown in Figure 4.1-1.


Page 20 of 361

Toxicity or adverse reaction - toxicity / adverse reaction evident Summary Problem; Patient experiencing adverse effect whilst taking bupropion Male patient (>65) presents a repeat prescription for bupropion. The patient mentions to the pharmacist that when he takes the second dose of bupropion he has been experiencing dizziness, visual disturbances and the sensation of spinning. The pharmacist recommended that the patient stop taking the bupropion. The patient accepted this recommendation Outcome; Patient no longer experiencing adverse effects of medication Category Toxicity or adverse reaction Subcategory Toxicity / adverse reaction evident Actions Investigation: Patient History Recommendations Drug change Discussion with patient or carer Outcome Accepted Significance Moderate

Figure 4.1-1: Example of "Toxicity Evident" Intervention

For each of the subcategories detailed scope notes were prepared. These notes include a definition of

the subcategory, details and examples of when the subcategory should be selected and examples of

when other categories are more appropriate. For each modification made during the review of the

DOCUMENT system, the relevant scope notes were also updated. The scope notes form the basis of

the ‘help’ section in both the online training and the recording interface.

The full version of DOCUMENT including the scope notes used for the PROMISe intervention study

can be seen in Appendix 2. An abbreviated version of the categories and subcategories can be seen

in Table 4.1-3.


Page 21 of 361

Drug selection

Problems related to the choice of drug prescribed or taken

Duplication (D1) Drug interaction (D2) Wrong drug (D3) Wrong dosage form (D4) Other drug selection problem (D0)

Over or underdose prescribed

Problems related to the prescribed dose or schedule of the drug

Dose too high (O1) Dose too low (O2) Other Dose Problem (O0)

Compliance

Problems related to the way the patient takes the medication

Taking too little (C1) Taking too much (C2) Intentional drug misuse (C3) Difficulty using dosage form (C4) Other Compliance Problem (C0)

Untreated indications

Problems relating to actual or potential conditions that require management

Condition not adequately treated (U1) Preventive therapy required (U2) Other Untreated indication Problem (U0)

Monitoring

Problems related to monitoring the efficacy or adverse effects of a drug

Laboratory Monitoring (M2) Non-Laboratory monitoring (M3) Other Monitoring Problem (M0)

Education or Information

Problems related to knowledge of the disease or its management

Patient drug information request (E1) Confusion about therapy (E2) Demonstration of device (E3) Disease management or advice (E4) Other Education or Information Problem (E0)

Non-clinical

Problems related to administrative aspects of the prescription

Toxicity or Adverse reaction

Problems related to the presence of signs or symptoms which are suspected to be related to an adverse effect of the drug

Toxicity caused by dose (T1) Toxicity caused by drug interaction (T2) Toxicity evident (T3) Other Toxicity/Adverse Effect problem (T0)

Table 4.1-3: DOCUMENT Categories and Subcategories Used for the PROMISe Intervention Study


Page 22 of 361

4.1.1.2 Actions

Once a drug related problem has been identified, actions are undertaken to clarify and investigate the

problem. In the DOCUMENT system, any number of such actions can be documented for each

intervention. The frequency of actions recorded in the PROMISe pilot study are shown in Table 4.1-4.

Investigation : Written MaterialsInvestigation: SoftwareInvestigation: InternetInvestigation: Drug Information CentreInvestigation: OtherContact PrescriberDiscuss with PatientNo DiscussionOther

Table 4.1-4: Frequency of Actions to Investigate Problems from PROMISe Pilot Study

A review of the action codes used in the Pilot study was undertaken. The version used in the

Melbourne trial can be seen in Table 4.1-5.

Investigation: written material (A1) Investigation: Software (A2) Investigation: Patient History (A3) Investigation: Other (A4) Contacted prescriber (A5) Discussion with patient or carer (A6) Corrected without discussion (A7)

Table 4.1-5: Action Definitions Used in PROMISe Intervention Study

During examination of the “Action Investigation: other” category in the PROMISe pilot study, it was

found that reviewing the patients' history was a common action. A new code and definition was added

for this action for inclusion into the categorisation system used in the PROMISe intervention study. As

in the pilot study, multiple actions could be recorded for each problem, and if the same action was

undertaken more than once the pharmacist could record it on multiple occasions.


Page 23 of 361

4.1.1.3 Recommendations

If, after investigation, the drug related problem requires resolution, then a recommendation for a change or modification needs to be made. In the DOCUMENT classification system, there are

a number of recommendations that can be selected from. There were 15 recommendation types available for use in the PROMISe pilot study and the frequency of their use is shown in

Table 4.1-6.

Education/Counselling SessionChange Dose of DrugChange DrugCease DrugChange Formulation of DrugNon- Laboratory MonitoringAddition of a DrugChange BrandChange Frequency or Schedule of DrugRefer to PrescriberLaboratory MonitoringMedication ReviewDose Administration AidNo Recommendation requiredOther

Table 4.1-6: Frequency of Recommendations to Resolve Problems from PROMISe Pilot Study

Again, modification of the recommendation types was undertaken based on feedback and frequency

of selection. Pharmacists who participated in the PROMISe pilot study commented that there was an

extensive list of recommendations to select from. To improve this aspect of the recording, the

recommendations were grouped into five main types;

• a change in therapy

• a referral required

• provision of information

• monitoring

• no recommendation necessary

The specifics of the recommended changes before and after the intervention, were recorded in the

PROMISe study. In order to achieve this, software modifications were made (see section 4.1.3.4). This

modification allowed the pharmacist to record which medication they altered, and how. For example,

ceasing celecoxib where it was found the patient was taking duplicate non-steroidal anti-inflammatory

agents (NSAIDs).


Page 24 of 361

The recommendation selections available for the PROMISe intervention study are outlined in Table

4.1-7.

A Change in Therapy

Dose change (R1) * Drug change (R2) * Drug formulation change(R3) * Drug brand change (R4) * Dose frequency/schedule change (R5) * Prescription not dispensed (R6) Other changes to therapy (R7)

A Referral Required

Refer to prescriber (R8) Refer to hospital (R9) Refer for medication review (R10) Other referral required (R11)

Provision of information

Education/counselling session (R12) Written summary of medications (R13) Commence dose administration aid (R14) Other written information (R15)

Monitoring Required

Monitoring: non-laboratory (R16) * Monitoring: Laboratory test (R17) *

No recommendation

No recommendation necessary (R18)

*these recommendations are linked to “before and after” changes recorded automatically elsewhere in the system

Table 4.1-7: Recommendation Definitions as Used for PROMISe Intervention Study

4.1.1.4 Outcome

If a recommendation is made to resolve a drug related problem, it is appropriate to document whether

the recommendation has been accepted (by the doctor or patient) or not (i.e. the outcome of the

recommendation). There were no changes made to the categories for acceptance of pharmacist’s

resolution of the problem (outcome) used in the PROMISe pilot study.

The potential outcomes are that the recommendation(s) are accepted, partially accepted (that is, only

some of the multiple recommendations were accepted), not accepted or the outcome may be unknown

at the time of recording the intervention. Definitions and scope notes for outcomes are shown in Table

4.1-8.


Page 25 of 361

Unknown

Definition: When the pharmacist is unaware of what happened after he made the recommendation(s).

Accepted

Definition: When all of the recommendation(s) that the pharmacist makes are accepted

Partially Accepted

Definition: When the pharmacist makes multiple recommendations, and only some of the recommendations that were made are accepted.

Not accepted

Definition: When all of the recommendation(s) that the pharmacist makes are rejected Table 4.1-8: Outcome Definitions Used for PROMISe Intervention Study

4.1.1.5 Clinical Significance

There were only minor changes made from the definitions for the different levels of clinical significance

used in the PROMISe pilot study. The final definitions used in the PROMISe intervention study are

shown in Table 4.1-9.

Nil (S0)

Definition: When there is no consequence to the patient.

Low (S1)

Definition: When the consequence to the patient are related to costs or information only

Mild (S2)

Definition: When the consequences to the patient are that they have improved a minor symptom, or if the intervention had not occurred they would have developed a minor symptom. The symptom should be such that it does not require a doctor’s visit to treat.

Moderate (S3)

Definition: When if the intervention did not occur, it was likely that the patient would have had to go to the doctor because of the consequences. Also covers the situation where you need to refer the patient to the doctor because of the seriousness of the situation.

High (S4)

Definition: When if the intervention did not occur, it was likely that the patient would have had to go to a hospital because of the consequences. Also covers the situation where you need to refer the patient to a hospital because of the seriousness of the situation. When if the intervention did not occur, it was likely the patient would have had to receive assistance from a regular nurse visit, or would have had to been placed into residential care of some sort. Also includes the situation where the intervention prevents the additional nursing care or delays the admission to residential care.

Table 4.1-9: Significance Definitions USed in PROMISe Intervention Study


Page 26 of 361

4.1.1.6 Identification of the Problem (Proactiveness)

In considering the results from the PROMISe pilot study, another form of classification of the

intervention became evident, whether the intervention was proactive or reactive. Reactive

interventions were those that were either not initiated by the pharmacist, or required addressing before

the prescription was able to be dispensed. Proactive interventions were initiated by the pharmacist and

were not necessary to be undertaken in order to dispense the medication.

Using this definition, 254 of the 352 interventions in the pilot study were reviewed and coded as

reactive or proactive. The results from the PROMISe pilot study indicated that 33.5% (85) of the

interventions were proactive in nature and 66.5% (169) of the interventions were reactive in nature.

Proactive interventions were found to be of higher significance and were more “discretionary” in the

pilot study.

Given the predicted numbers of interventions for the PROMISe intervention study, it would not be

possible to individually re-code each intervention. Therefore, pharmacists who participated in the

PROMISE intervention study were asked to indicate who initially identified the problem as a guide to

the proactiveness of the intervention. Training on this aspect of the classification of the intervention

was included in the on-line DOCUMENT training (see section 4.4.1).

4.1.2 Modifications to Intervention Recording Software

The recording software for the DOCUMENT classification system was incorporated into both the Rex

and WiniFRED dispensing programs for the pilot study. From a review of the functionality and use of

the software during the pilot study, a number of modifications were formulated. These changes were

collated and itemised before being forwarded to the software developers. Appendix 3 outlines the

specific requirements provided to PCA NU Systems and Phoenix for the project.

Based on this document, the software developed in the pilot study was modified and re-incorporated

into the PCA NU Systems WiniFRED dispensing program. The basic outline of the system architecture

is shown in Figure 4.1-2. The system consists of a user interface and local management of

information (at the pharmacy level), encryption, de-identification and secure communication to a

remote repository of data. Full details of technical and functional specifications for the user interface,

the repository database and HL7 messaging formats are included in Appendix 4, Appendix 5 and

Appendix 6 respectively. Additional technical information regarding changes made specifically by PCA

NU Systems are shown in Appendix 7.


Page 27 of 361

PROMISeSERVER

PROMISeDATABASE

CommServer

FIREWALLINTERNET

Pharmacy:

Dispense System

•WiniFRED

•PROMISe Interface

•Record Intervention Info

•HL7 Message Building

CommServer:

•PKI Encryption

•Transmission of data

Communication:

HL7:

•Encrypted HL7 Intervention

PKI:

•PKI Encryption

SMIME Packeting:

•Data Transmission

Server:

Secure Firewall:

•Authorised access only

PROMISe Server:

•User Authentication

•PKI Decryption

•HL7 Structure Rule Checks

•HL7 Business Rule Checks

Database:

•De-identified Intervention repository

WiniFRED Dispense

PROMISeSERVER

PROMISeDATABASE

CommServer

FIREWALLINTERNET

Pharmacy:

Dispense System

•WiniFRED

•PROMISe Interface

•Record Intervention Info

•HL7 Message Building

CommServer:

•PKI Encryption

•Transmission of data

Communication:

HL7:

•Encrypted HL7 Intervention

PKI:

•PKI Encryption

SMIME Packeting:

•Data Transmission

Server:

Secure Firewall:

•Authorised access only

PROMISe Server:

•User Authentication

•PKI Decryption

•HL7 Structure Rule Checks

•HL7 Business Rule Checks

Database:

•De-identified Intervention repository

WiniFRED Dispense

Figure 4.1-2: System Architecture for Data Collection, Transmission and Storage for the PROMISe Intervention Study

4.1.2.1 Data Management (Local Storage, Communication and Repository Storage)

To facilitate the transfer of information from the pharmacy to the remote repository, a communications

server application (Comm Server) was developed (see Figure 4.1-3).

The Comm Server utilised in the PROMISe project was adapted from an existing PCA NU Systems

application. The Comm Server was designed to be extendable through the use of “plug-in” architecture

that allows functionality to be added without changes to the application. For the PROMISe project, an

intervention recording plug-in was created. The plug-in was split into three separate components;

• “Control Module”

o This module interfaced between the Comm Server and the other two components.

This component also used the Health Insurance Commission’s Public Key

Infrastructure (HIC PKI) library to perform the public key encryption of messages

before they were sent over the Internet to the University server and decryption of

return messages.

• “Data Module”

o this module interfaced with the local WiniFRED data store in the pharmacy computer

to extract required information about interventions for the building of the HL7

messages


Page 28 of 361

• “HL7 Module”

o This module was used to construct the HL7 messages and return the message to the

“Control Module”. This module was also responsible for “deconstructing” the HL7

messages returned from the University server to determine success or failure of the

data transmission.

UTAS server

Control Module

Data Module

Data Store

HL7 Module

CommServer

Encrypted data transfer

Figure 4.1-3: Transfer of Encrypted Data and Storage Within the Pharmacy System

4.1.2.2 Communication: External Data Transfer

The architecture of the system used for transfer of the data from the pharmacy to the remote

repository is shown in Figure 4.1-4.


Page 29 of 361

Figure 4.1-4: Transfer of Encrypted Information to the Server

The pharmacy dispensing system incorporates integrated features that allow it to record an

intervention locally and then send the details to the Repository Server. The pharmacy’s de-identified

daily dispensing history was also sent to the server in the same fashion. Information sent to the

repository is formatted using HL7 messaging as it is rapidly emerging as the accepted standard for

transfer of health-related data both in Australia and internationally. These messages are encrypted for

security with PKI, then encapsulated in a HTTP message as an S/MIME attachment. This

methodology ensures ease of transmission of the message data over the Internet using existing

transport protocols and technologies (TCP/IP).

The main firewall is administered by the University of Tasmania. Its purpose is to block unauthorised

access to the internal networks and services. The firewall has been configured to allow standard

incoming HTTP requests to the project’s server. This grants access to the demilitarised zone (DMZ)

where the Apache Proxy Server operates. HTTP-encapsulated S/MIME requests come into the

Apache Proxy Server. The proxy server can view header information within this packet in order to

determine the appropriate destination of the encrypted message. Once the destination has been

resolved the proxy server retransmits the request through a specific port opened in the University inner

firewall to the repository S/MIME server for processing. The operation of the University inner firewall

blocks access to the internal network from any of the DMZ systems unless specified otherwise.


Page 30 of 361

Messages from pharmacy dispensing systems are sent as HL7 messages encrypted in S/MIME

packets and encapsulated in HTTP. The S/MIME server unwraps and decrypts these requests to

retrieve the original HL7 messages. In the process it retrieves the PKI certificate details from the

encrypted message. The HL7 and PKI details are then sent to the repository server for processing of

the transmitted data.

All processing of HL7 messages from the dispensing systems happens in the repository server. Once

the HL7 message is retrieved from the incoming packet it is imported into the HL7 parser. This parser

prepares the message for validation and decoding, allowing for faster processing of the message.

Once the message has been successfully parsed it is then checked for compliance against the HL7

structure and business rules. This enforces compliance from the sending application and ensures

integrity of the data within the message. If the message successfully passes these checks, the

contents are then processed in order to store the new data into the SQL Database. When this is

complete, a reply HL7 message is generated and returned to the sending application. If any errors

occur during these processes, an appropriate HL7 or HTTP error response message will be generated

and returned. All incoming and outgoing messages can be logged to file along with errors and

transaction logs.

The SQL database server manages the different databases used within the system. These databases

hold all the data received from the pharmacy dispensing systems. The database allows for efficient

central storage of all data collected and also assists reporting and analysis of the captured information.

Any modifications to data are tracked using internal audit techniques.

4.1.3 Documenting Interventions in WiniFRED Dispense

A number of technical changes were made to the “behind the scenes” aspects of the WiniFRED

dispense system to accommodate the PROMISe project requirements.

The major new functionality in WiniFRED dispense for the PROMISe project was the new “Alt+I”

intervention screen. This is a multi-tabbed screen allowing the user to enter considerably more

information about interventions than was previously possible in WiniFRED.

A number of additional database tables were created in the local WiniFRED data store (see Figure

4.1-3) at each pharmacy to hold the additional information being recorded about interventions.

To communicate with Comm Server, WiniFRED used a modified version of the messaging

components developed for a previous project where secure transmission was required. The existing

“WFMessenger” component was adapted for PROMISe to enable the WiniFRED information return /

error messages that may have been returned from the University server via Comm Server to be

viewed.

The intervention recording function in WiniFRED could be accessed by two methods:

• Through the 'Activities' menu by selecting intervention or

• by pressing Alt+I.


Page 31 of 361

4.1.3.1 Preliminary Steps

The basic commencement screen is shown in Figure 4.1-5. This consists of some demographic and

patient specific information above a series of tabs regarding the intervention. Where the intervention is

associated with a prescription, information concerning the prescription is imported automatically from

the dispensing system (see Figure 4.1-6). Information that is imported into the intervention module

automatically consists of:

• the patient's name

• the prescriber's name and prescriber number

• the drug involved

• the prescription number

• the number of different medications that the patient has had in the previous six months

(medication count)

Other information needs to be manually entered for each intervention. The recording pharmacist must

enter:

• the gender of the patient (unless previously stored in the dispensing software)

• the age group of the patient

• who initiated the problem (proactiveness)

• pharmacist identifier (initials)

When the clinical activity was not related to a prescription, the information had to be entered manually.

For situations where the patient is not listed in the pharmacy database, an unlisted patient could be

created. This allowed for the documentation of “off the street” interventions relating to customers who

had not received prescriptions from that particular pharmacy. The adopted “tab” structure allows this

demographic information to be in view while different aspects of the intervention documentation are

completed. The titles on these tabs change from red to green when information has been entered.


Page 32 of 361

Figure 4.1-5: Introduction Screen for the WiniFRED Intervention Interface

Figure 4.1-6: Populated Introduction Screen For the WiniFRED Intervention Interface


Page 33 of 361

4.1.3.2 Recording Category and Subcategory

The first step in the recording of an intervention in this system is selection of the main category of the

clinical activity. By selecting the main category on the left-hand side the appropriate subcategories

appear on the right-hand side of the entry screen. Help screens are available for each of the different

selections.

Once the initial category and subcategory has been selected, a draft of the intervention can be saved

and completed at a later time. A list of all interventions that are in draft mode can be accessed by

selecting activities; draft interventions (see Figure 4.1-7).

Figure 4.1-7: Saving and Re-Accessing a Draft Intervention

At any time during the recording process, help can be accessed using the button at the bottom of the

screen. Contextual help corresponding to the item selected by the recording pharmacist is

automatically provided (see Figure 4.1-8).


Page 34 of 361

Figure 4.1-8: Help Screen For the WiniFRED Intervention Interface

4.1.3.3 Recording Action(s)

Once the category and subcategory have been selected, the actions taken to investigate the problem

are entered. In the action entry screen, the actions taken in investigating the drug-related problem can

be selected (see Figure 4.1-9). Selection occurs by using the arrows between the two selection boxes.

Items that will be recorded against the intervention are on the right-hand side of the selection screen.

As multiple actions are possible for each clinical activity, the process allows for multiple selections to

be made. The system also allows for multiple recording of the same action. So, for example, if the

pharmacist speaks to a patient, then contacts the prescriber, then returns to speak to the patient

further, he may record the action “discussion with patient or carer” on two separate occasions.


Page 35 of 361

Figure 4.1-9: Recording of Actions on the WiniFRED Intervention Interface

4.1.3.4 Recording Recommendation(s) and Outcome

Recommendations made to resolve the drug-related problem are entered into the system in a similar

manner to the actions using a left-to-right selection arrow in the centre of the screen (see Figure

4.1-10).

Although multiple recommendations can be recorded against each intervention, it is not possible to

make the same recommendation more than once. The system has been designed to allow for multiple

unique recommendation selections.

Where the recommendation relates to a direct change in therapy, information concerning the before

and after situation was collected to allow for clarification of what changes were recommended. To

collect this information, an 'edit' key was created. So, for example, if a change in drug was

recommended, the drug could be entered with an explanation of the change made (see Figure 4.1-10

and Figure 4.1-11).

The outcome of the recommendation is included on this same screen and relates to a composite

outcome for all of the recommendations made.


Page 36 of 361

Figure 4.1-10: Recording Recommendations In the WiniFRED Intervention Interface

Figure 4.1-11: Recording Recommendations In the WiniFRED Intervention Interface: Details Of Drug Change


Page 37 of 361

4.1.3.5 Recording Clinical Significance and Time Taken

The clinical significance screen, (see Figure 4.1-12) allows a selection to be made to indicate the

appropriate level of clinical significance for the activity. For interventions that were rated as moderate

or high, additional information concerning the patient’s other medical conditions was requested. This

additional information was collected in order to enable reconstruction of the clinical situation which in

turn would enable adequate examination of the activity by an external clinical panel. A reminder of the

importance of this additional information was automatically presented when either moderate or high

clinical significance was selected (see red reminder message in Figure 4.1-13).

The time taken to perform the intervention was also recorded on this screen. Individual buttons for

particular brackets of time were made available to enable quicker data entry (see Figure 4.1-13).

Figure 4.1-12: Recording Of Significance In the WiniFRED Interventions Interface (Mild Significance)


Page 38 of 361

Figure 4.1-13: Recording Of Significance In the WiniFRED Interventions Interface (Moderate Or High Significance)

4.1.3.6 Local Intervention Record: Patient history and Summary

Once an intervention has been recorded, an entry concerning the intervention appears in the patient

history file (see Figure 4.1-14). The full details of the intervention can be accessed by double-clicking

on this entry. In addition, a summary of the intervention can be generated (see Figure 4.1-15).


Page 39 of 361

Figure 4.1-14: Intervention Note In Patient History

Figure 4.1-15: Printable Individual Intervention Summary


Page 40 of 361

4.1.3.7 Local Intervention Record: Reports

A number of reports can be produced for the recording pharmacist (see Figure 4.1-16). These include

a detailed report, a summary report and a statistical report of interventions. The detailed report is

shown in Figure 4.1-17. These reports were filterable, and reports could be generated for all

interventions or criteria could be specified (for example a date range or intervention category could be

specified).

Figure 4.1-16: Reports For Interventions Available at Each Pharmacy

Figure 4.1-17: Pharmacy-Based Detailed Report Of Interventions


Page 41 of 361

4.2 PROMISe Intervention Study Design

Consideration and review of the methods used in the PROMISe pilot study lead to the development of

the methods for data collection in the larger sample of pharmacies proposed for the PROMISe

intervention study. This information was used to determine an appropriate sample size and other

aspects of the main study.

4.2.1 Sample Size

As many parameters were measured and analysed during the trial it was challenging to determine a

sample size for this study. An important consideration was to ensure the sample gathered adequate

examples of pharmacist interventions that are associated with health or economic benefit.

In the PROMISe pilot study, 352 clinical interventions were recorded in 7 pharmacies during the

dispensing of around 9000 prescriptions. Of these interventions, 13 (3.7%) were classified as of high

significance, likely to prevent hospitalisation, by the recording pharmacist. A sample of interventions

(including these “high” interventions) was subsequently reviewed by a six-member clinical panel and

the probability of a hospitalisation from these interventions was rated as 9.9%.3 Thus, in order to

accumulate sufficient examples of high-level interventions that are confirmed as high by a clinical

panel (20), approximately 200 high level interventions would need to be collected in the PROMISe

intervention study.

Pilot Study

(PROMISe I)

PROMISe

Intervention

Study*Pharmacies 7 60

Pharmacists 21 150

Duration 2 weeks 4 weeks

Prescriptions 9012 139500

Rate of Interventions 3.90% 3.90%

Total Number of Interventions 352 5441

Rate of High interventions 3.70% 3.70%

Number of High Interventions 13 201* estimate

Table 4.2-1: Initial Estimates for Recruitment for PROMISe INtervention Study

During the course of the data collection it became apparent that the intervention rate was lower than

that in the pilot study. The data collection period was extended in an attempt to reach the predicted

number of interventions (see Phase 3 below).

3 Community Pharmacy Medication Incident Reporting and Management Systems (CPMIRMS) also

known as PROMISe: Pharmacy Recording of Medication Incidents and Services electronic

documentation system: Final Report 2004


Page 42 of 361

4.2.2 Overall Design of PROMISe Intervention Study

Given the overall objectives of the study, an observational controlled design was used. Each element

of the design was intended to enable differentiation of the effects of remuneration, the presence of an

observer and the presence of an intervention prompt.

The initial intended design included a crossover of remuneration over two, two week periods, with

randomised allocation of the intervention prompt and the presence of observer pharmacists between

these arms (see Figure 4.2-1).

Pharmacies Enrolled

Paid in

Phase 1

Unpaid in

Phase 1

Observation

Pharmacist

No

observation Pharmacist

No observation Pharmacist

Observation Pharmacist

No Intervention Prompt

Intervention

Prompt

No

Intervention Prompt

Intervention Prompt


Intervention Prompt

No

Intervention Prompt

Intervention

Prompt

Phase 1(2 weeks)

Phase 2(2 weeks)

Unpaid in Phase 2

Paid in Phase 2 No Intervention

Prompt

Intervention Prompt


Intervention Prompt



Cro

sso

ver

Figure 4.2-1: Intended Design for PROMISe Intervention Study However, as there were some initial installation difficulties with the communications software, some

pharmacies were delayed in commencing their data collection. The intervention prompt that was

installed also caused a small overlap problem. When the intervention prompt was de-commisioned,

this occurred with the routine monthly software update. The update was available to pharmacies from

the 24th of the month, and could be installed by individual pharmacies at any time before the beginning

of the next month. Thus some pharmacies “turned off” their intervention prompts up to a week earlier

than others (see section 5.9). In addition, the number of interventions documented was lower than

expected and extension of the study and the observation process was required in order to collect

information concerning a sufficient number of interventions.

The final schema, randomisation outcome and duration of observation and intervention prompt for the

participating pharmacies are shown in Figure 4.2-2.


Page 43 of 361

52 Pharmacies

Enrolled

23 Paid in Phase 1



12 No observation

Pharmacist

18 No observation

Pharmacist


9 No Intervention

Prompt

9 Intervention

Prompt

6 No Intervention

Prompt

5 Intervention

Prompt

2 No Intervention

Prompt

9 Intervention

Prompt

5 No Intervention

Prompt

7 Intervention

Prompt

52 Pharmacies

Paid in Phase 3

Phase 1(2 weeks)

Phase 2(2 weeks)

Phase 3(4 weeks)


29 Paid in Phase 2


12 No observation

Pharmacist

18 No observation

Pharmacist


9 No Intervention

Prompt

9 Intervention

Prompt

6 No Intervention

Prompt

5 Intervention

Prompt

2 No Intervention

Prompt

9 Intervention

Prompt

5 No Intervention

Prompt

7 Intervention

Prompt



Figure 4.2-2: Randomisation Schema for PROMISe Intervention Study Despite the difficulties and overlaps, there was still sufficient information obtained to indicate the

effects of observation, remuneration and the intervention prompt. The effects of these could be

assessed independently and together using univariate and multivariate analyses. The combined effect

of all three techniques (adjusted for non-recording of interventions) was used to determine and

optimum level of interventions.

Figure 4.2-3 represents the steps involved in conducting the project in the Melbourne pharmacies.

Each of these steps is explained in the following sections.


Page 44 of 361

Recruitment and Enrolment of Pharmacies and

Pharmacists

On - Line Training Exercise for All Pharmacists

Commencement and Enrolment Meeting for All Pharmacists

Randomisation to Payment First or Second Half of Period One

Allocation and Scheduling of Seven Observers

Randomisation to Educational Drug Alert

Data Collection Begins

Project team visit and

provide phone support

Pharmacists Record Clinical Services Electronically

Figure 4.2-3: Overview Of Methodology For Melbourne PROMISe Trial

4.3 Selection and Randomisation of Pharmacies

4.3.1 Project Promotion

Pharmacies registered as users of the WiniFRED dispensing software in metropolitan Melbourne were

contacted and provided with information on the PROMISe project. Expressions of interest were sought

from the 260 registered users, with the opportunity to reply via email or fax. Information concerning

the project was also generally distributed through AusPharmList and the Australian Pharmacist.

A flyer was developed to increase awareness of the project and distributed in the Australian

Pharmacist, the Victorian Guild Newsletter, the AACP Newsletter and the SHPA Newsletter (see

Figure 4.3-1 and Figure 4.3-2).


Page 45 of 361

Figure 4.3-1: PROMISe Project Promotional Brochure (Part 1)

Figure 4.3-2: PROMISe Project Promotional Brochure (Part 2)


Page 46 of 361

Once the pharmacies had expressed interest in the project, preliminary information was obtained,

relating to:

• computer hardware,

• software configuration and

• the number of staff likely to be involved.

This information provided detail for the logistical planning of the project and ensured the sample

provided a good representation of community pharmacies. Appendix 8 contains information

distributed to the potential pharmacies. Only pharmacies who had installed the Windows operating

systems Windows 2000 or Windows XP were accepted for enrolment.

4.3.2 Enrolment

Once each pharmacy had received information on the project, had agreed to participate and met the

software requirements, they were enrolled into the project. This involved signing an agreement with

the University to participate in the project. The agreement covered aspects of non-disclosure, privacy,

ethics and use of the data. A copy of the agreement can be found in Appendix 9.

4.3.3 PROMISe Pharmacy and Pharmacist Questionnaires

Questionnaires concerning pharmacy and pharmacist demographics were designed to be able to be

linked with information from previously published surveys of Australian community pharmacy. These

included The National Pharmacy Database Project4 and the Evaluation of the Quality Care Pharmacy

Program5 2005.

From these studies and the information gathered during the PROMISe pilot study, four questionnaires

were developed which were intended to gain information concerning particular aspects of the

pharmacy or pharmacist.

The four questionnaire information sources for the project were:

• Pharmacy ethos and logistics (The Pharmacy Owner/Manager questionnaire, section 4.3.3.1)

• Pharmacist demographics (completed online before the study, see section 4.3.3.2)

• Staff workload questionnaire (see section 4.3.3.3), and

• Pharmacist opinions of the study (The post-trial questionnaire, see section 4.5.2.1)

4 Berbatis CG, Sunderland VB, Mills CR, Bulsara M. National pharmacy database project. 2002.

5 Chapman J. Evaluation of the Quality Care Pharmacy Program/ Pharmacy Guild Project Number

92001-01. Available at http://beta.guild.org.au/research/project_display.asp?id=281


Page 47 of 361

4.3.3.1 Pharmacy Owner or Manager Questionnaire

The owner/manager questionnaire included questions designed to gain information concerning various

demographic characteristics of the pharmacy. These included:

• the size and location of the pharmacy,

• the resources available (staff, IT and financial),

• ownership,

• workload and staff mix.

In addition, a set of 14 opinion based statements were developed from Doucette's entrepreneurial

orientation assessments.6 These statements addressed the parameters which are involved in

determining a pharmacy’s entrepreneurial orientation (EO). These are innovativeness, risk-taking,

work ethic, proactiveness, autonomy and competitiveness. The Pharmacy Owner or Manager

Questionnaire is included in Appendix 10.

4.3.3.1.1 Calculation of Entrepreneurial Orientation (EO)

Doucette has proposed a link between EO and entrepreneurship, which in turn can improve

competitive positioning. When considering EO, the environment in which the business operates also

provides influence over the success of the business (see Figure 4.3-4).

Entrepreneurial Orientation

Performance

Market shareProfitability

Financial strength

Environment

Figure 4.3-3: The Impact Of Entrepreneurial Orientation On Business Performance

6 The Influence of Environmental Attributes on the Relationship between Entrepreneurial Orientation

and Performance in Independent Community Pharmacies. WR Doucette, MS Shrividya Iyer. University of Iowa


Page 48 of 361

In order to determine EO it is necessary to consider if the company is innovative and prepared to

make changes and, where necessary, act in a competitive nature. The proactiveness of their changes

and the work structure of the business also play a part in EO.

The work by Doucette et al.7 has lead to the development of a scale to assess EO. The areas

considered in the scale are listed below and shown in Figure 4.3-4;

• Proactiveness (P)

• Autonomy (A)

• Innovativeness (I)

• Competitive Aggressiveness (C)

• Work Ethic (W)

• Risk taking (R)


Proactiveness

Work EthicCompetitiveness

aggressive

Innovativeness

Risk Taking

Autonomy

Figure 4.3-4: Entrepreneurial Orientation Factors

Doucette developed the following formula from his work:

EO= (P x 0.98) + (A x 0.89) + (I x 0.82) + (C x 0.66) + (W x 0.64) + (R x 0.55)

Doucette explored Entrepreneurial Orientation in American community pharmacies and found that

whilst EO does not always lead to a better performance, given an appropriate environment it can

influence performance. In particular, where the environment is competitive and aggressive having a

stronger entrepreneurial orientation will lead to better business performance. Also, where the

pharmacy had greater resources, whether they be financial or intellectual, the pharmacies tended to

perform better.

7 Doucette WR, Jambulingam T. Pharmacy Entrepreneurial Orientation: Antecedents and its effect on

the provision of innovative pharmacy services. J Soc Admin Pharm. 1999;16:26-37.


Page 49 of 361

Results of the 14 opinion-based statements were used to develop scores for each of the parameters

involved in determining the entrepreneurial orientation. Questions were asked that were indicative of

the parameters:

• Proactiveness (P) – Two statements addressing this factor

• Autonomy (A) - Two statements addressing this factor

• Innovativeness (I) - Two statements addressing this factor

• Competitive Aggressiveness (C) - Two statements addressing this factor

• Work Ethic (W) - One statement addressing this factor

• Risk taking (R) - Two statements addressing this factor In addition to these statements, three statements designed to assess the management structure of the

pharmacy were included. The statements used to evaluate these parameters are shown in Table

4.3-1.

Innovation

This pharmacy is known for innovation among the pharmacies in this area

This pharmacy and management encourages the development of innovative services

Proactive

Because conditions are changing we continually seek out new opportunities

The management team is closely able to predict the future needs of this business

Risk Taking

It is out business strategy to avoid taking too many chances

If there were a risky new project or service this pharmacy would be prepared to take it on

Autonomy

Ideas for new services in the pharmacy from staff are supported by the management team

At this pharmacy it is primarily the management team who identify new business opportunities

Work Ethic

At our pharmacy we are ambitious about the service we provide

Competitive Aggressiveness

Our actions towards competitors can be termed aggressively competitive

We are aware and responsive to changes that other pharmacies in our area make

Management Structure of the Pharmacy

Five-year plans for this pharmacy are a high priority

The majority of staff in this pharmacy have worked here for more than 5 years

There has been little change in our pharmacy over the last 10 years

Table 4.3-1: Statements Used To Assess Entrepreneurial Orientation


Page 50 of 361

Each of these statements were scored using 0-10 Likert scales. The responses for related questions

were grouped, and where appropriate reversed, to allow calculation of the separate parameter scores.

A separate “unsure” box was included with the scale.

The formula shown above for calculation of entrepreneurial orientation was used to amalgamate these

parameters into a single entrepreneurial orientation score for each pharmacy.

4.3.3.2 Pharmacist demographics

The online pharmacist demographic questionnaire is shown in full in Appendix 11, and addressed

three main areas:

• Section 1: the pharmacist’s background in terms of training and experience (Figure 4.3-5 and

Figure 4.3-6)

• Section 2: the pharmacist’s current practice profile (Figure 4.3-7)

• Section 3: the pharmacist’s opinions regarding the profession of pharmacy (Figure 4.3-8)

The opinion-based statements requested in section 3 of the questionnaire targeted aspects of the

pharmacist’s professional integrity, job satisfaction and readiness for change.

Scores for each of these three characteristics were determined from the responses to questions in

these areas. Thus, each pharmacist completing the survey could be assigned an indicative score for

professional integrity, job satisfaction and readiness for change, in addition to providing the

demographic information.

Figure 4.3-5: Online Pharmacist Demographics Questionnaire (Part 1)


Page 51 of 361

Figure 4.3-6: Online Pharmacist Demographics Questionnaire (Part 1 Continued)


Page 52 of 361

Figure 4.3-7: Online Pharmacist Demographics Questionnaire (Part 2: Current Role And Practice Profile)


Page 53 of 361

Figure 4.3-8: Online Pharmacist Demographics Questionnaire (Part 3: Personal Views Of Pharmacy)

4.3.3.3 Staff workload

In order to accurately assess workload and in order to establish a relationship between number of

prescriptions and number of staff members present (i.e. how busy a pharmacy was) on a particular

day, we collected detailed information on staffing levels at particular hours of particular days of the

week. This allowed us to calculate a workload factor, which was used in further analysis (see section

4.6 ).

4.3.4 Study Arm Allocation

Pharmacies were randomised, using a computer-generated list of random numbers, for early or late

remuneration and for the educational drug alert. The pharmacies were also allocated to receive either

regular half day visits by an observer (the observation arm) or a less intensive short follow-up visits by

members of the Project Team.


Page 54 of 361

4.3.4.1 Remuneration Arms

To establish any influence of remuneration on intervention documentation frequency, pharmacies were

randomised to receive payment for clinical interventions from the 20th April to the 6th of May, or from

the 6th of May to the 20th of May. For each clinical intervention submitted during a payment period the

pharmacy received $15. During the extension of the trial from the 23rd of May to the 17th June all

pharmacies received payment of $15 for clinical interventions. Hence each pharmacy experienced

both payment and no payment for their interventions (see Figure 4.3-9).

PaymentPayment

PaymentPayment

Phase 1 Phase 2 Phase 3

PaymentPayment

Group A

Group B

Payment, $15 for each clinical intervention

No paymentNo payment


Figure 4.3-9: Outline Of Remuneration Randomisation For Intervention Payments

4.3.4.2 Aspirin Intervention Prompt Arm

Although it was not possible to assess the impact of a complete clinical intervention educational

program in this study, we were able to institute a specific intervention prompt for a selected

intervention.

The prompt related to the use of antiplatelet agents in high-risk patients with diabetes, and was

activated when a prescription for an oral antidiabetic agent was selected for dispensing (see Figure

4.3-10)


Page 55 of 361

Figure 4.3-10: Automated Intervention Alert

Those pharmacies that were randomised to receive the aspirin intervention alert were provided with an

“Aspirin pack” (see Appendix 12). This contained information suitable for the patient (see Figure

4.3-11), and also some supporting information for the pharmacist (see Figure 4.3-12).


Page 56 of 361

Figure 4.3-11: Patient Handout For Aspirin Automatic Intervention Prompt


Page 57 of 361

Figure 4.3-12: Pharmacist Handout For Aspirin Automatic Intervention Prompt


Page 58 of 361

Pharmacies were randomly allocated (on a 3:2 ratio basis) to receive the aspirin intervention prompt

outlined in section 4.3.4.2. The prompt was developed to test the hypothesis that a computer-based

reminder could increase intervention rates relating to a particular type of intervention. The educational

alert used in the Melbourne PROMISe project related to the use of low-dose aspirin (or other

antiplatelet agent) for cardiovascular and cerebral vascular event prevention in high risk diabetic

patients. The alert was triggered when an oral antidiabetic agent was selected for dispensing (Section

4.2.3).

Thirty one of the 52 pharmacies had the drug alert installed, which provided information for the patient

and for the pharmacists on aspirin usage for risk reduction.

4.3.4.3 Observation Allocation

Observer pharmacists were placed in some of the pharmacies for the initial three weeks of the study. It

was the role of the observers to assist in the documentation process and provide a reminder for

documentation. They also created a record of documentable events and hence were able to

demonstrate what was recordable and assist pharmacists to record the event. The observers were

also able to accurately time events involved in the intervention.

Support was provided to these observers by the Project Team and regular feedback meetings were

arranged. Full training was provided for all observers in the classification system and also in the

WiniFRED dispensing Software. All observers were experienced pharmacists who had used the

WiniFRED dispensing system. All observers completed the online training scenarios and a detailed

discussion took place during the observer training day in order to clarify the most appropriate

classification of the scenarios. A manual of information for each observer was prepared (see Appendix

13). This contained details of the visiting schedule, contact details for the assigned pharmacies and

the project team and recording sheets for activities within each pharmacy.

All observers carried mobile telephones and were contacted daily by the project team to determine if

any issues had arisen. The project team also visited some pharmacies while observation was in

progress. Any queries from the observers were either handled on an individual basis, or referred to the

project team if the issue was occurring in more than one pharmacy.

These observers visited 21 of the pharmacies on one weekday in three from the 21st of April to the 11

th

of May. Their visits were scheduled to ensure appropriate contact time and travel time, with each visit

lasting for 3 hours. The schedule can be seen in Figure 4.3-13 , and were designed to provide

additional information, assist with documentation or software difficulties, and carry out timing of

specific events.


Page 59 of 361

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Sunday, 17 April 2005

Monday, 18 April 2005

Tuesday, 19 April 2005

Wednesday, 20 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Thursday, 21 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Friday, 22 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Saturday, 23 April 2005

Sunday, 24 April 2005

Monday, 25 April 2005

Tuesday, 26 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Wednesday, 27 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Thursday, 28 April 2005 1 2 3 4 5 6 7

Friday, 29 April 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Saturday, 30 April 2005

Sunday, 1 May 2005

Monday, 2 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Tuesday, 3 May 2005 1 2 3 4 5 6 7

Wednesday, 4 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Thursday, 5 May 2005 1 2 3 4 5 6 7

Friday, 6 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Saturday, 7 May 2005

Sunday, 8 May 2005

Monday, 9 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Tuesday, 10 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Wednesday, 11 May 2005 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Observers 1-7 am pm

Observed Pharmacies

7-9pm Training Session for Pharmacies

9am to 4pm Observer Training Sessions and 6:30-8:30pm Training Session for

Pharmacies

7-9pm Training Session for Pharmacies

Meet with PROMISe Team

Meet with PROMISe Team

Figure 4.3-13: Observation Schedule For PROMISe Visits (Numbers Represent Different Observers)

To enable a practical observation visit schedule to be developed, the pharmacies which were to have

an observer present were selected based on location. The pharmacies which had completed the

enrolment process (n = 52) were plotted onto a map of greater metropolitan Melbourne. From this

map, 21 pharmacies were selected to receive regular observer visits. Using this technique pharmacies

were selected across a range of locations (see Figure 4.3-14).


Page 60 of 361

Figure 4.3-14: Locations Of Pharmacies Selected For Regular Observer Visits

The observation visits consisted of half-day sessions for support and facilitation of documentation of

interventions. Each observer usually visited one pharmacy in the morning, and a different (usually

nearby) pharmacy in the afternoon. The schedule for the visits is shown in Figure 4.3-13.

4.4 Induction and preliminary training

Training for the PROMISe Intervention Study involved education in both the DOCUMENT classification

system and in the WiniFRED PROMISe user interface. The training for the DOCUMENT classification

system was undertaken online, and demographic information concerning the participant’s background,

opinions of pharmacy and clinical skills was also collected during this process (see Section 4.3.3.2).

4.4.1 On-Line Training for the PROMISe project

Training in the DOCUMENT system was completed online. It was designed to familiarise the

pharmacists with the selection of appropriate categories during the process of documenting

interventions. The training module was available on the PROMISe website (www.promise.id.au). This

website was also used for ongoing feedback and information for PROMISe trial pharmacists.


Page 61 of 361

Figure 4.4-1: Online Classification Of Interventions Using DOCUMENT (The PROMISe Website)

The online training included three activities of which the first two were clinical problem solving skill

tests. The third activity consisted of 20 case-based scenarios in which the problem and the steps

taken to resolve the problem were described.

For each of the 20 cases the pharmacist assigned a DOCUMENT category and recorded the

appropriate actions and recommendations. The pharmacist also assigned the significance of the

intervention and considered who initially identified the problem.

Feedback was provided immediately after assessment of each scenario, with the suggested answers

and a discussion about each aspect. The complete scenario and the preferred classification codes

and the feedback given are shown in detail in Appendix 14.

The process of registration is depicted in Figure 4.4-2. Once registered, the participant could complete

the training. This registration also allowed the participant to complete the training in different stages if

they needed to.


Page 62 of 361

Figure 4.4-2: PROMISe DOCUMENT Training Commencement Screens


Page 63 of 361

Once registered within the training package there were three activities to complete: a tick appears

when each task is complete (see Figure 4.4-3).

Figure 4.4-3: PROMISe DOCUMENT Training And Clinical Entry Screen

4.4.1.1 Clinical Problem Solving Skills

As well as the web-based DOCUMENT training, pharmacists were asked to complete two clinical

problem solving scenarios. These interactive scenarios allowed the identification and assessment of a

number of clinical problems within a patient scenario.

Each step involved multiple choice questions or lines of investigation, which then provided additional

information depending on the path chosen (see Figure 4.4-4). Individual participants in the clinical

problem solving skills case studies received an overall score, which was based on:

• the number of problems that they identified,

• the appropriateness of the recommendations they made, and

• the number of investigative steps that they used to detect the problems.

It was possible as a result of the design of the system to track the number of pages that the

pharmacist visited in order to identify the underlying drug-related problems.

Thus, the total score was collated from an investigation score, a problem identification score, and an

appropriateness of recommendation score. It was felt that pharmacists with different levels of clinical

skills would have different ratios of scores within these three components.


Page 64 of 361

Figure 4.4-4: On-Line Clinical Problem Solving Skills Assessment

A complete summary of the choices and details of the clinical skills assessment scenario is shown in

Appendix 15.

4.4.1.2 DOCUMENT Training Scenarios

The 20 scenarios were carefully designed to encompass a variety of the problem types, actions and

recommendations available in the DOCUMENT system. Some particularly difficult to categorise

scenarios were included with explanations for why particular selections should be made. This allowed

the participant to have greater exposure to the different selection options (and their explanations)

available.

Once the user has selected training for the DOCUMENT classification system, a welcome screen for

their first assessment appears. There is an option on this screen to view an online demonstration of

how to categorise an intervention using the DOCUMENT system (see Figure 4.4-5 and Figure 4.4-6).


Page 65 of 361

Figure 4.4-5: PROMISe DOCUMENT Training Welcome Screen


Page 66 of 361

Figure 4.4-6: Online DOCUMENT Training: Example Of How To Classify A Scenario

Once the user selects a particular scenario, they are presented with a scenario summary and

instructions for completing the first step of the categorisation process (see Figure 4.4-7). The

scenario summary remains active whilst assigning categories, actions and recommendations.

For PROMISe trial participants, an audio visual presentation was also made available to facilitate the

training process; this was accessible through the PROMISe website.


Page 67 of 361

Figure 4.4-7: PROMISe DOCUMENT Training: Assigning Category And Subcategory To The Intervention

The previously mentioned scope notes and detailed definitions of each subcategory were available by

selecting the help section. Once selection of the category and subcategory was completed actions

could be selected.

Pharmacists were asked to document the actions that were undertaken by the pharmacist in the

scenario presented (see Figure 4.4-8). However, as the results of the training came to light, it became

clear that many pharmacists recorded the actions that they considered appropriate i.e. ‘what they

would have done’, rather than those demonstrated in the scenario.


Page 68 of 361

Figure 4.4-8: PROMISe DOCUMENT Training: Assigning Action For The Scenario

Selecting appropriate recommendation(s) for the scenario was the next step in the classification

process. As with action selection, where multiple recommendations were made by the pharmacist in

the scenario, these could be recorded (see Figure 4.4-9 ).


Page 69 of 361

Figure 4.4-9: PROMISe DOCUMENT Training: Assigning Recommendation(s) For The Scenario

The significance of the situation was then assessed as nil, low, mild, moderate or high, with the

relevant definitions available (see Figure 4.4-10 ).


Page 70 of 361

Figure 4.4-10: PROMISe DOCUMENT Training: Assigning Significance For The Scenario

The final step in the categorisation process was to ask the pharmacist to nominate, who in the

scenario originally identified the drug-related problem. This was done in order to train PROMISe

participants in the assignment of proactive and reactive interventions (see Figure 4.4-11 ).


Page 71 of 361

Figure 4.4-11: PROMISe DOCUMENT Training: Assigning Proactive Or Reactive Nature Of Scenario

The comments and feedback section (see Figure 4.4-12 ) was an enhancement made to the online

training for PROMISe II. The correct categorisation selections were displayed and additional

comments were formulated based on the definitions for the categories and comments made from

PROMISe I training.

One important difference between the training in the pilot study and the training in this program, was

that the feedback was provided to the participant immediately after categorisation of each scenario.

The intent of this was to facilitate an incremental improvement of classification ability. An example of

the feedback screen is shown in Figure 4.4-12.


Page 72 of 361

Figure 4.4-12: PROMISe DOCUMENT Training: Feedback Provided After Each Scenario

Being web-based, it was possible to automatically record the time taken for completion of each case

(and for the entire training process). It should be noted that there was no compulsion on pharmacists

undertaking the training that they complete all training scenarios in one sitting.


Page 73 of 361

4.4.2 WiniFRED Interface Training

A familiarisation evening was conducted to provide background information on the project and brief

training on the WiniFRED PROMISe interface. These sessions were conducted over three

consecutive nights to enable each pharmacist to have an ample opportunity to attend a session. Each

pharmacy was provided with written reference material (The PROMISe manual) for the project.

Information on the schedule for visits and other aspects of randomisation was also provided.

For those pharmacists who could not attend the training sessions, a CD-ROM presentation was either

mailed out or personally delivered. In addition, an audiovisual presentation was made available on the

PROMISe website.

Included in the PROMISe manual provided was a questionnaire on the ethos and logistics of the

pharmacy (see pharmacy demographics in section 4.3.3.1 ). This questionnaire was also distributed to

all WiniFRED pharmacies in metropolitan Melbourne who were not participating in the project in order

to gain insight into the representativeness of our sample of pharmacies.

4.5 PROMISe Data Sources and Data Processing

The nature of information collected for the PROMISe project from different sources is outlined in

Figure 4.5-1.


Page 74 of 361

4. PROMISe SQL Database (~13,000 interventions and ~430,000 prescriptions)

•Age group and gender of Intervention patient•Other prescriptions dispensed for intervention patient•Intervention Category, Action, Recommendation, Outcome•Intervention Significance and Proactive Assessment•Pharmacists Time taken for Intervention•Drug(s) Involved•Scripts Dispensed (Number, Drug, Prescriber, Directions)•Patient Encounters

3. PROMISe Pharmacist Demographics (125)

•Gender, Age, Registration year•Practice Profile•Job Satisfaction•Self-Reported Workload•Professional Integrity•Change Readiness•Clinical Skills

5. Clinical Panel Assessments(20 members, ~250 interventions)

•Risk Reduction of consequences of Interventions•Attributability to pharmacist•Consequences of interventions

7. PROMISe Pharmacists Feedback (~80)

•Opinions of Intervention Documentation•Barriers and Facilitators to Interventions•Software Feedback

•Remuneration Opinions

•Focus Group Results•Depth Interview Results

1. PROMISe Pharmacy Demographics (52)

•Entrepreneurial Orientation•IT Facilities and resources•Location and Size•Date of QCPP Accreditation

•Staffing levels

2. Non- PROMISe WiniFRED Pharmacy Demographics

(~40)•Entrepreneurial Orientation•IT Facilities and resources•Location and Size•Date of QCPP Accreditation

8. Non-PROMISe Pharmacists Opinions (~400 phone interviews)

•Opinions of Intervention Documentation•Barriers and Facilitators to Intervention identification•Barriers and Facilitators to Intervention Documentation

6. Direct Observation Visits (21 visits)

•Time for Intervention Entry•Identification of Barriers and Facilitators

Figure 4.5-1: Grouped Data Sources For PROMISe Intervention Study

4.5.1 Accumulation of Recorded Interventions

After completion of training, each participating pharmacy had the Comm Server installed and the

changes to their dispensing software made. The installation process involved a visit to each pharmacy

in order to load and test the software. A number of different system configurations were encountered,

and a number of pharmacies did not have the operating system that they indicated on their enrolment


Page 75 of 361

form. Many pharmacists were not aware of how their systems were set up and, on some occasions, a

second visit or a consultation with the WiniFRED support team was required. Once the Comm Server

was successfully installed, most updates could be conducted automatically over the internet.

The pharmacists were then able to record interventions using the PROMISe software and continued to

record during phase one, two and three of the study. The observer visit schedule, activation of drug

alert, and crossover remuneration were only present for phase one and two of the project. During

phase three, the pharmacies were provided with ongoing feedback and were contacted by the Project

Team. The PROMISe website was also updated at all stages of the project (see Figure 4.3-9).

Pa

ymen

t (23

)

Pa

ymen

t (23

)

Paym

en

t (52

)

Obse

rvatio

n S

upp

ort V

isits (2

1)

Auto

mate

d In

terv

entio

n P

rom

pt (3

1)

Sup

port V

isits

to N

on-O

bserv

ed

P

harm

acie

s (3

1)

No P

aym

en

t (29

)

No P

aym

en

t (29

)

Auto

mate

d B

rand

Substitu

tion

Re

cord

ing (5

2)

Te

leph

one

Support to

All P

ha

rmacie

s

Phase One21st April to 6th May

Phase Two7th May to 22nd May

Phase Three23rd May to 17th June

Pa

ymen

t (23

)

Pa

ymen

t (23

)

Paym

en

t (52

)

Obse

rvatio

n S

upp

ort V

isits (2

1)

Auto

mate

d In

terv

entio

n P

rom

pt (3

1)

Sup

port V

isits

to N

on-O

bserv

ed

P

harm

acie

s (3

1)

No P

aym

en

t (29

)

No P

aym

en

t (29

)

Auto

mate

d B

rand

Substitu

tion

Re

cord

ing (5

2)

Te

leph

one

Support to

All P

ha

rmacie

s

Phase One21st April to 6th May

Phase Two7th May to 22nd May

Phase Three23rd May to 17th June

(Numbers in brackets indicate number of pharmacies)

Figure 4.5-2: Outline Of Different Phases Of PROMISe Intervention Study Data Collection

4.5.1.1 Feedback to Pharmacists and Pharmacies During the Trial

4.5.1.1.1 PROMISe website – trial participants

The PROMISe website incorporated the online training for the project and also provided a reference

point for pharmacists participating in the trial. The section for trial participants was continually updated

and included links to presentations created by the Project Team for the participants.


Page 76 of 361

Figure 4.5-3: Trial Participant Web Feedback And News Page

4.5.1.1.2 Individual Feedback Reports for Pharmacists and Pharmacies

A web-based reporting system was developed to allow individual participants to view their clinical

intervention rate. This system used information as it was submitted to the PROMISe Server. The

information was presented as shown in Figure 4.5-4.


Page 77 of 361

Pharmacy Name

Pharmacist Name

Pharmacy Name

Pharmacist Name

Figure 4.5-4: Individual Pharmacy And Pharmacists Intervention Rate Report

The information on this feedback report was calculated from information available in the database at

the time of viewing. There were some initial difficulties with the amount of information sent to the

server from each pharmacy, as the hardware in place at the University was insufficient to cope with

the message load. Although changes were made to the hardware, messages from several days

previously remained “queued up” for a number of pharmacies. To compound this problem, 10 days

into the trial, it was discovered that there was a misreading of a data field (dispensed date) in the

dispensing history information being sent to the Server. As a result of the misreading, the incorrect

date of dispensing of prescriptions was submitted to the PROMISe database. This meant that the

number of prescriptions on each day for each pharmacists and pharmacy would therefore be

unreliable. The number of prescriptions was critical in order to calculate a rate of interventions (the

number of prescriptions was the main denominator used). This error required that the dispensing

history be re-sent from all pharmacies in the trial. This meant that although all data from these

pharmacies was queued up to send, the Server was not consistently up to date with the history files

from each pharmacy.

Due to the problem outlined above, the intervention rate report was often not accurate. The intent was

to use this report as a feedback mechanism for each pharmacy and pharmacist, but doubts about its

currency meant that it was not used. In lieu of this rate-based feedback, feedback was provided on

the number of interventions completed by each pharmacy and pharmacist (see Figure 4.5-5) . This

information was distributed by email, by observers and during PROMISe pharmacy visits by the

Project Team. Each pharmacy received three reports, one during each phase of the study. A final

report was presented to each pharmacy and was the basis for generating an invoice for payment for

interventions.


Page 78 of 361

Pharmacist 1 Name

Pharmacist 1 Name

Pharmacist 1 Name

Pharmacist 1 Name

Pharmacist 2 Name

Pharmacist 2 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacist 3 Name

Pharmacy Name

Figure 4.5-5: Example Of Pharmacy Feedback Report Used In PROMISe Study

4.5.2 Post Trial Information Collection

After completion of the PROMISe trial, feedback was obtained from participating pharmacists in a

number of different ways.

4.5.2.1 Pharmacist opinions of the study

Each pharmacist involved in the trial was given the opportunity to provide feedback on the PROMISe

system and any software issues they had experienced. In particular, questions relating to barriers and

facilitators to recording and performing interventions in community pharmacy were used. They were

also asked to nominate any further improvements in the system. Details of the questions included in

the post-trial questionnaire can be seen in Appendix 16.

4.5.2.2 Focus group sessions

Individuals involved in the project were invited to participate in focus group sessions. This focussed

feedback was facilitated by DeBoos Associates in consultation with the Project Team. These focus


Page 79 of 361

groups sessions were held at a site specifically designed for such sessions with appropriate recording

facilities.

There were two focus group sessions with pharmacists who had used the system extensively and

recorded a significant number of interventions (that is, pharmacists who recorded more interventions

than other pharmacists in the study).

During these sessions, discussion focussed on the areas of:

• documenting interventions,

• attitudes to the interface,

• potential improvements to the interface,

• models for remuneration and

• opinions on the likely adoption of the PROMISe system.

The detailed discussion guide for the PROMISe pharmacist focus groups is included in Appendix 17. A

member of the Project Team attended each of the focus group sessions, but did not participate in the

discussion and the facilitator ensured that they did not influence the outcome of the group.

It was planned to gain feedback from pharmacists who did not use the intervention documentation

system to any great extent, in order to determine what factors influenced this. Pharmacists who

documented less than five interventions over the entire period of the study were contacted in order to

gain their feedback. Unfortunately, we were not able to convince any pharmacists from this group to

participate in a focus group to discuss these issues. In order to obtain this critical feedback, we

conducted face to face interviews with some of these pharmacists (see section 4.5.2.3).

We also conducted a focus group session with the PROMISe observers. This discussion was able to

also focus on the observed characteristics of pharmacists and pharmacies, and the relationship of

these characteristics to recording rates. Each of the seven observers drew on experience from the 27

half-day visits to the three pharmacies they were allocated (nine visits per pharmacy). Hence, the

observers were able to obtain a sound general impression of the functioning of these pharmacies and

their pharmacists’ opinions and attitudes toward intervention detection and documentation. Any

barriers and facilitators that were observed during the visits were explored. The detailed discussion

guide for the PROMISe observer focus groups is included in Appendix 18.

4.5.2.3 In-Depth Interviews and Discussion

Individual in-depth interviews were scheduled with 10 pharmacists who had participated in the trial, but

who declined to attend a focus group session. These interviews were conducted by pharmacists

employed by DeBoos Associates, and no members of the PROMISe Project Team were present

during the interview.

The aim of these interviews was to gain the opinions of pharmacists who had the opportunity to use

the software but did not fully utilise its capabilities. Similar issues were discussed in these interviews

as in the focus groups; however, the format allowed the pharmacists to express their opinions in a one


Page 80 of 361

on one format at a time convenient to them. The detailed discussion guide used for these interviews is

shown in Appendix 19.

4.5.2.4 Further Exploration of Barriers and Facilitators to Community Pharmacy Interventions

In order to further examine possible barriers to recording and performing clinical interventions, a

telephone survey was designed to be administered to:

• the remaining pharmacists in the PROMISe project (those who had not participated in either of the

focus groups or the individual depth interviews; approximately 110 pharmacists), and

• a national sample of pharmacies who had not participated in the PROMISe project at all (sufficient

pharmacists to make the sample for the telephone interview up to 500, sampled according to

population in each state).

Questions were formulated by DeBoos Associates in consultation with the Project Team and aimed to

elucidate the barriers and facilitators to performing and recording interventions in community

pharmacy.

Basic demographics of the respondents were also gathered during the interview. The telephone

interviews were conducted by IView, a company specialising in these types of surveys. The interview

protocol is shown in detail in Appendix 20.

4.5.3 Clinical Panel Assessment of Interventions

Each pharmacist who recorded an intervention was required to make an assessment of the clinical

significance of the intervention which was entered and transferred to the PROMISe database. There

are a number of factors that will influence the pharmacist’s perception of the clinical significance of the

intervention, and during the sample on-line scenarios, it was clear that different pharmacists may rate

the same intervention differently. It was therefore considered inappropriate to use the recording

pharmacist’s indication of clinincal significance as a way of describing the potential value of the

intervention. We chose to select a sample of interventions and have them assessed by a clinical

panel.

There were two reasons for making this additional assessment of the clinical significance of the

interventions.

Firstly, the pilot study data indicated that the recording pharmacists had a tendency to overestimate

the clinical significance of the interventions. This meant that many of the interventions classified as

being of severe clinical significance, when evaluated by other health professionals were not

considered so.

Secondly, and more importantly, the Project Team felt that the assessment of the clinical significance

of the intervention required more than a unilateral estimate of its severity. Although the potential

severity of a particular outcome or consequence from the intervention was an important factor, there

were other factors that should be considered in the assessment of the potential consequences of an

intervention.


Page 81 of 361

The factors to be considered were:

• the likelihood that any particular event or consequence would actually occur after an intervention

• the likelihood that the same consequence may occur despite the intervention

• the likelihood of a less or more severe manifestation of the same consequence

• the likelihood that another health professional may detect and resolve the problem

4.5.3.1 Development of Clinical Panel Assessment Methods

The Project Team have developed a unique clinical assessment method which takes into account the

probability of a consequence occurring (with the intervention and also without the intervention) and

also the “attributability” of the intervention to the pharmacist.

For example, if a patient was taking two different non-steroidal anti-inflammatory agents (NSAIDs) and

the pharmacist intervened to cease one of the agents, one possible prevented outcome would be a

gastrointestinal bleed. The probability that a bleed would have actually occurred, however, would

depend on a range of other factors that may or may not be present in the particular situation being

considered. In the assessment system we have developed, the clinical expert assessor assigns their

estimate of the probability of a gastrointestinal bleed before the intervention (that is, while the person

was taking both NSAIDs), and also assigns the probability that a gastrointestinal bleed would occur

after the intervention (that is, once the person went back to just taking one NSAID, which does still

carry some risk of a gastrointestinal bleed). The expert’s assessments would take into account

whatever other factors are known to be present in the patient, and the potential benefit of the

intervention can be considered in term of the reduced probability of a gastrointestinal bleed.

There is, however, another aspect to consider in the assessment of interventions, that is the severity

of the outcome. To continue with the example above, while there is some probability that the patient

may have a gastrointestinal bleed requiring admission to hospital, blood transfusion and endoscopy,

the probability of a less severe manifestation of the same pathological process may be greater (for

example, a mild sub-clinical bleed that would not require immediate medical management). Thus, we

felt that is was necessary to consider the probability of different levels of severity (severe, moderate

and mild levels) of any consequence. In this example, the expert assessor would be asked to assign

the before and after probability of a severe gastrointestinal bleed, a moderate gastrointestinal bleed

and a mild gastrointestinal bleed. The overall potential clinical significance of the intervention would

therefore be considered in term of the changes in probability of the different levels of severity of the

different potential consequences of the intervention.

4.5.3.1.1 Consequences Impacted upon by Clinical Interventions

A set of clinical consequences for interventions was developed by the Project Team. Consequences

were common diseases or signs or symptoms that were impacted on by interventions. The

consequences were grouped into main diagnostic categories (MDCs) and definitions were prepared to

describe the severe, moderate and mild levels for each consequence. Examples of two clinical

consequences and their definitions are shown in Table 4.5-1.


Page 82 of 361

MDC Code

MDC heading

Sub-group Code

Subgroup Sub-Group

Severity Code Subgroup Severity Description

5 circulatory system

5.02 Hypertension 05.02Mild Mild signs or symptoms which resolve without intervention


05.02 Hypertension 05.02Moderate

Moderate elevation of blood pressure requiring modification of or commencement of medical management


05.02 Hypertension 05.02Severe

Acute injury to target organs (e.g. renal, ocular or cerebral) requiring prompt medical management

6 digestive system

06.01 Gastrointestinal bleeding

06.01Mild

Occult gastrointestinal bleeding likely to require medical management only if persistent

6 digestive system


06.01Moderate Overt gastrointestinal bleeding requiring medical management

6 digestive system


06.01Severe

Overt gastrointestinal bleeding with haemodynamic consequences requiring admission to hospital and prompt medical management

Table 4.5-1: Examples Of Clinical Consequences And Their Severity Descriptions

Each consequence of an intervention (for example a severe gastrointestinal bleed) has a level of

disability and expense associated with it. We expanded the consequences table to include a value in

terms of a number of different economic and non-economic parameters.

The different parameters were:

• Impact on Health Status

o This was a scale of impact based on the severity of the particular consequence, with 1

being mild impact on health and 3 being a severe impact on health.

• Duration of health status impact

o This was a value in days of the duration of the health impact. For chronic conditions, a

one year timeframe was considered.

• Duration of Admission

o The duration in days of any admission associated with the consequence. Where the

consequence definition matched that of an existing AR-DRG definition, the information

was obtained from the National Hospital Cost Data Collection Cost Weights for ARE-

DRG Version 4.2, Round 7 (2002-2003). Where no matching definition existed, the

average duration of admission was used.

• Cost of Admission

o A value in dollars for any admission associated with the consequence. This was

determined from the same information as the Duration of Admission.


Page 83 of 361

• Number of General Practitioner Consultations

o The number of community based general practitioner consultations required to

manage the particular consequence.

• Cost of General Practitioner Consultations

o The total cost of the general practitioner consultations, based on an average of 3:1

Level B (Item 23) to Level C (Item 26) consultations as per the 2005 Medicare

Schedule.

• Number of Specialist Consultations

o The number of specialist consultations required to manage the particular

consequence

• Cost of Specialist Consultations

o The total cost of the specialist consultations, based on an initial consultation cost

according to MBF Item 110 and subsequent consultation costs according to MBF Item

26.

• Investigation and Pathology Costs

o The costs of typically required investigation or pathology tests required in the

management of the particular consequence. These were based on the schedule fee

for the appropriate item.

The initial estimates for these parameters were made by the Project Team, and then reviewed and

modified by a consensus group process that included a physician, a general practitioner and two

experienced clinical pharmacists. The values assigned to the examples outlined in Table 4.5-1 are

shown in Table 4.5-2.

A complete copy of the consequences table that was used in the study, the definitions for each level of

severity for each consequence, the value for each level of severity, and the source for the value

assigned can be found in Appendix 21.


Page 84 of 361

Subgroup Health Status Impact

Dur’n of Health Status Impact

Duration of

Admiss’n

Cost of Admiss’n

No. of GP

Cons.

Cost of GP Cons.

Number of

Specialist Consults

Cost of Specialist Consults

Investi-gation Cost

Hypertension

05.02Mild 1 360 0.00 $0 3 $113

Hypertension

05.02Moderate 2 360 0.00 $0 8 $302 $85

Hypertension

05.02Severe 3 90 3.65 $2,381 4 $151 4 $320 $85

Gastrointestinal bleeding

06.01Mild 1 180 0.00 $0 2 $76 $35


06.01Moderate 2 60 1.68 $1,199 1 $38 1 $128 $1,847


06.01Severe 3 90 5.62 $3,881 2 $76 2 $192

Table 4.5-2: Examples Of Values Assigned To Consequences

4.5.3.1.2 Attributability of the Intervention

Each expert clinical assessor was asked to assign a value for the “attributability” of each reviewed

intervention to the pharmacist. This involved consideration of the probability that another health

professional (the patient’s GP or someone else) would have detected the problem and performed the

same intervention. This value was based on each assessor’s own clinical experience and

understanding of the likelihood of another healthcare professional being involved.

Thus, the final process for assessment of clinical interventions involved each assessor assigning a

before and after probability for consequences that they selected from the consequences table and

also an assessment of the attributability of the intervention to the pharmacist. In this way, the value

associated with each sampled intervention could be estimated and then this value could be

extrapolated back to the entire PROMISe dataset. This procedure is represented in Figure 4.5-6.


Page 85 of 361

PROMISe Dataset of

Clinical Interventions

Sample of Clinical Ints

Expert Assessment

Panel Members

Consequences Table

(Value of each consequence)

Value of Difference in Probabilities

Value of Sample of

Clinical Ints

Before and After Probabilities

Attribution

Extrapolation

Selection and assessment of

probability

Figure 4.5-6: Outline Of Expert Assessment Process For Clinical Interventions

4.5.3.2 Preparation of Information for Assessment by Clinical Panels

When the participating pharmacists submitted a clinical intervention, they selected the clinical

significance of their intervention (nil, low, mild, moderate, high). As outlined in section 4.1.3.5, in those

interventions that the pharmacists nominated as of moderate or severe clinical significance,

pharmacists were asked to provide additional information concerning the patient’s medical condition

and also any other clarifying notes. From these notes, and other information available concerning the

clinical intervention, each of the interventions were reviewed and assessed by two clinical

pharmacists.

This assessment involved review of the information recorded and the pharmacy’s dispensing history

for that patient. Where further clarification was required, the recording pharmacist was contacted by

the clinical pharmacists undertaking the preliminary assessment. This initial assessment was made on

a daily basis during the data collection period. One major component of the preliminary assessment

was to develop a written summary of the reconstructed situation that was documented. To facilitate

this review, a web-based formatted summary of the intervention was developed (see Figure 4.5-7).

Once all moderate and severe significance clinical interventions were assessed and a summary was

prepared, a selection process was used to determine which of these would be further assessed by a

clinical panel (see section 4.5.4).


Page 86 of 361

Figure 4.5-7: Interface For Review Intervention Submitted By Pharmacist

4.5.3.3 Clinical Panel Composition and Panel Access to Intervention Information

Four clinical panels were organised each containing 2 pharmacists, 2 general practitioners and where

possible a physician (see Table 4.5-3). The panel members were recruited by email and telephone

follow-up with the aim to attract members with a variety of clinical experience.

To allow the assessment of the interventions a web-based interface was developed (see Figure

4.5-8). The interface included information for the assessors and allowed them to submit their opinions.

The information that was made available to each assessor consisted of a summary of the intervention

scenario and a dispensing history for that patient from the pharmacy that the intervention took place in.


Page 87 of 361

Panel Number Composition Interventions Assessed

Panel 1 1 Physician 2 General Practitioners 2 Pharmacists

• 51 Common type interventions (reviewed by all panels)

• 60 Random additional interventions (unique to panel 1)

Panel 2 1 Physician 2 General Practitioners 2 Pharmacists



Panel 3 2 General Practitioners 2 Pharmacists



Panel 4 2 Pharmacists



Total 16 Assessors

• 51 Common interventions (16 opinions)

• 60 Panel 1 Interventions (5 opinions)



• 60 Panel 4 Intervnetions (2 opinions)

Table 4.5-3: Composition Of Clinical Panels To Assess Intervention

4.5.3.4 Clinical Panel Assessment Process

Each assessor was provided with an audiovisual presentation outlining the exact process for entering

their assessments. In addition, a written set of instructions were provided (see Appendix 22), and each

assessor was personally contacted and guided through their first intervention assessment with the

project manager.

In addition to the practical aspects of how to provide an assessment, the instructions provided to each

assessor also addressed some of the underlying assumptions that needed to be taken into account in

determining the consequences of the intervention. Assumptions made were designed to ensure that

the assessments were consistent between assessors.

The assumptions which the assessors were asked to keep in mind were:

• that if the person has other pathologies, that these are adequately controlled

• that for interventions where preventative therapy is added, the timeframe for the consideration of

consequences is one year.


Page 88 of 361

Summary provides detail of the intervention

List of medications dispensed

Selecting Consequence(s) of the intervention







Figure 4.5-8: Clinical Panel Assessment Information

Once a consequence was selected, the assessor considered the probability of a severe, moderate or

mild event occurring. Definitions relating to each consequence could be viewed by hovering the mouse

over the level of severity. For example, for the consequence of seizure (seen below Figure 4.5-9), the

severe level consequence relates to the likelihood of "severe seizure resulting in hospitalisation and

intravenous anticonvulsants".


Page 89 of 361

Figure 4.5-9: Clinical Panel Assigning Likelihood Of Different Events

Figure 4.5-10: Clinical Panel Estimate The Attributability To The Pharmacist

The second step in the assessment of the intervention is to review the probability that no other health

care provider would have identified the issue if the pharmacist had not intervened, that is, attributability

(see Figure 4.5-10). This attribution value was used during the analysis of the interventions to assess

the impact of the pharmacists’ actions.

On completing the assessments, it was possible for each assessor to view the overall results for a

particular intervention (see Figure 4.5-11).


Page 90 of 361

Figure 4.5-11: Clinical Assessment Panel Display Of Individual Interventions Results (Example Only)

A comments section was also included in the interface to allow the assessor to incorporate issues that

may have been relevant to their assessment. The individual results could be viewed Figure 4.5-11 in

reference to the other assessors’ opinions. The assessors were given approximately 2 weeks to

complete the assessment of 111 interventions each.

4.5.4 Sampling of interventions for assessment

The clinical interventions were reviewed for similarities, and a grouping system based on common

types of interventions was prepared. Examples of these common types of interventions were included

in the clinical panel assessment, as representative of the other interventions of the same type within

the dataset.

The common types of interventions, and the frequency with which they occurred in the assessment

sample and the entire dataset are shown in Table 4.5-4. The 64 interventions included in the

assessment panel that were of these common types were specifically chosen rather than randomly

selected. These interventions were chosen as they were typical of the remainder of the interventions

of the same common type.

In all, 291 interventions were assessed by the assessment panels, and the remaining 227

interventions were randomly selected from the remainder of the dataset (not including those that were

of a common type). Information concerning the common type of the intervention and other


Page 91 of 361

characteristics of the intervention was used to extrapolate the results of the clinical assessment panel

to the entire dataset (see section 6.3.4)

Common Type

CodeIntervention Common Type Description

Assessment

Panel

Remainder of

DatasetA Excess Antihypertensive Agent 5 68

B Insufficient Antihypertensive Agent 3 68

C Previous Allergy to Prescribed Agent 2 26

DPaediatric Antibiotic Dose Below

Recommendation2 13

E Antidepressant Crossover or Cessation Issue 1 8

F Excessive Serotonin Risk (Serotonin Syndrome) 1 22

G Prophylactic Aspirin Recommended 2 200

H Compliance Issue with Statins 1 11

I General Compliance Issue with Multiple Agents 5 56

J Duplication of NSAIDs 3 20

K Excessive Paracetamol Dosing (adult) 2 13

L Excessive Paracetamol Dosing (paediatric) 2 4

M NSAID Prescribed with a Gastrointestinal Disease 6 9

N Respiratory Device Usage Issue 3 56

O Preventive Vaccination Issue 1 17

P Osteoporosis Preventive Issue 2 24

Q Compliance Issue with Antidiabetic Agents 4 32

R Statin Dose Prescribed Incorrectly 3 19

S Oral Contraceptive Efficacy Issue 4 20

T Paediatric Dose Too High 4 11

U Respiratory Compliance Issue 4 19

V Respiratory Management Issue 4 38

64 754

227 1351

291 2105

Subtotal

Remaining Interventions

Total

Table 4.5-4: Common Types Of Interventions And Number Of Each In Assessment Sample

4.6 Economic Analysis

The overall purpose of the economic analysis was to estimate the economic value of community

pharmacists’ clinical interventions related to prescription medications.

The three specific objectives of the economic analysis were:

1) To estimate the economic value to the pharmacy of the pharmacists’ intervention – essentially

the opportunity cost of the pharmacists’ time

2) To estimate the economic value to the patients and the health care system of the clinical

interventions.

3) To estimate the economic value of changing the rate of clinical interventions.


Page 92 of 361

The analysis assessed three main components:

• Capacity to perform interventions– what is the capacity for intervention?

• Activity within the pharmacy – what is the rate of intervention?

• Value of the interventions– what is the value of intervention?

Capacity is determined by pharmacy staff hours and skill mix (pharmacist, graduate pharmacist and

dispensing technician).

Data collection: A form was sent to each pharmacy to ascertain what their staff mix was over the hours

which the pharmacy was open. This was then to be faxed back to the researchers; where information

was not returned, each pharmacy was contacted individually to gather the information on staff mix

(see 4.3.3.3)

Capacity is of interest because it may be a determinant of the rate at which pharmacists perform

proactive clinical interventions. While prescriptions per day are an indicator of activity in a pharmacy

and the opportunity to perform interventions, if a pharmacy is busy in terms of prescription numbers,

an individual pharmacist’s capacity for performing interventions depends upon the number of

pharmacists, graduate pharmacists and dispensing technicians who are available. If a pharmacist’s

workload is a determinant of rate of interventions then changing pharmacist workload may change rate

of interventions.

Capacity was calculated in three ways:

• Pharmacist hours per day

• Pharmacy hours and pharmacy graduate hours

• The above plus dispensing technician hours.

The ratio of prescriptions to pharmacy capacity (calculated three ways as defined above) was used to

determine how busy each pharmacy was on a given day and then compared to the rate of clinical

interventions, and the rate of proactive and reactive interventions per 100 prescriptions.

The number of proactive, reactive and aspirin-prompted interventions were ascertained for each

unique pharmacy-day of activity. Total prescriptions, total new prescriptions and total number of

patients, as well as patients with different numbers of medications were also included in this dataset.

Interventions were sampled according to the methods outlined in section 4.5.4 and an assessment

was made by several clinical panels as outlined in section 4.5.3.1. Extrapolation of the results of the

panel to the remainder of the dataset is outlined in section 6.


Page 93 of 361

4.6.1 Detailed Economic Analysis Methods

The analysis we performed had five sequential and dependent parts.

1) Estimate the rate of opportunity for interventions viz. the rate of interventions under perfect

conditions (100% identification and action by pharmacists). That is, estimate the capacity for

interventions.

2) Estimate the actual rate of intervention in community pharmacies, per 1000 prescriptions. That

is, estimate the intervention activity in each pharmacy.

3) Estimate the average value of an intervention and combine with 2) above to estimate the total

value of interventions recorded in the PROMISe study.

4) Estimate the rate and value of interventions nationally by extrapolating to national

prescriptions dispensed, using 3).

5) Estimate the potential for increased or improved rate of intervention, combining each of the

above steps.

The key results from each of these steps, and the key methodological issues, are presented in section

6 with a more detailed summary of the method we used.


Page 94 of 361

5 Results and Discussion Part 1: Nature and Frequency of Interventions and Factors Affecting Intervention Rate

5.1 Data Collection

Pharmacists recorded their clinical activities from the 21st April to the 17th June 2005. They were

encouraged to record any professional activity related to prescription medicines (outside basic

dispensing and counselling procedures) directed toward improving health outcomes, the quality use of

medicines or the provision of health-related information. Automation of recording of brand substitution

allowed separate analysis of this commonly occurring situation. The pharmacists were provided with

ongoing feedback and support during the data collection.

5.2 PROMISe Pharmacy and Pharmacist Recruitment

Pharmacies which used the WiniFRED dispensing software in the greater metropolitan area of

Melbourne were contacted for recruitment (260 pharmacies). Of these, 75 (29%) responded to the

initial invitation of interest, which was disseminated by mail and email. From the 75 expressions of

interest 23 of the pharmacies were excluded because of software or location requirements. Hence,

there were 52 pharmacies enrolled in the PROMISe Intervention study. These pharmacies were

located in the greater metropolitan area of Melbourne, with pharmacies dispersed from Werribee to

Lilydale taking part in the project (see Figure 5.2-1).

Each pharmacy provided the names and contact details of pharmacists who would be working in their

pharmacy during the trial period. Each of these pharmacists was contacted and provided with

information on the project and given access codes and information regarding the on-line DOCUMENT

training. Within the first five days of the study, 125 of these pharmacists had completed the 20

scenarios and other aspects of the DOCUMENT training.

Figure 5.2-1: PROMISe Intervention Study Pharmacy Locations


Page 95 of 361

5.2.1 PROMISe Pharmacist Training

There were three evening orientation and training sessions held for the PROMISe Project. Overall,

150 participants were either able to attend one of the training evenings or received an audiovisual

presentation on CD-ROM, which provided the information in a reviewable form.

Attendance at training sessions

Sunday 17th April 55

Monday 18th April 49

Tuesday 19th April 27

Unavailable for training provided with electronic presentation

19

Total 150

Table 5.2-1: Attendance At Training Evenings

Figure 5.2-2: Orientation And Training Evenings For PROMISe Project

Those pharmacists who worked at least one full day when the project was running were encouraged to

undertake the training. In all, 301 different pharmacists dispensed prescriptions during the PROMISe

project. Of the 150 who attended the training session, 125 completed the online training.

Therefore, a number of non-PROMISe trained pharmacists worked in the pharmacies during the study.

Although every effort was made to provide support for these pharmacists, a number of these

pharmacists dispensed prescriptions without documenting any interventions. During the trial 58

pharmacists dispensed prescriptions but did not record any interventions. A further 19 pharmacists

recorded only 1 non-clinical intervention. These 76 pharmacists dispensed a total of 8405

prescriptions (~2% of all prescriptions in the study).


Page 96 of 361

5.3 Pharmacy Demographics

The Melbourne PROMISe trial involved 52 pharmacies. A range of demographic information was

collected by providing a questionnaire to the manager or owner of each pharmacy participating in the

study. In order to ascertain the representativeness of the pharmacies in the PROMISe study, we also

invited a sample of 200 Victorian metropolitan pharmacies to complete the same owner/manager

questionnaire. These were all pharmacies which used the same WiniFRED dispensing system, and

had been sent preliminary information on the PROMISe project, but had not responded that they

wished to participate. Of these non-participating pharmacies, 38 responses were received (19% return

rate).

Section 5.3.1 outlines the demographic caharacterisitcs of the PROMISe participant pharmacies, while

section 5.3.2 outlines the same information specific to the non-participating group of pharmacies. A

comparison is made between the PROMISe and non-participant group in section 5.3.3.

5.3.1 Characteristics of the PROMISe Pharmacies

5.3.1.1 Location and Size of the Pharmacy

The type of location of each pharmacy is shown in Figure 5.3-1. The most common location was

within a strip of other suburban shops (23 pharmacies, 47%).

10, 21%

7, 15%

8, 17%

23, 47%

Local centre <25

Major centre >25

Medical centre

Strip of shops

Figure 5.3-1: The Location Of Pharmacies Involved In The PROMISe Project

With respect to size of the pharmacies, 13 of the 48 were <100m2 (see Figure 5.3-2). Of these “small”

pharmacies, seven were located in street strips of shops, and 4 in smaller local shopping centres.


Page 97 of 361

13

10

13

8

4

0

2

4

6

8

10

12

14

<100 101-150 151-250 251-500 >500

Area metres2

No

. P

harm

acie

s

Victorian Metropolitan

Average 160m2

Figure 5.3-2: The Size Of Pharmacies Involved In The PROMISe Project (m2)

The average size of pharmacies in the promise study was 150-220m2, which is slightly larger than the

Victorian metropolitan average size of 160m2 (Guild Digest 2004).

The largest single combination of location type and size was the small (less than 100m2) pharmacy

located in a shopping strip (see Table 5.3-1).

<100 101-150 151-250 251-500 >500

Local <25 4 1 2 3 10

Major >25 1 4 2 7

Medical centre 2 3 3 8

Strip 7 5 4 3 4 23

Total 13 10 13 8 4 48

Size (square metres)Location Type Total

Table 5.3-1: Type Of Location Versus Size Of PROMISe Pharmacies.

5.3.1.2 Workload of the pharmacy

The pharmacies involved in the trial fell into two main groups according to their opening hours:

• those that were open mostly weekdays, or

• those that were open seven days a week.

The histogram of opening hours (see Figure 5.3-3) shows the two groups, with those pharmacies

open for seven days predominantly being open for 85 hours (~12 hours per day) and those

pharmacies that were open for five days, predominantly being open for 55 hours (~ 10hours per day).

The average number of days open each week was 6.5 (Range 5 – 7; Standard deviation 0.6).


Page 98 of 361

The minimum total hours open each week was 43 and the maximum was 94 hours. This is in keeping

with the average total opening hours per week of 63 hours based on the Guild Digest (2004) results

for Victorian metropolitan pharmacies.

The PROMISe sample seems representative in terms of size and opening hours.

2

18

9

5

11

3

0

2

4

6

8

10

12

14

16

18

20

45 55 65 75 85 95

Hours open per week

Victorian Metropolitan Average 63hrs

Figure 5.3-3: Weekly Hours Open for PROMISe Pharmacies

The pharmacies that participated in the PROMISe project were asked to estimate the number of

prescriptions dispensed each week. The categories available are seen in Figure 5.3-4.


Page 99 of 361

8

11

16

10

30

2

4

6

8

10

12

14

16

18

<400 400-750 750-1250 1250-2000 >2000

Average number of prescriptions dispensed per week

No

. o

f P

ha

rma

cie

s

Victorian Metropolitan Average 1058

Figure 5.3-4: The Average Number Of Prescriptions Dispensed By The PROMISe Pharmacies Each Week

The distribution of the prescription workload of the pharmacies generally reflects the average from the

Guild Digest for metropolitan Victorian pharmacies, of 1058 prescriptions per week. There were three

pharmacies within the sample which dispensed twice this Victorian average amount.

5.3.1.3 Staff mix of the pharmacy

Those pharmacies with the high workload necessarily employed more pharmacists. The number of

full-time equivalent (FTE) pharmacists employed in each pharmacy ranged from 1 to 5. Two peaks

are evident in the histogram of FTE pharmacists, indicating that the majority of pharmacies employed

either one or two full-time pharmacists.


Page 100 of 361

0

18

3

15

1

6

1 2 0 20

2

4

6

8

10

12

14

16

18

20

0.5 1

1.5 2

2.5 3

3.5 4

4.5

Mor

e

Number of Full Time Pharmacists

Nu

mb

er

of

Ph

arm

acie

s

Figure 5.3-5: Number Of Pharmacists Working Within PROMISe Pharmacies

The support staff within each pharmacy included dispensary technicians, graduate pharmacists and

pharmacy assistants. There were 10 pharmacies involved in the trial who employed one or more

graduate pharmacists. In the previous 2 years, 40% (20) of the pharmacies had employed a graduate

pharmacist. Approximately 60%(30) of the pharmacies employed a dispensary technician and most

pharmacies employed at least one other pharmacy assistant (mean was 3.5 FTE pharmacy

assistants). Of these support staff, there was a range of training levels. The number of staff with

Certificate II qualification or equivalent is displayed in Figure 5.3-6. It can be seen for a number of

pharmacies that their staff had not undergone formalised training.


Page 101 of 361

7

18

10

13

02468

101214161820

none <50% 50% All or most

Percent of Pharmacy Assistant Staff With Certificate II

Qualification

No

. P

ha

rma

cie

s

Figure 5.3-6: Staff With Certificate II Qualification In PROMISe Participating Pharmacies

5.3.1.4 Ownership of PROMISe Pharmacies

The length of time the pharmacy had been in the current ownership is displayed in Figure 5.3-7. There

were pharmacies which had been under their current ownership for a relatively short period of time (9

of the 48 pharmacies <1 year) and the range of time in the current ownership ranged from 0.25 to 40

years. The majority of the pharmacies were managed by individual owners or those in partnership (45

respondents; 19 individual, 26 partnerships).


Page 102 of 361

Bin Frequency

0.5 4

1 5

5 12

10 12

15 4

20 4

30 5

40 2

More 0

4

5

12 12

4 4

5

20

0

2

4

6

8

10

12

14

0.5 1 5 10 15 20 30 40 More

Number of Years of Current Ownership

Nu

mb

er

of

Ph

arm

acie

s

Figure 5.3-7: Duration of Current Ownership for PROMISe Pharmacies

Respondents were asked to nominate if they belonged to a banner or brand group. It was found that

18 of the 48 pharmacies (37.5%) did belong to a banner group, with representation from a number of

types (see Table 5.3-2).

Banner group Number of pharmacies Guardian 8

Amcal 3 Priceline 2

Chemmart 1 Other 1

Table 5.3-2: Banner Group Of Pharmacies Involved In The PROMISe Project

The average annual turnover of metropolitan Victorian pharmacies based on the Guild Digest for 2004

was $2.3 million. The pharmacies involved in the PROMISe project represented a range of annual

income with the majority around the average value (Figure 5.3-8).


Page 103 of 361

7

20

5

12

40

5

10

15

20

25

<1M 1-2M 2-3M 3-4M >5M

Annual Turnover

Nu

mb

er

of

Ph

arm

acie

s

Figure 5.3-8: Average Annual Turnover Of Participant Pharmacies

Thus, from the point of view of ownership and annual turnover, the pharmacies that participated in the

PROMISe project seemed representative.

5.3.1.5 Information Resources and Clinical Services Provided by the Pharmacy

Pharmacists are presented with, and resolve, a variety of clinical situations in their daily practice. The

information resources available in the pharmacy may influence the quality of the response to the

problem. The presence of up-to-date electronic or hard-copy resources, beyond those required by law,

is an indicator of the pharmacists’ attitude towards this aspect of pharmacy practice.

Certain reference materials (in book or electronic form) are compulsory as determined by The

Pharmacy Board of Victoria (published 1st February 2005). These include

• Australian Pharmaceutical Formulary and Handbook (APF) • Australian Medicines Handbook (AMH) • A reference work on prescription products

o APP guide 2005 o MIMS annual with bimonthly addenda

• The Pharmacy Board of Victoria Guidelines • A reference work on drug interactions • The Merck Manual • The Therapeutic Guidelines • The Royal Children’s Hospital Paediatric Pharmacopoeia


Page 104 of 361

The resources available within the PROMISe pharmacies generally reflected those which they were

required to keep (see Table 5.3-3). Approximately half of the pharmacies used the MIMS as their

primary prescription product reference, while the other half primarily used the APP guide.

The usage of the Pharmacy Self Care program was high, with 40 pharmacies (83%) indicating that

they used the program and 24 (60%) indicating that they used it on a regular basis, at least once a

day.

Internet-based drug information was available in 44 (91%) of the pharmacies, but only 12 of these

pharmacies (33%) indicated that they used the resource regularly.

Available used occasionally

Available used

regularly1

Not Currently available

Australian Medicines Handbook 2 26 19 2

APF 2 44 4 0

Pharmacy Self Care 16 24 8

Martindale 43 2 3

eMIMS 3 2 24 22

eAPP 3 8 24 16

eAMH 15 9 24

Therapeutic Guidelines 2 38 10 0

Internet based drug information 32 12 4 1 Defined as at least once a day

2 Compulsory References in Victorian Pharmacies

3 Compulsory to have either MIMS or APP Guide

Table 5.3-3: Resources Available In PROMISe Participating Pharmacies

Questions regarding which additional clinical services were offered by the pharmacies and whether

these were charged for, were asked. Of the pharmacies involved in the PROMISe trial, 8 of the 48

respondents (16%) supplied pharmaceuticals to nursing homes and most had been involved in

receiving payments through the Home Medicines Review (HMR) process. The HMR service was

clearly the main service that the pharmacists felt comfortable charging for (see Table 5.3-4). This is

likely to be related to the fact that the payments for HMRs are not directly from the patient, but rather a

formal payment process claimed through the Health Insurance Commission. The results of this

question seem to indicate that pharmacists are hesitant to charge patients for the clinical services

which they provide.


Page 105 of 361

# % # % # % # %

BP monitoring 10 20.8% 2 4.2% 35 72.9% 1 2.1%

Wound Care 22 45.8% 4 8.3% 21 43.8% 1 2.1%

Diabetes screening 32 66.7% 7 14.6% 9 18.8% 0 0.0%

Weight management

program22 45.8% 9 18.8% 15 31.3% 2 4.2%

Community Education 16 33.3% 7 14.6% 24 50.0% 1 2.1%

HMR's 7 14.6% 1 2.1% 0 0.0% 40 83.3%

Fee Charged

for serviceServiceNot provided

Planning to

provide

Service

provided at no additional cost

Table 5.3-4: Services Offered By The Pharmacies Which Participated In The PROMISe Trial

5.3.1.6 Quality Care Adopter Status of the Pharmacies

The Quality Care Pharmacy Program (QCPP) was initially launched in 1998. It was developed by The

Pharmacy Guild of Australia and is based around the development and implementation of service

standards for each area of community pharmacy. The program aimed to create a quality assurance

program, which in turn would raise the standards of service provided to the public as well as improving

profitability. For a pharmacy to become accredited within the QCPP they must go through a process

whereby they set and implement pharmacy standards based on reference material provided. If there

are areas where the pharmacy does not meet the current standard, then plans to bring them up to

standard are developed. Finally the pharmacy is assessed for accreditation.

Currently there are approximately 4200 accredited pharmacies in Australia.8 The year in which

different pharmacies became accredited during the program is shown in

Figure 5.3-9. The date that a pharmacy becomes accredited can be considered as a broad indicator of

various aspects of the pharmacy’s competitiveness and adaptation to change.

8 as per QCPP website with information current at 14/7/05


Page 106 of 361

3

176

669

931

700

1654

367

36

0

200

400

600

800

1000

1200

1400

1600

1800

1998 1999 2000 2001 2002 2003 2004 2005 to

Jul

Year of QCPP Accreditation

No

. p

ha

rma

cie

s Q

CP

P a

cc

red

ite

d

Figure 5.3-9: The Number Of Pharmacies QCPP-Accredited By Year

A review of the QCP Program was completed during 2005.9 This review developed a classification

system for uptake of the QCP Program based on Roger’s normal distribution theory for the uptake of

innovation. The normal distribution theory proposes that uptake of an innovation follows a standard

Normal Distribution in terms of the time taken for uptake (see Figure 5.3-10).

The first 2.5% of the population to adopt the innovation (greater than 2 standard deviations below the

mean time for uptake) are termed innovators. The next 13.5% (those more than one but less than two

standard deviation below the mean time) are termed early adopters. The next 34% to adopt the

innovation (one standard deviation from the mean) are termed the early majority. The following 34%

(within one standard deviation after the mean time for uptake) are termed the late majority, and the

remaining 16% (greater than one standard deviation after the mean time) are termed laggards.

9 An Evaluation of the Quality Care Practice Program, 2005


Page 107 of 361

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

2 .5 1 6 5 0 8 4 9 7 .5 1 0 0

Innovators2.5%

EarlyAdopters

13.5%

EarlyMajority

34%

LateMajority

34%

Laggards16%

Normal Distribution Theory for Innovation Uptake

Mean + 1 SD- 1 SD- 2 SD

Time for uptake of Innovation

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

2 .5 1 6 5 0 8 4 9 7 .5 1 0 0

Innovators2.5%

EarlyAdopters

13.5%

EarlyMajority

34%

LateMajority

34%

Laggards16%

Normal Distribution Theory for Innovation Uptake

Mean + 1 SD- 1 SD- 2 SD

Time for uptake of InnovationTime for uptake of Innovation

Figure 5.3-10: Normal Distribution Theory For Innovation Uptake

Figure 5.3-11 and Table 5.3-5 show how these proportions were applied to the pharmacies that

provided information to the QCPP review.


Page 108 of 361

3

176

669

931

700

1654

367

360

200

400

600

800

1000

1200

1400

1600

1800

1998 1999 2000 2001 2002 2003 2004 2005 to

Jul

No

. p

harm

acie

s Q

CP

P a

ccre

dit

ed

InnovatorEarly

AdopterEarly

MajorityLate

MajorityLaggards

QCPP Adopter Status

Figure 5.3-11: Determination Of QCPP Adopter Status

QCPP Category Date of Accreditation Innovator Before 1

st Dec 1999

Early adopter Between 1st Dec 1999 and 31

st Jan 2001

Early majority Between 1st Feb 2001 and 31

st Jan 2003

Late majority Between 1st Feb 2003 and 31

st Jan 2004

Laggards After 1st Feb 2004 or not yet accredited

Table 5.3-5: Category Of QCCP Adoption Created During The Review Of The QCP Program

Based on these definitions, the QCPP adopter status of the pharmacies participating in the PROMISe

study was determined from the date of accreditation, which was provided. On first examination (see

Figure 5.3-12 and Table 5.3-6), the PROMISe pharmacies appear to be predominantly (64.5%) the

type of pharmacies which adopt innovation earlier than others. There were, however, 16.7% of the

pharmacies who were placed in the laggard category as they had not yet completed the accreditation

process.


Page 109 of 361

10.4%8.3%

45.8%

18.8%

2.5%

13.5%

34.0% 34.0%

16.7% 16.0%

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

innovator early adopter early majority late majority laggards

PROMISe participants

Australian Pharmacies

Figure 5.3-12: QCPP Status Of The Pharmacies Compared To The Australian Average

Status Date QCPP Accredited PROMISe sites

(n=48)

Innovator Before Dec 99 5 Early adopter 1999 - 2001 4 Early majority 2001 - 2003 22 Late majority 2003- 2004 9

Laggards After Feb 04 or not accredited 8

Table 5.3-6: QCPP Status Of Pharmacies Participating In PROMISe

We reviewed of the length of time the pharmacy had been owned and compared this to the QCPP

adopter status. We found that six of the 17 pharmacies that had initially been classified as laggards or

late majority adopters had only been owned by their current owners for less than 12 months. Given

that the process for QCPP can take many months, we felt it best to remove these pharmacies from

this part of the analysis.

In Figure 5.3-13 those pharmacies which had been in the current ownership for less than one year are

excluded. The distribution of QCPP adopter status suggests that over 75% of the pharmacies

participating in the PROMISe study are more likely to take up an innovation earlier than the average

uptake time. Furthermore, 21% of the pharmacies were either innovators or early adopters.


Page 110 of 361

10.5% 10.5%

52.6%

21.1%

5.3%2.50%

13.50%

34.00% 34.00%

16.00%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%


Pharmacies adjusted for >1 year

Australian Pharmacies (Jan 04)n=4907

Figure 5.3-13: QCPP Status, Excluding The Pharmacies Owned For <1year

5.3.1.7 Entrepreneurial Orientation (EO) of the pharmacy

The Likert scale responses to the 14 statements were amalgamated to determine level of agreement

or disagreement with the statement. The results of these amalgamated scores are shown in Table

5.3-7.

There were particularly strong responses to five of the statements by the respondents.

• Forty of the respondents (87%) indicated that they strongly agreed with the statement that their

pharmacy encouraged the development of innovative services.

• Forty-two of the respondents (89%) strongly agreed that their pharmacy is constantly seeking out

new opportunities.

• Forty of the respondents (87%) strongly agreed with the statement that ideas from staff are

supported by management in their pharmacies.

• Forty-three of the respondents (93%) strongly agreed that their pharmacy was ambitious about the

service provided.

• Thirty-eight of the respondents (83%) strongly disagreed that there had been little change in their

pharmacy over the last 10 years.

Thus, it seemed that the pharmacies that participated in the PROMISe trial considered themselves to

be innovative, autonomous and ambitious.

A comparison of the EO of the PROMISe pharmacies to those pharmacies who declined participation

in the project can be found in section 5.3.3.2.2.


Page 111 of 361

Unsure

# % # % # % # % # % #

This pharmacy is known for

innovation among the

pharmacies in this area

9 21.4% 18 42.9% 7 16.7% 6 14.3% 2 4.8% 5 47

This pharmacy and

management encourages the

development of innovative

services

17 37.0% 23 50.0% 3 6.5% 2 4.3% 1 2.2% 1 47

Because conditions are

changing we continually seek

out new opportunities

19 40.4% 23 48.9% 1 2.1% 3 6.4% 1 2.1% 0 47

The management team is

closely able to predict the

future needs of this business

6 13.6% 22 50.0% 9 20.5% 6 13.6% 1 2.3% 3 47

It is out business strategy to

avoid taking too many chances0 0.0% 14 29.8% 9 19.1% 18 38.3% 6 12.8% 0 47

If there were a risky new

project or service this

pharmacy would be prepared to

take it on

12 27.9% 17 39.5% 7 16.3% 2 4.7% 5 11.6% 4 47

Ideas for new services in the

pharmacy from staff are

supported by the management

team

12 26.1% 28 60.9% 4 8.7% 2 4.3% 0 0.0% 1 47

At this pharmacy it is primarily

the management team who

identify new business

opportunities

11 23.4% 21 44.7% 6 12.8% 8 17.0% 1 2.1% 0 47

At our pharmacy we are

ambitious about the service we

provide

20 43.5% 23 50.0% 1 2.2% 1 2.2% 1 2.2% 1 47

Our actions towards

competitors can be termed

aggressively competitive

7 14.9% 15 31.9% 4 8.5% 17 36.2% 4 8.5% 0 47

We are aware and responsive

to changes that other

pharmacies in our area make

13 27.7% 24 51.1% 6 12.8% 3 6.4% 1 2.1% 0 47

Five-year plans for this

pharmacy are a high priority10 21.7% 24 52.2% 5 10.9% 4 8.7% 3 6.5% 1 47

The majority of staff in this

pharmacy have worked here for

more than 5 years

12 26.1% 11 23.9% 6 13.0% 8 17.4% 9 19.6% 1 47

There has been little change in

our pharmacy over the last 10

years

1 2.2% 6 13.0% 1 2.2% 18 39.1% 20 43.5% 1 47

% excludes the unsure respondants for each statement

Statement and


Parameter Adressed

Innovation

Proactiveness

Risk Taking

Autonomy

Work Ethic

Total

Strongly

DisagreeDisagree


Management Structure of Pharmacy

NeutralAgreeStrongly Agree

Table 5.3-7: Results For Statements Used To Evaluate Entrepreneurial Orientation

The responses to these statements have been used to calculate the Entrepreneurial Orientation

scorebased on Doucette et. al. formula (see section 4.3.3.2). The results of the EO Score are shown in

Figure 5.3-14.


Page 112 of 361

0 = Low Score, 10 = High Score

0 5 10 15 20 25 30

Entrepreneurial Orientation Score

0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 21.1493Std. D ev. = 3.11424N = 48

Study participant: yes

Figure 5.3-14: Entrepreneurial Orientation Score for PROMISe Pharmacies

The maximum EO score possible is 45.4, which would be scored if a pharmacy answered 100%

agreement with all of the appropriate statements.

5.3.2 Non- PROMISe Pharmacy Characteristics

In order to ascertain the representativeness of the pharmacies in the PROMISe study, we invited a

sample of 200 Victorian metropolitan pharmacies to complete the same owner/manager questionnaire.

These were all pharmacies which used the same WiniFRED dispensing system, and had been sent

preliminary information on the PROMISe project, but had not responded that they wished to

participate.

Of these pharmacies, 38 responses were received (19% return rate). Sections 5.3.2.1 to 5.3.2.6

outline the information specific to the non-PROMISe group of pharmacies, while a comparison is made

between the PROMISe and non-PROMISe group in section 5.3.3.

5.3.2.1 Location and Size of Non –Participant Pharmacies

Half of respondents were located within a strip of other shops.

In this sample of pharmacies there was representation from pharmacies located within medical

centres. As with the PROMISe participant sample there were no pharmacies located within a hospital.


Page 113 of 361

11, 29%

6, 16%

2, 5%

19, 50%

Local centre <25 shops

Major centre >25 shops

Medical centre

Strip of shops

Figure 5.3-15: Location Of Pharmacies (Non-Participant)

The size of the pharmacy in this sample is inline with the Victorian average (metropolitan average =

160 m2 ), with a spread of sizes across the range of areas provided.

5

12

8

1

12

0

2

4

6

8

10

12

14

<100 101-150 151-250 251-500 >500

Area metres2

No

. P

harm

acie

s

Victorian Metropolitan Average 160m2

Figure 5.3-16: Size Of Non Participant Pharmacies (m2)

5.3.2.2 Workload of the pharmacy (non-participant)

The hours open each week (see Figure 5.3-17) covers a range of responses. The single pharmacy at

the higher end of the histogram operated for 24 hours a day 7 days a week. The pharmacies were

open for between 5.5 and 7 days each week with the mean number of days open 6.6 (SD 0.6). The

pharmacies opening times varied between 48.5 and 168 hours each week (Median 80hrs per week).


Page 114 of 361

78

6

11

5

0 0 01

0

2

4

6

8

10

12

50 60 70 80 90 100 120 150 170

Victorian Metropolitan Average 63hrs

Figure 5.3-17: Weekly Hours Open for Non Participant Pharmacies

The prescription workload of the pharmacies which responded to the questionnaire reflects a range of

relatively busy pharmacies. The average number of prescriptions dispensed weekly in metropolitan

Victoria pharmacies was 1058 in 2003 (Guild Digest).

6

12

10

8

20

2

4

6

8

10

12

14

<400 400-750 750-1250 1250-2000 >2000

Average No. Prescriptions per week

No

. P

harm

acie

s

Victorian Metropolitan Average 1058

Figure 5.3-18: Average Number Of Prescriptions Dispensed Each Week

5.3.2.3 Staff Mix Of Non-PROMISe Pharmacies

Seventeen of the responding pharmacies employed one FTE pharmacist (45%) (see Figure 5.3-19).


Page 115 of 361

17

3

9

6

0

21

0

2

4

6

8

10

12

14

16

18

1.0 1.7 2.3 3.0 3.7 4.3 More

Figure 5.3-19: Number Of Full Time Equivalent Pharmacists In Non-Participant Pharmacies

Other staff employed within the pharmacy included a number of graduate pharmacists (9 of 38

pharmacies), of whom between 0.5 and 3 FTE were employed in these pharmacies. Seventeen (45%)

of respondent pharmacies had employed a graduate pharmacist within the last two years.

Dispensary technicians were employed approximately as often, with 18 of the 38 respondents

employing at least 1 FTE dispensary assistant (47%).

Pharmacy assistant staff with a high level of recognised qualifications (Certificate II) were only

represented in a relatively small proportion of the pharmacies (5 of 38 respondents; see Figure

5.3-20).


Page 116 of 361

6

18

9

5

0

2

4

6

8

10

12

14

16

18

20

none <50% 50% All or most

Percent of Assistant Staff

No

. P

harm

acie

s

Figure 5.3-20: Proportion Of Assistant Staff With Certificate II Level Training

5.3.2.4 Ownership of Non-Participant Pharmacies

The duration of current ownership of the non-participant pharmacies was between 0.8 to 84 years (an

organisation was listed as the owner of a pharmacy for 84 years). Generally, pharmacies within this

sample had been under the current ownership for less than 10 years (see Figure 5.3-21).

1

19

5

3

1

3

0

2

4

6

8

10

12

14

16

18

20

1 7.8 14.6 21.4 28.2 More

Figure 5.3-21: Duration Of Current Ownership for Non-Participant Pharmacies


Page 117 of 361

The were a number of respondents which belonged to a banner group, 20 of the 38 (52%) The groups

which were represented are outlined in Table 5.3-8.

Banner Group Number of pharmacies

Amcal 6

Chemmart 3

Guardian 3

Healthsense 1

My Chemist 1

Pharmacist advice 1

Pharmacy Plus 1

Priceline 3

Quality pharmacy 1

Table 5.3-8: Banner Groups Amount Non-Participant Pharmacies

5.3.2.5 Quality Care Adopter Status of Non-Participant Pharmacies

The quality care adopter status of non-participant pharmacies was determined by the method

previously outlined. As the distribution of the adopter status from this group of pharmacies matched

the distribution of the Australian average adopter status. A comparison will be made with PROMISe

pharmacies in section 5.3.3.

18.4%

44.7%

21.1%

13.5%

34.0% 34.0%

2.6%

13.2%16.0%

2.5%

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

early adopter early majority late majority laggards innovator

Non Participants

Australian AVG

Figure 5.3-22: Quality Care Adopter Status for Non-Participant Pharmacies

5.3.2.6 Entrepreneurial Orientation of Non-Participant Pharmacies

Pharmacies that were offered participation in the PROMISe project but declined, completed the same

opinion based statements as those that participated in the project. This allowed calculation of the

various parameters used for calculation of entrepreneurial orientation.


Page 118 of 361

The Likert scale scores of zero to 10 were converted into agreement and disagreement levels, and the

results of these are shown in Table 5.3-9. As with the PROMISe pharmacies, a number of strong

opinions were expressed.

• Thirty of the pharmacies (81%) felt that new ideas for services from the staff was supported by the

management team

• Thirty-one of the pharmacies (84%) felt that their pharmacy was ambitious about the service they

provided

• Thirty-one of the pharmacies (82%) felt that they were aware it and responsive to changes that are

the pharmacies in the area make.

Thus, non-participant pharmacies felt that they had elements of autonomy, ambition and competitive

aggressiveness.

A comparison of these parameters with the PROMISe pharmacies will be carried out in section 5.3.3.

Table 5.3-9: Entrepreneurial Orientation Statements For Non-Participant Pharmacies

Unsure Total

# % # % # % # % # % # #

This pharmacy is known for innovation

among the pharmacies in this area3 10.7% 5 17.9% 9 32.1% 9 32.1% 2 7.1% 10 38

This pharmacy and management

encourages the development of

innovative services

9 23.7% 19 50.0% 8 21.1% 2 5.3% 0 0.0% 0 38

Because conditions are changing we

continually seek out new opportunities11 28.9% 20 52.6% 4 10.5% 3 7.9% 0 0.0% 0 38

The management team is closely able

to predict the future needs of this

business

4 11.8% 16 47.1% 9 26.5% 3 8.8% 2 5.9% 4 38

It is out business strategy to avoid

taking too many chances3 9.4% 9 28.1% 5 15.6% 10 31.3% 5 15.6% 6 38

If there were a risky new project or

service this pharmacy would be

prepared to take it on

2 6.3% 16 50.0% 6 18.8% 7 21.9% 1 3.1% 6 38

Ideas for new services in the pharmacy

from staff are supported by the

management team

12 32.4% 18 48.6% 4 10.8% 2 5.4% 1 2.7% 1 38

At this pharmacy it is primarily the

management team who identify new

business opportunities

10 26.3% 18 47.4% 4 10.5% 5 13.2% 1 2.6% 0 38

At our pharmacy we are ambitious

about the service we provide12 32.4% 19 51.4% 2 5.4% 2 5.4% 2 5.4% 1 38

Our actions towards competitors can be

termed aggressively competitive 6 16.2% 11 29.7% 10 27.0% 6 16.2% 4 10.8% 1 38

We are aware and responsive to

changes that other pharmacies in our

area make

7 18.4% 24 63.2% 5 13.2% 2 5.3% 0 0.0% 0 38

Five-year plans for this pharmacy are a

high priority5 14.3% 14 40.0% 6 17.1% 8 22.9% 2 5.7% 3 38

The majority of staff in this pharmacy

have worked here for more than 5

years

8 22.2% 10 27.8% 1 2.8% 7 19.4% 10 27.8% 2 38

There has been little change in our

pharmacy over the last 10 years 2 5.7% 4 11.4% 2 5.7% 15 42.9% 12 34.3% 3 38

Management Structure of the Pharmacy

Risk Taking

Autonomy

Work Ethic


Disagree Strongly

Innovation

Proactiveness

Statement and Entrepreneurial

Orientation Parameter Addressed

Strongly agree Agree Neutral


Page 119 of 361

5.3.3 Comparison of Characteristics between PROMISe and Non-Participant Pharmacies

5.3.3.1 Areas of similarity between the pharmacies

When comparing the pharmacy demographics assessed in the PROMISe owner/manager

questionnaire with those of the non-participating pharmacies, there were a number of similarities.

There was little discernable difference in the size and location of the pharmacies. The largest

proportion of pharmacies in both cases were located in a strip of other shops, with representation of

both smaller pharmacies (<100m2) and larger pharmacies (>500m

2).

The number of full time equivalent pharmacists displayed similar patterns between the participants and

non-participants. Pharmacies commonly employed either one or two pharmacists.

In both groups of pharmacies, the staff mix within the pharmacy was similar. In the group of

pharmacies which participated in the trial, 9 of the 48 employed a graduate pharmacist. In the non-

participants, 10 of the 38 employed a graduate pharmacist.

The uptake of recognised training for pharmacy staff was around the same between the two groups.

Those with pharmacy staff who had not completed formalised Certificate II training represented 7 of 48

participants and 6 of the 38 non-participants.

5.3.3.2 Areas of Difference

When considering differences between these groups, we need to be aware that those pharmacies that

choose to return questionnaires are only a subset of all pharmacies, and there may be characteristics

within these pharmacies that make them similar to pharmacies that participated in the research

project.

The sample of non-participant pharmacies in general represented businesses with longer opening

hours, with the maximum number of hours including a pharmacy open 24hours 7 days a week.

Those pharmacies which participated in the project had typically been in the current ownership for a

shorter period of time and were less likely to be a member of a banner group.

5.3.3.2.1 QCPP adopter status

Whilst the non-participants closely reflected the national average, it can be seen that those

pharmacies which participated in the PROMISe trial had a larger number of innovators (accreditation

before 1st Dec 1999) . When ownership of the pharmacy was taken into account, the proportion of

laggards is also well below the national average (see Figure 5.3-13).


Page 120 of 361

10.4%8.3%

18.8%18.4%

34.0%

16.7%

45.8%

21.1%

44.7%

2.6%

13.2%16.0%

13.5%

34.0%

2.5%

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%


PROMISe

Non-participant

Australia

Figure 5.3-23: QCPP Adopter Status for PROMISe and Non-Participant Pharmacies

5.3.3.2.2 Entrepreneurial Orientation

Of the 14 statements, there was no difference in the pattern of responses between the PROMISe

pharmacies and non-participant pharmacies in all but one of the questions. The results for the

statement relating to innovation are shown in Figure 5.3-24. There was a mean difference of 1.5 point

score (p = 0.012), with the PROMISe pharmacies scoring higher in terms of self-perceived innovation.


Page 121 of 361

This pharmacy is known for innovation among the pharmacies in this area

.138 .712 -2.593 69 .012 -1.52741 .58896 -2.70236 -.35246

-2.601 58.429 .012 -1.52741 .58719 -2.70261 -.35221

Equal variances

assumed

Equal variances

not assumed

Q1 InnovationF Sig.

Levene's Test for

Equality of Variances

t df Sig. (2-tailed)Mean

DifferenceStd. ErrorDifference Lower Upper

95% ConfidenceInterval of the

Difference

t-test for Equality of Means

0 5 10

Q1 Innovation

0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 4.8214Std . D ev. = 2.40453N = 28

Study participant: no

2.00 4.00 6.00 8.00 10.00

Q1 Innovation

0

2

4

6

8

10

Fre

qu

en

cy

Mean = 6.3488Std. D ev. = 2.43862N = 43


0 = Strongly disagree, 10 = Strongly agree0 = Strongly disagree, 10 = Strongly agree

Figure 5.3-24: Innovation Statement For PROMISe And Non-Participant Pharmacies

The six parameters that are used to determine entrepreneurial orientation are compared between

PROMISe and non-participant pharmacies from Figure 5.3-25 to Figure 5.3-30. Note that the

respondents were able to indicate they were unsure about a particular statement, and these were not

included in the analysis for the score, and therefore the number of valid cases for the different scores

varies.

There was a significance difference between the PROMISe pharmacies and the non-participant

pharmacies in their innovativeness score. PROMISe pharmacies scored 0.87 points higher (p = 0.013)

in terms this score. There were no other statistical differences between the other parameters used to

calculate the entrepreneurial orientation score or in the overall management structure of pharmacy

score.


Page 122 of 361

Innovativeness Score

1.478 .228 -2.539 84 .013 -.87500 .34465 -1.56038 -.18962

-2.509 75.472 .014 -.87500 .34877 -1.56972 -.18028

Equal variancesassumed

Equal variancesnot assumed

Innovativeness ScoreF Sig.

Levene's Test forEquality of Variances




Difference


0 = Low Score, 10 = High Score0 = Low Score, 10 = High Score

0.00 2.00 4.00 6.00 8.00


0

2

4

6

8

10

Fre

qu

en

cy

Mean = 3.25Std . D ev. = 1.67554N = 38


0.00 1.00 2 .00 3 .00 4.00 5.00 6.00 7.00


0

3

6

9

12

15

Fre

qu

en

cy

Mean = 4.125Std . D ev. = 1.51412N = 48


Figure 5.3-25: Comparison of Innovativeness Score between PROMISe and Non-Participant Pharmacies


Page 123 of 361

Proactiveness Score

.082 .775 -1.010 84 .315 -.38925 .38545 -1.15577 .37726

-1.006 78.216 .318 -.38925 .38706 -1.15979 .38128



Proactiveness ScoreF Sig.





Difference



0 2 4 6 8 10

Proactiveness Score

0

2

4

6

8

10

Fre

qu

en

cy

Mean = 6.3816Std. D ev. = 1.81001N = 38


0 2 4 6 8 10

Proactiveness Score

0

2

4

6

8

10

Fre

qu

en

cy

Mean = 6.7708Std. D ev. = 1.74721N = 48


Figure 5.3-26: Comparison of Proactiveness Score between PROMISe and Non-Participant Pharmacies


Page 124 of 361

Risk Taking Score

2.623 .109 -1.064 84 .290 -.39474 .37089 -1.13230 .34283

-1.034 68.746 .305 -.39474 .38162 -1.15610 .36663



Risk Taking ScoreF Sig.




95% Confidence

Interval of theDifference



0 2 4 6 8 10

Risk Taking Score

0

2

4

6

8

Fre

qu

en

cy

Mean = 4.1053Std. D ev. = 1.9318N = 38


0 2 4 6 8 10

Risk Taking Score

0

2

4

6

8

10

12

14

Fre

qu

en

cy

Mean = 4.50Std. D ev. = 1.50884N = 48


Figure 5.3-27: Comparison of Risk Taking Score between PROMISe and Non-Participant Pharmacies


Page 125 of 361

Autonomy Score

1.666 .200 -.717 84 .475 -.24726 .34474 -.93282 .43830

-.691 65.174 .492 -.24726 .35771 -.96161 .46709



Autonomy ScoreF Sig.




95% Confidence

Interval of theDifference



0.00 2.00 4 .00 6.00 8.00 10.00

Autonomy Score

0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 5.0132Std . D ev. = 1.85446N = 38


0 2 4 6 8 10

Autonomy Score

0

5

10

15

20

Fre

qu

en

cy

Mean = 5.2604Std. D ev. = 1.34081N = 48


Figure 5.3-28: Comparison of Autonomy Score between PROMISe and Non-Participant Pharmacies


Page 126 of 361

Work Ethic Score

2.812 .097 1.605 84 .112 .71601 .44598 -.17086 1.60288

1.569 70.963 .121 .71601 .45645 -.19414 1.62616



Work Ethic ScoreF Sig.





Difference



0 5 10

Work Ethic Score

0

1

2

3

4

5

6

7

Fre

qu

en

cy

Mean = 2.7368Std. D ev. = 2.27423N = 38


0 .00 5.00 10.00

Work Ethic Score

0

2

4

6

8

10

12

14

Fre

qu

en

cy

Mean = 2.0208Std. D ev. = 1.86216N = 48


Figure 5.3-29: Comparison of Work Ethic Score between PROMISe and Non-Participant Pharmacies


Page 127 of 361

Competitive Aggressiveness Score

.251 .618 .255 84 .799 .06908 .27081 -.46945 .60761

.252 75.330 .802 .06908 .27415 -.47701 .61517



CompetitiveAggressiveness Score

F Sig.





Difference



0 2 4 6 8 10


0

2

4

6

8

10

Fre

qu

en

cy

Mean = 4.1316Std . D ev. = 1.3187N = 38


0 2 4 6 8 10


0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 4.0625Std. D ev. = 1.18781N = 48


Figure 5.3-30: Comparison of Competitive Aggressiveness Score between PROMISe and Non-Participant Pharmacies


Page 128 of 361

Management Structure of Pharmacy Score

1.785 .185 -.904 84 .369 -.41155 .45523 -1.31683 .49373

-.882 70.267 .381 -.41155 .46671 -1.34231 .51921



Structure ofPharmacy Score

F Sig.





Difference



0.00 2 .00 4.00 6.00 8.00 10 .00

Structure of Pharmacy Score

0

2

4

6

8

Fre

qu

en

cy

Mean = 5.6579Std. D ev. = 2.33717N = 38


2.00 4.00 6.00 8.00 10.00

Structure of Pharmacy Score

0

2

4

6

8

10

Fre

qu

en

cy

Mean = 6.0694Std. D ev. = 1.88557N = 48


Figure 5.3-31: Comparison of Management Structure Score between PROMISe and Non-Participant Pharmacies

Using the formula previously outlined it was possible to calculate the entrepreneurial orientation score

of both the PROMISe pharmacies and those who responded to the survey. The results of these

calculations, and a comparison between these two groups is shown in Figure 5.3-32 and Figure

5.3-33. There was no statistical difference between the entrepreneurial orientation score of the two

groups.


Page 129 of 361


.615 .435 -1.376 84 .172 -1.03230 .75005 -2.52386 .45927

-1.343 70.476 .184 -1.03230 .76858 -2.56500 .50040



EntrepreneurialOrientation Score

F Sig.





Difference


0 5 10 15 20 25 30


0

3

6

9

12

15

Fre

qu

en

cy

Mean = 20 .117Std. D ev. = 3.84305N = 38


0 5 10 15 20 25 30


0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 21 .1493Std. D ev. = 3.11424N = 48


Figure 5.3-32: Comparison of Entrepreneurial Orientation Score between PROMISe and Non-Participant Pharmacies (Histogram)

no yes

Study participant

10.00

15.00

20.00

25.00

En

tre

pre

neu

ria

l O

rie

nta

tio

n S

co

re

34

Figure 5.3-33: Comparison of Entrepreneurial Orientation Score between PROMISe and Non-Participant Pharmacies (Boxplot)


Page 130 of 361

Based on the possible responses to the questions and the loading developed around Entrepreneurial

Orientation, the minimum score would be 0, a completely neutral score would be 22.7 and the

maximum score would be 45.4. Both groups of pharmacies had a neutral mean score and there were

pharmacies above and below the mean (as expected, see Figure 5.3-34).

The individual scores for the PROMISe participants will be reviewed later in relation to the pharmacy’s

clinical intervention rate (see section 5.8.13.1).

01

5

10

26

29

13

2

0 0 0 0 00

5

10

15

20

25

30

35

9 12 15 18 21 24 27 30 33 36 39 42 45

Ma

xim

um

po

ss

ible

va

lue

45

.4

Ne

utra

l va

lue

22

.7

Min

imu

m p

os

sib

le v

alu

e 0

Figure 5.3-34: Distribution of EO score with maximum, minimum and average value displayed

5.4 Pharmacist Demographics

Of the 148 pharmacists initially enrolled in the study, 125 (84.5%) completed the online classification

system training either before or within the first 5 days of data collection.

The online training included a questionnaire which collected information on aspects of the

pharmacist’s demographics, educational background, work experience, and attitudes to pharmacy.

5.4.1 Age, Year of Graduation and Gender

Of the 125 pharmacists who completed the online training for the PROMISe project, 70 (56%) were

male and the majority (60.8%) were between the ages of 20 and 40 years (see Table 5.4-1).


Page 131 of 361

In the National Pharmacy Database Project conducted in 2002, 31.7% of pharmacists were urban

females, while 32% of the population were over 50 years old.10

The National Health Labour Force

series on Pharmacists, published in 2001, by the Australian Institute of Health and Welfare (AIHW)

found that in 1999, 46.9% of employed pharmacists were females. The proportion of females in the

pharmacist labour force had risen steadily, from 43.5% in 1994 to 46.9% in 1999. The labour force

grew by 12.5% over this 5-year period, but the number of male pharmacists rose by 5.8% and females

by 21.3%.11

Thus, the participants in the online training, and hence the trial, showed a greater

proportion of males than the national average.

#% within

Gender

% within

Age

Group

#% within

Gender

% within

Age

Group

#% within

Total

Under 20 years old 0 0.0% 0.0% 4 5.7% 100.0% 4 3.2%

20 to 30 years old 21 38.2% 44.7% 26 37.1% 55.3% 47 37.6%

31 to 40 years old 15 27.3% 51.7% 14 20.0% 48.3% 29 23.2%

41 to 50 years old 12 21.8% 54.5% 10 14.3% 45.5% 22 17.6%

51 to 60 years old 5 9.1% 29.4% 12 17.1% 70.6% 17 13.6%

Over 60 years old 2 3.6% 33.3% 4 5.7% 66.7% 6 4.8%

Total 55 100.0% 44.0% 70 100.0% 56.0% 125 100.0%

Age Group

Female Male Total

Table 5.4-1: Age And Gender Of PROMISe Pharmacists

The AIHW survey also showed that the pharmacy labour force has aged slightly, from an average of

45.1 years in 1994 to 45.5 years in 1996 and 46.1 years in 1999. There was a considerable age

difference between the sexes. The average age of females in 1999 was 41.7 years, substantially

younger than the average age of males (50.0 years). Over 60% of females were aged less than 45

years compared with only 34.7% of males, while 44.1% of males were aged 55 or more compared with

only 16.1% of females. Thus, the PROMISe cohort of males is significantly younger than the National

average, with only 22.8% being over 50 years old.

In Figure 5.4-1, the year of graduation of the PROMISe pharmacists is shown related to the gender.

As can be seen, a large number of the PROMISe pharmacists were recent graduates (less than five

years), and almost half of these were male.

10 Berbatis CG, Sunderland VB, Mills CR, Bulsara M. National pharmacy database project. 2002.

11 AIHW (Australian Institute of health and welfare) 2001. Pharmacy labour force to 2001: Canberra,

AIHW (National Health Labour Force Series No. 25)


Page 132 of 361

before 1975 1975 to 1979 1980 to 1984 1985 to 1989 1990 to 1994 1995 to 1999 2000 to 2005

Group of Graduation Year

0

10

20

30

40

Co

un

t

gender

f

m

Figure 5.4-1: Graduation Year Group And Gender Of PROMISe Pharmacists

5.4.2 Continuing Education and Qualifications

The majority of pharmacists who participated in the PROMISe study undertook 10 to 25 hours per year

of continuing education activities. Three of the respondents had achieved higher qualifications in

pharmacy (one PhD and two Fellows of the Society of Hospital Pharmacists of Australia).

Seventeen of 122 pharmacists (13.9%) indicated that they were accredited to conduct medication

reviews. This is higher than the average national rate of accredited pharmacists, possibly indicating

that the cohort involved in the PROMISe project includes a more professionally active group of

pharmacists.


Page 133 of 361

No CE at all Less than 10 hours per year

10 to 25 hours per year

26 to 50 hours per year

Over 50 hours per year

Don't Know

ce

0

10

20

30

40

50

60

70

Co

un

t

Figure 5.4-2: Number Of Hours Per Year Of Continuing Education For PROMISe Pharmacists

5.4.3 Practice Profile

Pharmacists who participated in the PROMISe study had a past history of predominantly community

practice experience (see Table 5.4-2). This question requested the number of full-time years

experience, so that pharmacists who worked part-time may have done so for many years, although the

number of full-time years would be lower than this.

# %

Community Pharmacy 22 23 76 121 63.4%

Hospital Pharmacy 27 11 10 48 25.1%

Medication Reviews 10 4 1 15 7.9%

Other 3 1 3 7 3.7%

Total 62 39 90 191 100.0%

TotalPractice Setting

Less than 2

years2-5 years over 5 years

Table 5.4-2: PROMISe Pharmacists’ Past Years In Different Areas Of Pharmacy Practice

Consequently 45 of 121 (37.2%) pharmacists indicated that they had five years or less experience in

community pharmacy, although many of these had a longer duration of experience (see below).

As expected, 116 of the 125 pharmacists (92.8%) indicated they spent the majority of their current

work-time in community practice settings.

Twenty four of 124 (19.4%) pharmacists had worked in pharmacy internationally, and 39 of 123

(31.7%) had worked in rural or remote areas of Australia.


Page 134 of 361

5.4.4 Role and Duration of Employment of Community Pharmacists

Pharmacists who participated in the PROMISe trial were predominantly employee pharmacists

(58.8%) or owners (34.5%) (see Table 5.4-3). Male pharmacists were more likely to be owners and

female pharmacists were more likely to be employee pharmacists (see Figure 5.4-3).

Role Number Percent

employee 70 58.8

locum 8 6.7

owner 41 34.5

Total 119 100.0

Table 5.4-3: Role Of PROMISe Pharmacists

f m

gender

0

10

20

30

40

50

60

70

Co

un

t

cp_role

employee

locum

owner

Figure 5.4-3: Gender Of PROMISe Pharmacists By Role

The majority of pharmacists who participated in the PROMISe project (59.2%) had been in the position

that they indicated for 10 years or more (see Table 5.4-4). More than half of these longer-serving

pharmacists were owners (see Figure 5.4-4).

8 6.4

4 3.2

39 31.2

29 23.2

45 36.0

125 100.0

2 years or less

2 to 5 years

5 to 10 years

10 to 20 years

20 years or more

Total

Frequency Percent

Table 5.4-4: PROMISe Pharmacists’ Years In Current Position


Page 135 of 361

2 years or less 2 to 5 years 5 to 10 years 10 to 20 years 20 years or more

Years in Current Position

0

10

20

30

40

50

Co

un

t

cp_role

employee

locum

owner

Figure 5.4-4: Duration Of Employment Of PROMISe Pharmacists In Current Role

The majority of pharmacists who participated in the PROMISe project worked 40 hours per week or

less in the pharmacy (see Figure 5.4-5). Pharmacists who worked 40 hours or more in the pharmacy

were more likely to be owners (see Figure 5.4-5) and males (see Figure 5.4-6).

Less than 10 hours/week 10 to 20 hours/week 20 to 40 hours/week Over 40 hours/week

cp_hours

0

10

20

30

40

50

Co

un

t

cp_role

employee

locum

owner

Figure 5.4-5: Hours Of Weekly Work Of PROMISe Pharmacists By Role


Page 136 of 361

Less than 10 hours/week

10 to 20 hours/week 20 to 40 hours/week Over 40 hours/week

cp_hours

0

10

20

30

40

50

Co

un

t

gender

f

m

Figure 5.4-6: Hours Of Work By Gender Of PROMISe Pharmacists

5.4.5 Self-Reported Workload of Community Pharmacists

Pharmacists who participated in the PROMISe study usually either worked alone or with one other

pharmacist (86.5%) (see Table 5.4-5). Sixty-one of the 122 pharmacists (50%) indicated that they

spend more than 50% of their time doing administrative tasks (see Table 5.4-6).

38 31.9

65 54.6

16 13.4

119 100.0

Sole Pharmacist

One other Pharmacist

2 to 4 other Pharmacists

Total

ValidFrequency Percent

Table 5.4-5: Number Of Pharmacists Working Together In PROMISe Pharmacies


Page 137 of 361

3 2.5

58 47.5

56 45.9

5 4.1

122 100.0

Less than 10%

10 to 50%

51 to 75%

Over 75%

Total

ValidFrequency Percent

Table 5.4-6: Proportion Of Time PROMISe Pharmacists Spend Doing Administrative Tasks

As expected, those pharmacists who spent more than 50% of their time doing administrative tasks

were owners (see Figure 5.4-7).

The self-reported estimate of number of prescriptions dispensed per pharmacist per day was very

high, with 88.6% of respondents indicating that they dispensed 150 or more prescriptions daily

(including 25.2% who dispensed over 250 prescriptions per day) (see Table 5.4-7). The majority of

pharmacists in this “high dispensing” group were employee pharmacists (see Figure 5.4-8).

10 to 50% 51 to 75% Over 75%

cp_admin

0

10

20

30

40

50

60

Co

un

t

cp_role

employee

locum

owner

Figure 5.4-7: Proportion Of Time PROMISe Pharmacists Spend Doing Administrative Tasks By Role In The Pharmacy

3 2.4

6 4.9

78 63.4

31 25.2

5 4.1

123 100.0

Less than 50 prescriptions

50 to 150 prescriptions


Over 250 prescriptions

Not applicable to role

Total

Frequency Percent

Table 5.4-7: Self-Reported Number Of Prescriptions Dispensed Daily By PROMISe Pharmacists


Page 138 of 361



Over 250 prescriptions

Not appropriate

0

20

40

60

80

Co

un

t

cp_role

employee

locum

owner

Figure 5.4-8: Self-Reported Number Of Prescriptions Dispensed Daily By Role In Pharmacy

5.4.6 Clinical Skill Assessment

Two interactive clinical scenarios were prepared to assess clinical problem solving skills. There were

some technical problems with the first scenario, and the results for this scenario will not be considered

in this report.

A clinical skills score was calculated for each pharmacist. This was based on a composite value from

the three component scores:

• The problem identification score

• The investigative steps score

• The appropriateness of recommendations score

The second clinical skills scenario (see Appendix 13) required the detection and resolution of three

major drug-related problems. The number of problems that the pharmacists detected determined their

problem identification score. There were several routes to the detection of each of the problems, and

the number of steps taken by each pharmacist to identify the problems (investigative steps) was

automatically recorded. Upon identifying a problem, the pharmacist was asked to select the most

appropriate recommendation for its resolution. Each selection was scored separately, with some being

more appropriate than others, and some being totally inappropriate. The selections that the

pharmacists made determined their appropriateness of recommendation score.


Page 139 of 361

Of the 125 pharmacists who completed the on-line training, 13 had technical problems with their

assessments, and exited from the case before scores could be registered against the case.

The majority of pharmacists identified all three potential drug-related problems in the scenario (see

Table 5.4-8), with only one pharmacist not identifying any drug-related problems in the scenario.

# %0 1 0.9%

1 30 26.8%

2 19 17.0%

3 62 55.4%

Total 112 100.0%

FrequencyNumber of

Problems

Identified

Table 5.4-8: Number Of Problems Identified In PROMISe Clinical Skills Assessment

The number of web pages the pharmacist visited in detecting the drug-related problems was recorded

for each user. As would be anticipated, the number of pages visited (that is the number of questions

asked or investigations made) increased with the number of problems detected (see Figure 5.4-9).

There were, however, a number of pharmacists who were able to identify all three drug-related

problems with a minimal amount of investigation (see purple shading in Figure 5.4-10).

0 1 2 3

Number of Problems Identified

0

5

10

15

20

25

30

35

40

45

Nu

mb

er

of

Investi

gati

on

Pa

ge

s

Figure 5.4-9: Number Of Investigative Pages Visited And Number Of Problems Identified In PROMISe Clinical Skill Assessment


Page 140 of 361

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 30 33 37 40 42

Number of Investigation Pages

0

2

4

6

8

10C

ou

nt

Number of Problems Identified

0

1

2

3

Figure 5.4-10: Problems Identified By Number Of Investigation Pages Visited In PROMISe Clinical Skills Assessment

Our premise is that pharmacists who are able to identify problems with less investigative steps are

more likely to be more highly clinically skilled. We were able to calculate the ratio of number of

problems identified to the number of investigational steps for each pharmacist (number of problems

per 10 investigational steps; see Figure 5.4-11).


Page 141 of 361

0.00 1.00 2.00 3.00 4.00 5.00 6.00

Problems Identified per 10 pages visited

0

10

20

30

40

Fre

qu

en

cy

Mean = 1.5291Std. Dev. = 1.03556N = 112

Figure 5.4-11: Ratio Of Problems Identified By Number Of Investigative Steps Taken In PROMISe Clinical Skills Assessment

The clinical skills scenario required that once a problem was identified, a recommendation was made.

There were five options for recommendations for each of the three different drug-related problems.

Each of these recommendations were scored for appropriateness (out of either 20 or 15). The scores

for each of the three problems, and the combined score for all problems identified by the participant,

are shown in Figure 5.4-12. The maximum score possible for appropriateness of recommendation

was 55. The total recommendation score could easily be converted to a percentage.


Page 142 of 361

0 5 10 15 20

Problem 2 Recommendation Score

0

20

40

60

80

Fre

qu

en

cy

Mean = 14.87Std. D ev. = 8.15N = 112

0 5 10 15

Problem 3 Recommendation Score

0

20

40

60

80

Fre

qu

en

cy

Mean = 5.8Std. D ev. = 6.629N = 112

0.00 10.00 20.00 30.00 40.00 50.00

Total Recommendation Score

0

5

10

15

20

25

Fre

qu

en

cy

Mean = 33.6161Std. Dev. = 17.31023N = 112

Figure 5.4-12: Individual Problem And Total Appropriateness Of Recommendation Scores For Participants In PROMISe Clinical Skills Assessment

The product of the total recommendation score (as a percentage) and ratio of the number of clinical

problems to investigation steps was termed a clinical skills score. This score increases with number of

problems identified, increases with the appropriateness of recommendations and increases with fewer

investigative steps. The distribution of the scores within our sample follows an expected pattern for

clinical skills (see Figure 5.4-13).


Page 143 of 361

0.00 100.00 200.00 300.00 400.00 500.00 600.00

Clinical Skill Score

0

10

20

30

40

Fre

qu

en

cy

Mean = 106.6545Std. Dev. = 107.07734N = 112

Figure 5.4-13: Clinical Skill Score For PROMISe Participants

5.4.7 Personal Views of Pharmacy

A series of eight questions were asked to determine overall personal satisfaction and professional

opinions of pharmacy.

1. I am comfortable with the technology changes taking place in pharmacy

2. I see the traditional sources of pharmacy remuneration as the only source of funds for the future

3. For me pharmacy is just a job

4. I believe that pharmacists currently have too much to do and there is no capacity for change

5. I generally enjoy my work

6. A pharmacist's most important job is dispensing

7. I am comfortable selling a product that may be of questionable therapeutic value as long as it is

not harmful

8. I am able to perform my expected duties without interruption

Each of these questions was scored on a scale of 1 to 7, with 1 being strongly agree, 4 being neutral

and 7 being strongly disagree (see scale on each histogram). Histograms, mean values and number of

respondents scoring each question are shown from Figure 5.4-14 to Figure 5.4-21 below.


Page 144 of 361

1 2 3 4 5 6 7

I am comfortable with the technology changes taking place in pharmacy

0

10

20

30

40

50

Fre

qu

en

cy

Mean = 2.0252Std. Dev. = 1 .00391N = 119

Figure 5.4-14: PROMISe Pharmacists’ Attitude Towards Technology Change

1.00 2.00 3.00 4.00 5.00 6.00 7.00

I see the traditional sources of pharmacy remuneration as the only source of funds for the future

0

5

10

15

20

25

30

Fre

qu

en

cy

Mean = 5 .1927Std. Dev. = 1 .46235N = 109

Figure 5.4-15: PROMISe Pharmacists’ Attitude Towards Traditional Sources Of Remuneration


Page 145 of 361

1 .00 2 .0 0 3.0 0 4.0 0 5.0 0 6 .0 0 7.00

F or m e p harm ac y is jus t a job

0

10

20

30

40

50

Fre

qu

en

cy

M ea n = 5 .6 2 81S td. D e v . = 1 .4 4 99N = 1 21

Figure 5.4-16: PROMISe Pharmacists’ Attitude Towards Pharmacy As A “Job”

1.00 2.00 3.00 4.00 5.00 6.00 7.00

I believe that pharmacists currently have too much to do and there is no capacity for change

0

5

10

15

20

25

30

Fre

qu

en

cy

M ean = 4 .7627Std. Dev. = 1.62606N = 118

Figure 5.4-17: PROMISe Pharmacists’ Attitude Towards Capacity For Change


Page 146 of 361

1 2 3 4 5 6 7

I generally enjoy my work

0

10

20

30

40

50

60

70

Fre

qu

en

cy

M ean = 1.6421Std. Dev. = 1.04082N = 95

Figure 5.4-18: PROMISe Pharmacists’ Attitude Towards Enjoyment Of Their Work

1 2 3 4 5 6 7

A pharmacist's most important job is dispensing

0

10

20

30

40

Fre

qu

en

cy

M ean = 3.9244

Std. Dev. = 1.46229N = 119

Figure 5.4-19: PROMISe Pharmacists’ Attitude Towards Dispensing As A Primary Role


Page 147 of 361

1.00 2.00 3.00 4.00 5.00 6.00 7.00

I am comfortable selling a product that may be of questionable therapeutic value as long as it is not harmful

0

10

20

30

40

50

Fre

qu

en

cy

M ean = 5.3898Std. Dev. = 1.35895N = 118

Figure 5.4-20: PROMISe Pharmacists’ Attitude Towards Selling Items Of Questionable Therapeutic Value

1.00 2.0 0 3 .00 4 .00 5 .00 6 .00 7.0 0

I am able to perform my expected duties without interruption

0

10

20

30

40

Fre

qu

en

cy

Me an = 5.0331Std . D ev. = 1 .57 551N = 121

Figure 5.4-21: PROMISe Pharmacists’ Opinion Of Interruption


Page 148 of 361

Questions 1 (Figure 5.4-14), 3 (Figure 5.4-15) and 4 (Figure 5.4-17) were all related to current or

potential change and capacity for change. Generally, there were positive attitudes to change, with

pharmacists indicating that they were comfortable with the technological changes taking place (mean

value 2.0, agree with the statement), and that traditional sources of remuneration are not the only

source of income for the future (mean value 5.2, disagree with the statement). There was slight

disagreement with the statement that there was no capacity for change by pharmacists because they

had too much to do (see Figure 5.4-17).

By appropriate reversal and combination of the scores for these three questions, an overall “change

preparedness” score can be constructed. As can be seen in Figure 5.4-22, pharmacists who

participated in the PROMISe study, generally had a high preparedness for change (mean score 5.14).

0 2 4 6

Change Preparadness Score

0

10

20

30

40

50

Fre

qu

en

cy

Mean = 5.1429Std. Dev. = 1.13709N = 119

Figure 5.4-22: Change Preparedness Score For PROMISe Participants

Questions 2 and 5 relate to “job satisfaction”. PROMISe pharmacists generally disagreed that

pharmacy was “just a job” (mean score 5.63, see Figure 5.4-16) and generally enjoyed their work

(mean score 1.64, see Figure 5.4-18). Of interest, only 95 respondents answered this question,

compared to 119 to 123 respondents for other questions in this section. This may indicate a reticence

to commit to an answer to this question.

By reversing the scores for general enjoyment of their work (question 5), a “job satisfaction” construct

can be determined. As can be seen in Figure 5.4-23, PROMISe pharmacists had a relatively high job

satisfaction score as determined by this method (mean score 5.85).


Page 149 of 361

Questions 6 and 7 relate to the “professional integrity” of the pharmacist. The first of these questions

states that “a pharmacist’s most important job is dispensing”. The overall response to this statement

was neutral (mean score 3.92), with approximately one third of the respondents agreeing, one third

disagreeing and one third neutral. The second of these questions concerns the pharmacist’s comfort in

selling a product of questionable therapeutic efficacy. There was a general response that indicated

that these pharmacists would not be comfortable selling such a product (mean score 5.03).

1 2 3 4 5 6 7

Job Satisfaction Score

0

10

20

30

40

Fre

qu

en

cy

Mean = 5.8474Std. Dev. = 1.15794N = 95

Figure 5.4-23: Job Satisfaction Score For PROMISe Participants

5.5 DOCUMENT Classification System Training

5.5.1 Pharmacists’ Competency with the DOCUMENT Classification System

Of the 148 pharmacists initially enrolled in the study, 125 (84.5%) completed the online classification

system exercises either before or within the first 5 days of data collection. Of those pharmacists who

participated in the trial, 86 completed the online sample scenarios for a second time at the completion

of data recording.

The time taken by each pharmacist to complete the categorisation of the scenarios before and after

the project is shown in Figure 5.5-1 and Table 5.5-1. It was possible for pharmacists to complete one

or a few scenarios and then leave the session and come back to it at another time. The figure

demonstrates that a number of pharmacists who took greater than 48 hours to complete the scenarios.

These pharmacists obviously completed their assessment in multiple sittings. However, of those

pharmacists who completed the initial training within one day, 43 (34%) did so in two hours or less.

This compares favourably with the number of pharmacists who were able to complete the scenarios in

less than two hours after the project (48; 56%).


Page 150 of 361

Figure 5.5-1: Time To Complete The Online Scenario Classifications Before and After the Project

Hours to complete training

Pre-trial online

training

Post-trial online

training

1 7.2% 32.6%

2 27.2% 23.3%

3 6.4% 11.6%

6 6.4% 4.7%

9 3.2% 3.5%

12 0.8% 0.0%

24 8.8% 4.7%

36 7.2% 5.8%

48 4.8% 3.5%

More 28.0% 10.5%

Table 5.5-1: Time To Complete Training Pre-Trial And Post-Trial

In the following sections, we review the appropriateness of the responses for the scenarios before and

after the project. These sections consider the 20 scenarios either together or as a series. Information

concerning these parameters for each individual scenario are included in Appendix 23.

5.5.1.1 Categories and subcategories recorded for the 20 scenarios

For each scenario the pharmacist considered the situation and assigned a category and subcategory

to that situation. After each scenario, the pharmacists were provided with information on the most

appropriate category with explanatory notes where applicable. Therefore, even allowing for variation

in the ease of selecting a category for different scenarios, it can be seen ( see Table 5.5-2 and Figure

5.5-2) that as the pharmacists progressed through scenarios during their initial training (pre-trial), their

correct selection of category improved gradually.

9

34

8 8

4

1

119

6

35

0

5

10

15

20

25

30

35

40

1 2 3 6 9 12 24 36 48 More

Hours to complete training pre-trial

28

20

10

43

0

45

3

9

0

5

10

15

20

25

30

35

40

1 2 3 6 9 12 24 36 48 More

Hours to complete training post-trial


Page 151 of 361

Also displayed in Table 5.5-2 is the proportion of pharmacists who selected the correct subcategory of

intervention. Where the appropriate category was selected, more often than not the pharmacist also

selected the most appropriate subcategory.

Number % Number % Number % Number %

1 58 45.7 38 29.9 59 65.6 52 57.8

2 110 88 109 87.2 87 97.8 87 97.8

3 104 82.5 103 81.7 84 94.4 84 94.4

4 45 36 45 36 62 71.3 61 70.1

5 99 79.8 99 79.8 76 85.4 76 85.4

6 106 87.6 104 86 83 93.3 82 92.1

7 84 67.2 56 44.8 68 77.3 53 60.2

8 119 96 119 96 88 100 88 100

9 101 80.8 99 79.2 80 92 79 90.8

10 110 88 107 85.6 82 94.3 81 93.1

11 85 68 84 67.2 54 62.1 54 62.1

12 102 81.6 99 79.2 70 80.5 65 74.7

13 78 62.4 27 21.6 70 80.5 38 43.7

14 108 86.4 68 54.4 78 89.7 53 60.9

15 52 41.6 49 39.2 39 44.8 36 41.4

16 111 88.8 97 77.6 80 93 74 86

17 86 68.8 62 68.8 65 75.6 43 50

18 102 82.3 86 69.4 67 77.9 64 74.4

19 121 96.8 121 96.8 86 100 86 100

20 114 91.2 103 82.4 81 94.2 74 86

Average 76.0 68.1 83.5 76.0

Correct CategoryCorrect

Subcategory

After Use of Classification System in

Trial

ScenarioCorrect Category

Correct

Subcategory

Initial Training

Table 5.5-2: Number And Percent Of Correct Category Classifications After Initial Training And After Trial Participation


Page 152 of 361

0

10

20

30

40

50

60

70

80

90

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Scenario

PreTrialPost TrialPoly. (PreTrial)Poly. (Post Trial)

Figure 5.5-2: Correct Category of Intervention Selected Before and After PROMISe Project

In Table 5.5-2 the assignment of category is displayed for before and after the trial. Each pharmacist

who participated in the trial was encouraged to complete the online training at the end of the recording

period. There were 86 pharmacists who completed the training a second time. Once the pharmacists

had had experience with the DOCUMENT classifications, the average rate of assigning the correct

category increased from 76% to 83.5%, with similar increases in the assignment of the most

appropriate sub-category (68.1% to 76.0%).

In Figure 5.5-3 the change in allocation of the appropriate category from pre-trial to post-trial is

displayed. It can be seen that for the majority (11 of the 20) of scenarios there was at least a 10%

increase in accuracy.


Page 153 of 361

1

2

6

5

3

2

1

0

1

2

3

4

5

6

7

-5 0 5 10 15 20 More

Percent change in category pre-trial to post-trial

Figure 5.5-3: The Change In Correct Allocation Of Category From Pre-Trial To Post-Trial

5.5.1.2 Actions recorded for the 20 scenarios

The online training for the DOCUMENT categorisation system involved assigning the actions to the

scenario as described. For example:

Scenario; a prescription is presented for a 12 year old boy for amoxycillin 250mg/5mL, 4mL three

times a day for acute otitis media. You check the dose in the product information* and find that it is

meant to be 500mg three times a day. You discuss the situation with the boy's mother*** and contact

the prescriber**; you suggest that the dose is increased.

The appropriate actions to record would be:

1) Investigate written material (see *)

2) Contact prescriber (see **)

3) Discussion with patient or carer (see ***)

These actions are all that need to be recorded for this section.


Page 154 of 361

In Table 5.5-3 the assignment of the appropriate action was on average 80.3% for those who had had

limited exposure to the DOCUMENT system. In this pre-trial group the proportion of responses

considered as an incorrect action was 19.7%. Where pharmacists recorded incorrect actions it was

commonly the pharmacist attempting to indicate what ‘they would have done’ in the situation. The

percentage of correct responses increased slightly from 80.3% to 82.8% once the pharmacists had

had some exposure to the DOCUMENT system through the trial.

# % # %

Number of Correct Actions

Identified3919 80.3 2859 82.8

Number of Incorrect Actions

Identified962 19.7 592 17.2

Total 4881 3451

Correct or Incorrect

Action

Pre Trial Post Trial

Table 5.5-3: Number And Percent Correct For The Actions Recorded In Both The Pre-Trial And Post-Trial Online Training.

5.5.1.3 Recommendations

When recording the recommendations within the scenario the pharmacist was instructed to record

those recommendations as specified in the scenario. As with the recording of the actions there was a

proportion of respondents who included the recommendations which they would have carried out.

In Table 5.5-4 the frequency of correct allocation of the recommendations both pre-trial and post-trial

can be seen.


Page 155 of 361

Scenario number correct % correct number correct % correct

1 240 64.2% 184 85.2%

2 167 61.2% 98 74.8%

3 293 82.1% 211 90.2%

4 266 83.6% 205 92.3%

5 190 64.8% 133 75.1%6 92 32.6% 60 36.8%

7 173 53.2% 122 62.6%

8 83 63.4% 79 90.8%

9 212 83.8% 144 90.6%

10 112 56.9% 79 69.9%11 165 49.5% 126 58.3%

12 219 63.1% 151 68.6%

13 120 48.4% 85 54.8%

14 322 87.5% 224 91.8%

15 177 74.1% 142 91.6%16 220 69.8% 155 71.8%

17 232 65.7% 170 69.4%

18 214 86.6% 161 91.5%

19 123 69.5% 84 82.4%

20 335 80.5% 230 84.2%

Average 67.1% 76.6%

Post-trial responses Pre-trial responses

Table 5.5-4: Allocation of recommendation pre-trial and post-trial

The difference in appropriateness of recommendation selection between the post-trial to pre-trial is

positive for each scenario. This indicates that the pharmacists became familiar with the

recommendations and were able to record the most appropriate recommendation more frequently.

The increase in the improvement of recommendation was as much as 27% in one particular scenario.

The improvement in allocation of the recommendations can be seen in Figure 5.5-4.


Page 156 of 361

0

6

7

4

1

2

0

1

2

3

4

5

6

7

8

0.0% 5.0% 10.0% 15.0% 20.0% More

Figure 5.5-4: The Improvement In Recommendation Allocation (Post-Trial Compared With Pre-Trial)

5.5.1.4 Clinical Significance

For each scenario the pharmacist was asked to assign the relevant significance to the situation. The

percentage of correct responses can be seen in detail in Table 5.5-6 and is summarised in Table

5.5-6.


Page 157 of 361

# % # % # % # % # % # % # % # % # % # %

1 1 1.1 6 4.8 5 5.6 67 53.6 74 82.2 52 41.6 10 11.1

2 13 10.4 7 7.8 72 57.6 65 72.2 40 32 17 18.9

3 2 1.6 28 22.4 17 18.9 95 76 72 80.0

4 9 7.2 8 8.9 113 90.4 75 83.3 3 2.4 6 6.7

5 3 2.4 2 2.2 85 68 51 56.7 33 26.4 34 37.8 4 3.2 2 2.2

6 11 8.8 8 8.9 96 76.8 76 84.4 18 14.4 5 5.6

7 1 0.8 1 1.1 69 55.2 41 45.6 52 41.6 45 50.0 3 2.4 1 1.1

8 113 90.4 82 91.1 4 3.2 6 6.7 5 4 2 1.6 1 0.8

9 2 1.6 35 28 19 21.1 81 64.8 65 72.2 7 5.6 3 3.3

10 4 3.2 1 1.1 1 0.8 10 8 78 62.4 70 77.8 32 25.6 16 17.8

11 2 1.6 42 33.6 23 25.6 81 64.8 64 71.1

12 1 0.8 1 0.8 2 2.2 52 41.6 41 45.6 71 56.8 44 48.9

13 25 20 7 7.8 83 66.4 75 83.3 17 13.6 5 5.6

14 3 2.4 1 1.1 46 36.8 27 30.0 76 60.8 59 65.6

15 1 0.8 1 1.1 2 1.6 1 1.1 98 78.4 71 78.9 24 19.2 14 15.6

16 3 2.4 64 51.2 40 44.4 54 43.2 43 47.8 4 3.2 3 3.3

17 1 0.8 59 47.2 34 37.8 49 39.2 46 51.1 16 12.8 6 6.7

18 2 1.6 1 1.1 43 34.4 18 20.0 79 63.2 67 74.4

19 1 0.8 1 1.1 70 56 31 34.4 54 43.2 54 60.0

20 26 20.8 14 15.6 89 71.2 64 71.1 10 8 8 8.9

Pre Trial Post Trial Pre Trial Post Trial Post TrialPre Trial Post Trial Pre Trial Post Trial

Nil

Scenario

Number

SevereModerateMildLow

Pre Trial

Table 5.5-5: Assignment of significance in the 20 scenarios (shaded = most appropriate response)


Page 158 of 361

In some instances, the difference in assigning the appropriate significance decreased when comparing

post-trial to pre-trial results. Therefore, in a histogram of the difference between post- and pre-trial

results (see Figure 5.5-5) the change includes a negative change for three of the scenarios.

There is an obvious subjective nature to the assignment of a unilateral significance category to a

scenario where there are a number of unknown comorbidities and other factors. For this reason

significance has been used only as a guide to the importance of the situation.

Pre Trial Post Trial Scenario

% Correct % Correct

1 53.6 82.2

2 57.6 72.2

3 76 80.0

4 90.4 83.3

5 94.4 94.4

6 76.8 84.4

7 55.2 45.6

8 90.4 91.1

9 64.8 72.2

10 88 95.6

11 64.8 71.1

12 98.4 94.4

13 66.4 83.3

14 60.8 65.6

15 78.4 78.9

16 94.4 92.2

17 47.2 37.8

18 63.2 74.4

19 43.2 60.0

20 71.2 71.1

Table 5.5-6: Percent Correct Responses To Clinical Significance Of The Scenarios


Page 159 of 361

0

3

4 4 4

2 2

1

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

-10 -5 0 5 10 15 20 More

Percent correct for significance post vs pre-trial

No

. cases

Figure 5.5-5: The Difference In Responses To Clinical Significance Post-Trial Vs Pre-Trial

5.5.1.5 Proactive or Reactive Situations

The pharmacists involved in the online training were asked to assign ‘who initially identified the

problem?’. Whether the intervention was proactive or not was determined from the possible responses

as to who identified the problem. If the pharmacist identified the problem. It was usually proactive and

if another person identified the problem, it was usually reactive (see section 4.1.1.6). The assignment

of proactiveness in the test scenarios was intended to familiarise the pharmacist with what sorts of

situations are proactive or reactive.

There are some instances and categories of intervention where nominating ‘who identified the

problem’ does not correlate with proactive or reactive. For example scenario 8 where the problem is of

a non-clinical nature (an out of date prescription) this situation is of a reactive nature even though it is

the pharmacist who identifies the problem. This explains the low results seen for scenario 8 (see

Table 5.5-7) which are actually not poor results but an example of where interpretation is required to

determine whether the intervention is of a proactive or reactive nature.


Page 160 of 361

# % # % # % # %

1 Proactive 108 86.4 77 85.6 17 13.6 13 14.4

2 Proactive 123 99.2 89 100.0 1 0.8 0.0

3 Reactive 59 47.2 21 23.6 66 52.8 68 76.4

4 Proactive 44 35.2 30 33.7 81 64.8 59 66.3

5 Proactive 111 88.8 83 93.3 14 11.2 6 6.7

6 Proactive 124 99.2 89 100.0 1 0.8 0.0

7 Reactive 45 36.0 16 18.2 80 64.0 72 81.8

8 Reactive 120 100.0 81 97.6 2 2.4

9 Reactive 30 24.0 14 16.1 95 76.0 73 83.9

10 Proactive 125 100.0 86 98.9 1 1.1

11 Proactive 122 97.6 84 96.6 3 2.4 3 3.4

12 Proactive 115 92.7 82 94.3 9 7.3 5 5.7

13 Proactive 111 88.8 78 89.7 14 11.2 9 10.3

14 Proactive 95 76.0 68 78.2 30 24.0 19 21.8

15 Reactive 52 41.6 37 42.5 73 58.4 50 57.5

16 Proactive 79 63.2 65 75.6 46 36.8 21 24.4

17 Proactive 122 97.6 85 98.8 3 2.4 1 1.2

18 Proactive 125 100.0 86 100.0 0.0

19 Proactive 125 100.0 84 97.7 2 2.3

20 Proactive 125 100.0 85 98.8 1 1.2

Scenario

NumberPre Trial Post Trial

Proactive Reactive

Pre Trial Post TrialCorrect

Category

(shading = appropriate response)

Table 5.5-7: Allocation Of ‘Who Identified The Problem’ For Pre-Trial And Post-Trial


Page 161 of 361

5.6 Issues with Use of Intervention Recording Software

The software used for the recording of interventions had been modified significantly from that used in

the pilot study. As a result, there were a number of issues that required attention during the early part

of the data collection period.

As a result of the tight development timeline, it was not possible to fully test the modified interface, and

modifications to the interface were still being made while training sessions were being undertaken. It

therefore became important to have face-to-face visits and feedback on the website to inform users of

some of the more important changes, and how to undertake some of the common procedures (such

as saving a draft of the intervention and then coming back to it later).

One area where the software was consistently not used appropriately was in the section relating to

changes in drug therapy. It was intended that pharmacists should use the method developed to record

changes made during the intervention that would allow calculation of the costs associated with the

prescription before the intervention and again after the intervention (see section 4.1.3.4). If, for

example, a drug’s dose was increased or decreased, or a drug was ceased or commenced, there

would be a change in the costs associated with the prescriptions for these situations. We intended to

compare the before and after situation using these entered descriptions of changes in order to

determine the changes in costs. However, only some of the pharmacists in the study were aware of

this facility, and in some pharmacies activation of the facility resulted in system crashes or freezing,

requiring a re-start of the computer. This discouraged use of the “edit key” facility and as a result, the

data from this section of the software was considered unreliable.

There were a number of different computer system setups encountered when installing the software,

and some of these systems proved unstable once the software was installed. In two or three

pharmacies, the commencement of data collection resulted in system crashes which required

technical assistance to overcome. A number of pharmacies reported slowing of their systems while the

Comm Server was operational, and a few pharmacies chose to only turn on the Comm Server for

selected periods of the day to overcome this issue.

There were six pharmacies in the study who did not have broadband access, and were connected to

the internet using a dial-up connection. All of these pharmacies had issues with sending of dispensing

history data, and it was necessary to visit each of these at the completion of the trial in order to

download their data manually.

Despite these issues, the software was reasonably well received (see section 8.1) and those

pharmacists who used it regularly rapidly became proficient at recording interventions.

5.6.1 Time Taken to Enter an Intervention

In a post-trial questionnaire, pharmacists were asked what the average time was for them to perform

and record an intervention. The majority indicated that they could do this within 2-5 minutes (see

Figure 5.6-1).


Page 162 of 361

2

59

43

10

1

0

10

20

30

40

50

60

70

< 1 min 2 to 5 6 to 10 11 to 30 >30

Figure 5.6-1: Time to record and perform interventions during the Melbourne PROMISe trial

For the pharmacies that had an observer present (21 of the 52 pharmacies, 40.4%), the time taken to

record the intervention on the software for a sample of the interventions was further assessed.

Each observer aimed to record the time taken to record five interventions during the first week of the

trial and five interventions during the third week of the trial (the observers finished their time within the

pharmacy during this week). This time taken to record the intervention was able to be accurately (with

a digital stop watch) and impartially assessed by the observer. Results are shown in Table 5.6-1

Sampling Date Number Average

(min) Median (min)

Range (min)

Early in the trial (Before 27

th April to 3

rd May)

94 2.28 2.00 0.3-10.5

Week three of the trial (After 9

th to 12th May)

48 1.66 1.05 0.6-6.1

Overall 142 2.07 1.73 0.3-10.5

Table 5.6-1: The Average Time Recorded By The Observers Of Pharmacist Recording Time

The range of time taken to record interventions during the early part of the trial was between 20

seconds and 10.5 minutes. This fell to between 0.6 and 6.1 minutes during the third week of the trial.

Half of the pharmacists were able to record their interventions in 1.05 minutes or less during the third

week of the trial (approximately twice as fast as early in the trial).


Page 163 of 361

5.7 Non-Clinical Intervention Data

The non-clinical intervention data referred to in this section applies to the automatically recorded brand

substitution interventions.

The generation of automatic brand substitution records was developed by PCA NU Systems for this

phase of the PROMISe project.

It should be noted that there are multiple ways to substitute a generic brand. The automatic feature

was only activated for one of these methods. The pharmacists were encouraged to use this method

during the trial to ensure an accurate record of the rate of intervention. Thus, although a significant

number of brand substitutions were identified during the study, it is unlikely that all brand substitutions

were detected and documented.

The automation of brand substitution was deactivated in the June update of the WiniFRED software.

This update was installed in the pharmacies between the 26th of May and the 1

st of June. Hence there

was a gradual decline in recording rate during these dates as can be seen in Table 5.7-1 and Figure

5.7-1.

5.7.1 Frequency and Rate of brand substitution

Overall 11,493 brand substitutions were documented from 305,519 prescriptions (an average rate of

3.76%).

The daily rate of brand substitution fluctuated during the data collection period (see Figure 5.7-2).

There was the initial increase in rate, associated with the dissemination of information required on how

to make this documentation easier. This was followed by several weeks of routine documentation at

approximately 4.5% of all prescriptions, and then a tail off as the facility was withdrawn from

pharmacies.


Page 164 of 361

DateNon Clinical

InterventionsPrescriptions

Non Clinical

Intervention Rate (per

100 Prescriptions)

Thursday, April 21, 2005 48 9032 0.53

Friday, April 22, 2005 147 9127 1.61

Saturday, April 23, 2005 117 5455 2.14

Sunday, April 24, 2005 54 2647 2.04

Monday, April 25, 2005 46 2621 1.76

Tuesday, April 26, 2005 287 10217 2.81

Wednesday, April 27, 2005 316 8958 3.53

Thursday, April 28, 2005 289 8807 3.28

Friday, April 29, 2005 362 9000 4.02

Saturday, April 30, 2005 171 5363 3.19

Sunday, May 01, 2005 85 2688 3.16

Monday, May 02, 2005 340 8924 3.81

Tuesday, May 03, 2005 423 8423 5.02

Wednesday, May 04, 2005 328 8332 3.94

Thursday, May 05, 2005 401 9203 4.36

Friday, May 06, 2005 410 9026 4.54

Saturday, May 07, 2005 199 5239 3.80

Sunday, May 08, 2005 96 2363 4.06

Monday, May 09, 2005 468 9601 4.87

Tuesday, May 10, 2005 450 8850 5.08

Wednesday, May 11, 2005 405 8948 4.53

Thursday, May 12, 2005 373 8907 4.19

Friday, May 13, 2005 480 9151 5.25

Saturday, May 14, 2005 197 5662 3.48

Sunday, May 15, 2005 131 2850 4.60

Monday, May 16, 2005 462 10218 4.52

Tuesday, May 17, 2005 428 9028 4.74

Wednesday, May 18, 2005 364 9114 3.99

Thursday, May 19, 2005 407 9801 4.15

Friday, May 20, 2005 451 8893 5.07

Saturday, May 21, 2005 213 5713 3.73

Sunday, May 22, 2005 125 2784 4.49

Monday, May 23, 2005 482 9649 5.00

Tuesday, May 24, 2005 434 8943 4.85

Wednesday, May 25, 2005 360 8351 4.31

Thursday, May 26, 2005 320 8627 3.71

Friday, May 27, 2005 297 8422 3.53

Saturday, May 28, 2005 114 5249 2.17

Sunday, May 29, 2005 77 2789 2.76

Monday, May 30, 2005 192 9572 2.01

Tuesday, May 31, 2005 144 8972 1.60

Total 11493 305519 3.76

Table 5.7-1: Frequency and Rate of Non-Clinical Interventions


Page 165 of 361

Figure 5.7-1: Accumulation of Non-Clinical Interventions

Non Clinical Intervention Rates

0.00

1.00

2.00

3.00

4.00

5.00

6.00Cumulative Non Clinical Intevention Rate (per 100Prescriptions)Non Clinical Intervention Rate (per 100 Prescriptions)

Plateau Rate (4.4%)

Figure 5.7-2: Cumulative and Daily Non-Clinical Intervention Rate

As previously mentioned, there was a change in procedure required in order to automatically record

brand substitution interventions (involving using the keys GS). It seems clear that, although the intent

was that all pharmacies would record in this manner, some pharmacies did not alter their methods and

only recorded very few brand substitutions (see Figure 5.7-3). Twenty seven of the 52 pharmacies

only documented brand substitutions at a rate of 2% or less. This rate was lower than expected

through normal variation in substitution rates.

Cumulative Non-Clinical Interventions

11493

3055.19

0

2000

4000

6000

8000

10000

12000

14000

21/0

4/200

5

23/0

4/200

5

25/0

4/200

5

27/0

4/200

5

29/0

4/200

5

1/05

/200

5

3/05

/200

5

5/05

/200

5

7/05

/200

5

9/05

/200

5

11/0

5/200

5

13/0

5/200

5

15/0

5/200

5

17/0

5/200

5

19/0

5/200

5

21/0

5/200

5

23/0

5/200

5

25/0

5/200

5

27/0

5/200

5

29/0

5/200

5

31/0

5/200

5

Cumulat ive Non Clinical Intervent ions

Cumulat ive Prescript ions (x100)


Page 166 of 361

16

11

5

8

65

01

00

2

4

6

8

10

12

14

16

18

1 2 3 4 5 6 7 8 More

Brand Substitutions per 100 Prescriptions

Nu

mb

er

of

Ph

arm

ac

ies

Figure 5.7-3: Number of Pharmacies with Different Rates of Brand Substitution Interventions

In order to further examine this, the number and rate of brand substitution interventions documented

by each pharmacist was calculated. As previously indicated, a number of pharmacists did not actively

participate in the study and merely dispensed without documenting interventions.

During the period where brand substitution interventions were being documented (up to 31st May), 286

different pharmacists dispensed prescriptions at the 52 pharmacies. Of these, 50 did not record any

brand substitution interventions, and 20 only recorded one. As can be seen in Figure 5.7-4, 100 of the

pharmacists only recorded less than one brand substitution interventions per 100 prescriptions.


Page 167 of 361

100

53

34

25

1511

1411

83

11

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 More

Brand Substitution Interventions per 100 Prescriptions

Nu

mb

er

of

Ph

arm

ac

ists

Figure 5.7-4: Number of Pharmacists with Different Rates of Brand Substitution Interventions

These “poor uptake” pharmacists had a significant influence on the total number of brand substitution

interventions performed. (see Table 5.7-2). If these pharmacists had documented at the same rate as

the active participant pharmacists, the 57,867 prescriptions they dispensed would have resulted in

2,609 brand substitution interventions compared to the 350 interventions documented (2259 “missed

interventions”; 19.6% of the total).

Number of

Pharmacists

Brand

Substitution

Interventions

PrescriptionsRate (per 100

Prescriptions)Median Rate

Poor Uptake 100 350 57867 0.60 0.14

Active Participant 186 11153 247347 4.51 3.11

Total 286 11503 305214 3.77 1.78

Table 5.7-2: Brand Substitution Intervention Rates for Poor Uptake and Active Participant Pharmacists


Page 168 of 361

5.7.2 Drugs Involved in brand substitution

Of the brand substitutions recorded 99.4% (11,442 of 11493) could be linked to a specific medication.

The drug involved was specific for the brand of the product and overall there were 1,250 different

brands of drugs involved. The most frequent brand substitution interventions were with Panadeine

Forte (630 occasions) and a range of antibiotics (Amoxil, Augmentin Duo Forte and Keflex).

The top 100 brands involved in brand substitution interventions are shown from Table 5.7-3 to Table

5.7-5.

Rank Brand Name Number Brand Premium

Total Savings

1 PANADEINE FORTE TAB 500-30mg 20 630 $1.58 $995.40

2 AMOXIL CAP 500mg 539 $1.00 $539.00

3 KEFLEX CAP 500mg 453 $1.87 $847.11

4 AUGMENTIN DUO FORTE TAB 367 $1.30 $477.10

5 VALIUM TAB 5mg 251 $1.32 $331.32

6 VOLTAREN EC-TABS 50mg 201 $3.21 $645.21

7 CECLOR CD CR-TAB 375mg 194 $1.54 $298.76

8 RULIDE TAB 300mg 164 $2.27 $372.28

9 VENTOLIN CFC FREE MET-AERO 158 $1.00 $158.00

10 STEMETIL TAB 5mg 25 150 $1.95 $292.50

11 AMOXIL FORTE SYRP 250mg/5mL 149 $1.01 $150.49

12 LOSEC TAB 20mg (base) 148 $2.00 $296.00

13 MOBIC TAB 15mg 130 $1.84 $239.20

14 TRAMAL CAP 50mg 20 124 $0.60 $74.40

15 RULIDE TAB 150mg 123 $2.27 $279.21

16 AUGMENTIN DUO TAB 500mg/125mg 115 $0.99 $113.85

17 CECLOR SUSP 250mg/5mL 115 $1.55 $178.25

18 AMOXIL SYRP 125mg/5mL 114 $1.00 $114.00

19 EES TAB 400MG 107 $2.89 $309.23

20 ZOLOFT TAB 100mg 30 99 $0.55 $54.45

21 SEREPAX TAB 30mg 96 $1.45 $139.20

22 ZOLOFT TAB 50mg 30 94 $0.55 $51.70

23 CIPRAMIL TAB 20mg 85 $3.20 $272.00

24 AMOXIL CAP 250mg 85 $1.00 $85.00

25 LASIX TAB 40mg 83 $1.11 $92.13

26 AUGMENTIN DUO 400 SYRP 81 $0.98 $79.38

27 ASTRIX TAB 100mg 112 80 $1.36 $108.80

28 AROPAX TAB 20mg 78 $1.10 $85.80

29 ZOCOR TAB 40mg 73 $0.70 $51.10

30 MOGADON TAB 5mg 73 $2.02 $147.46

Table 5.7-3: Brands of Drugs Involved in Brand Substitution Interventions (Ranks 1 to 30)


Page 169 of 361


Total Savings

31 MOBIC TAB 7.5mg 70 $1.83 $128.10

32 LIPEX TAB 40mg 69 $0.70 $48.30

33 KEFLEX CAP 250mg 66 $1.76 $116.16

34 KENACOMB EAR-OINT 5g 65 $1.10 $71.50

35 ZOCOR TAB 20MG 63 $0.70 $44.10

36 MICROGYNON 30 ED TAB 150/30mcg 62 $9.76 $605.12

37 CECLOR SUSP 125mg/5mL 61 $1.53 $93.33

38 FLAGYL TAB 400mg 60 $2.07 $124.20

39 VIBRAMYCIN TAB 100mg 60 $1.60 $96.00

40 KEFLEX SYRP 250mg/5mL 60 $1.76 $105.60

41 TENORMIN TAB 50mg 57 $3.55 $202.35

42 TRAMAL SR-TAB 200mg 56 $0.79 $44.24

43 MOXACIN CAP 500mg 56 $1.00 $56.00

44 AVANZA TAB 30mg 55 $1.25 $68.75

45 LOMOTIL TAB 2.5mg 20 53 $1.83 $96.99

46 IBILEX CAP 500mg 52 $1.87 $97.24

47 SIMVASTATIN TAB 40mg 51 $0.70 $35.70

48 MINOMYCIN TAB 50mg 51 $1.12 $57.12

49 SOFRADEX EAR-DRP 50 $1.50 $75.00

50 LIPEX TAB 20mg 49 $0.70 $34.30

51 TRAMAL SR-TAB 100mg 49 $0.79 $38.71

52 TRITACE TAB 5mg 49 $2.00 $98.00

53 MAXAMOX TAB 1g 47 $1.49 $70.03

54 TRIPRIM TAB 300mg 46 $1.36 $62.56

55 CILAMOX CAP 500mg 46 $1.00 $46.00

56 EES SUSP 400mg/5mL 100mL 45 $1.75 $78.75

57 ZANTAC TAB 150mg 44 $2.17 $95.48

58 MONOFEME TAB 150/30mcg 28 41 $9.76 $400.16

59 KLACID TAB 250mg 14 39 $1.58 $61.62

60 FLAGYL TAB 200mg 38 $1.98 $75.24

61 DIABEX TAB 500mg 37 $1.80 $66.60

62 NORMISON TAB 10mg 35 $1.06 $37.10

63 INDOCID CAP 25mg 35 $3.06 $107.10

64 TEMAZE TAB 10mg 34 $1.06 $36.04

65 DIAMICRON TAB 80mg 32 $1.35 $43.20

66 ZYLOPRIM TAB 300mg 31 $2.28 $70.68

67 ARISTOCORT CRM 0.02% 100g 31 $2.60 $80.60

68 EES SUSP 200mg/5mL 100mL 30 $2.19 $65.70

69 CARDIZEM CD SR-CAP 240mg 30 30 $2.64 $79.20

70 AMOXYCILLIN GENRX SYRP 29 $1.01 $29.29



Page 170 of 361


Total Savings

71 AMOXIL DUO TAB 1g 28 $1.49 $41.72

72 QUINATE TAB 300mg 28 $2.19 $61.32

73 FASIGYN TAB 500mg 27 $2.50 $67.50

74 NAPROSYN SR-TAB 1000mg 27 $1.77 $47.79

75 TRAMAL SR-TAB 150mg 27 $0.81 $21.87

76 ALPHAMOX CAP 500mg 26 $1.00 $26.00

77 PANAMAX TAB 500mg 100 26 $2.20 $57.20

78 VENTOLIN NEB 5mg 30 26 $2.20 $57.20

79 OROXINE TAB 50mcg 24 $1.30 $31.20

80 ZOCOR TAB 80mg 24 $0.71 $17.04

81 NOROXIN TAB 400mg 24 $3.73 $89.52

82 TRITACE TAB 2.5mg 23 $2.01 $46.23

83 MOXACIN SYRP 250mg/5mL 23 $1.01 $23.23

84 OROXINE TAB 100mcg 23 $1.30 $29.90

85 CEFACLOR CR-TAB 375mg 22 $1.54 $33.88

86 SIMVASTATIN TAB 20mg 22 $0.70 $15.40

87 REDIPRED O-LIQ 5mg/mL 30mL 22 $1.32 $29.04

88 CO57 22 $1.88 $41.31

89 LUVOX TAB 100mg 22 $1.60 $35.20

90 FLOPEN CAP 500mg 22 $0.58 $12.76

91 NORDETTE TAB 150/30mcg 28 21 $9.24 $194.04

92 BETALOC TAB 50mg 21 $3.26 $68.46

93 DORYX CAP 100mg 21 $1.52 $31.92

94 VIBRA TAB 50mg 21 $1.65 $34.65

95 CLAVULIN DUO FORTE TAB 20 $1.30 $26.00

96 XANAX TAB 500mcg 20 $1.16 $23.20

97 DOTHEP TAB 75mg 20 $1.05 $21.00

98 IMODIUM CAP 2mg 20 $0.92 $18.40

99 XANAX TAB 1mg 20 $1.35 $27.00

100 TRIFEME TAB 28 20 $9.24 $184.80

OTHER DRUGS 3637 $1.87* $6,801.19

TOTAL 11422 $19,936.46

* Average brand premium


Each of the previous tables highlights antibiotics as a frequently substituted group of medications. The

most frequently substituted medication was amoxicillin, which represented 11.9% of the brand

substitutions; this on average saved the patient $1.00 per prescription. Therefore, the savings

induced by these changes would produce important savings to the patients.

The most savings to the patient were from the substitution of Panadeine Forte with an individual

saving of $1.58 with total savings of $995.40. It should be noted that where approval prescriptions for

increased quantities of Panadeine Forte are dispensed the patient will pay multiple brand premiums.


Page 171 of 361

In the list of the top 20 brands substituted Zoloft 100mg appears at 20th with Zoloft 50mg a little lower

at 22nd

. Although this premium is relatively low it should be noted that the uptake of this substitution is

relatively high as this premium was only introduced in April this year.

5.7.2.1 Grouping of the Drugs involved in Brand substitutions

It was possible to classify the vast majority of drugs involved by the World Health Organisation

anatomical therapeutic classification (ATC) codes. This five-tiered classification system, groups drugs

into larger and larger therapeutic classifications.

The drugs involved in the brand substitution interventions at four of these levels of classification are

shown from Table 5.7-6 to

Table 5.7-9.

Number PercentJ01CA04 Amoxicillin 1357 11.9

J01DB01 Cefalexin 751 6.6

N02AA59 Codeine, combinations excl. psycholeptics 731 6.4

J01CR02 Amoxicillin and enzyme inhibitor 675 5.9

C10AA01 Simvastatin 493 4.3

J01DC04 Cefaclor 444 3.9

N05BA01 Diazepam 337 3.0

J01FA06 Roxithromycin 312 2.7

N02AX02 Tramadol 293 2.6

M01AB05 Diclofenac 249 2.2

R03AC02 Salbutamol 244 2.1

N06AB06 Sertraline 230 2.0

M01AC06 Meloxicam 224 2.0

J01FA01 Erythromycin 204 1.8

A02BC01 Omeprazole 195 1.7

J01AA02 Doxycycline 165 1.4

N05AB04 Prochlorperazine 156 1.4

G03AA07 Levonorgestrel and estrogen 152 1.3

C03CA01 Furosemide 135 1.2

N05BA04 Oxazepam 128 1.1

N06AB04 Citalopram 124 1.1

C07AB03 Atenolol 122 1.1

P01AB01 Metronidazole 115 1.0

C09AA05 Ramipril 103 0.9

A10BA02 Metformin 98 0.9

Other 3356 29.5

Total 11393 100.0

ATC Code

(L5)DESCRIPTION

Brand Substitution

Interventions

Table 5.7-6: Generic Drug Groups Involved in Brand Substitution Interventions (ATC Level 5)


Page 172 of 361

Number PercentJ01CA Penicillins with extended spectrum 1357 11.9J01DB First-generation cephalosporins 751 6.6N02AA Natural opium alkaloids 736 6.5J01CR Combinations of penicillins, incl. beta-lactamase inhibitors 675 5.9J01FA Macrolides 566 5.0N05BA Benzodiazepine derivatives 531 4.7N06AB Selective serotonin reuptake inhibitors 518 4.5C10AA HMG CoA reductase inhibitors 493 4.3J01DC Second-generation cephalosporins 444 3.9M01AB Acetic acid derivatives and related substances 294 2.6N02AX Other opioids 293 2.6M01AC Oxicams 263 2.3R03AC Selective beta-2-adrenoreceptor agonists 244 2.1J01AA Tetracyclines 219 1.9C09AA ACE inhibitors, plain 208 1.8A02BC Proton pump inhibitors 199 1.7C07AB Beta blocking agents, selective 177 1.6G03AA Progestogens and estrogens, fixed combinations 174 1.5N05CD Benzodiazepine derivatives 174 1.5N05AB Phenothiazines with piperazine structure 156 1.4P01AB Nitroimidazole derivatives 145 1.3S02CA Corticosteroids and antiinfectives in combination 137 1.2C03CA Sulfonamides, plain 135 1.2D07AC Corticosteroids, potent (group III) 121 1.1A02BA H2-receptor antagonists 110 1.0A10BA Biguanides 98 0.9

Others 2175 19.1

Total 11393 100.0

Brand Substitution

InterventionsATC Code

(L4)DESCRIPTION



Page 173 of 361


Number PercentJ01C BETA-LACTAM ANTIBACTERIALS, PENICILLINS 2142 18.8

J01D OTHER BETA-LACTAM ANTIBACTERIALS 1195 10.5

N02A OPIOIDS 1029 9.0

N06A ANTIDEPRESSANTS 713 6.3

M01AANTIINFLAMMATORY AND ANTIRHEUMATIC PRODUCTS,

NON-STEROIDS643 5.6

J01F MACROLIDES, LINCOSAMIDES AND STREPTOGRAMINS 571 5.0

N05B ANXIOLYTICS 531 4.7

C10A CHOLESTEROL AND TRIGLYCERIDE REDUCERS 515 4.5

A02BDRUGS FOR PEPTIC ULCER AND GASTRO-

OESOPHAGEAL REFLUX DISEASE (GORD)311 2.7

G03A HORMONAL CONTRACEPTIVES FOR SYSTEMIC USE 269 2.4

R03A ADRENERGICS, INHALANTS 245 2.2

J01A TETRACYCLINES 219 1.9

C07A BETA BLOCKING AGENTS 214 1.9

D07A CORTICOSTEROIDS, PLAIN 209 1.8

C09A ACE INHIBITORS, PLAIN 208 1.8

A10B ORAL BLOOD GLUCOSE LOWERING DRUGS 180 1.6

N05C HYPNOTICS AND SEDATIVES 174 1.5

N05A ANTIPSYCHOTICS 160 1.4

P01AAGENTS AGAINST AMOEBIASIS AND OTHER PROTOZOAL

DISEASES145 1.3

S02CCORTICOSTEROIDS AND ANTIINFECTIVES IN

COMBINATION137 1.2

C03C HIGH-CEILING DIURETICS 135 1.2

J01E SULFONAMIDES AND TRIMETHOPRIM 102 0.9

B01A ANTITHROMBOTIC AGENTS 95 0.8

C08DSELECTIVE CALCIUM CHANNEL BLOCKERS WITH DIRECT

CARDIAC EFFECTS93 0.8

A07D ANTIPROPULSIVES 85 0.7

N03A ANTIEPILEPTICS 79 0.7

OTHERS 994 8.7

TOTAL 11393 100.0

Brand Substitution

InterventionsDESCRIPTIONATC Code

(L3)


Page 174 of 361


Based on this raw data, we can extrapolate the direct to consumer savings due to brand price

premium reductions to the Australian population. There were 220.1 million prescriptions dispensed in

Australia in 2003,12

compared to the 305,519 that resulted in the brand substitution interventions in the

PROMISe study. The extrapolated direct to consumer savings from brand substitution interventions

calculates out to approximately $14.3m per year. We believe this to be a very conservative estimate

based on the underuse of the brand substitution component of the software.

12 Statistics in Drug Use in Australia 2004

Number PercentJ01 ANTIBACTERIALS FOR SYSTEMIC USE 4277 37.5

N02 ANALGESICS 1099 9.6

N05 PSYCHOLEPTICS 865 7.6

N06 PSYCHOANALEPTICS 713 6.3

M01 ANTIINFLAMMATORY AND ANTIRHEUMATIC PRODUCTS 643 5.6

C10 SERUM LIPID REDUCING AGENTS 515 4.5

A02 DRUGS FOR ACID RELATED DISORDERS 312 2.7

G03SEX HORMONES AND MODULATORS OF THE GENITAL

SYSTEM301 2.6

R03 DRUGS FOR OBSTRUCTIVE AIRWAY DISEASES 268 2.4

C09 AGENTS ACTING ON THE RENIN-ANGIOTENSIN SYSTEM 215 1.9

C07 BETA BLOCKING AGENTS 214 1.9

D07 CORTICOSTEROIDS, DERMATOLOGICAL PREPARATIONS 209 1.8

C03 DIURETICS 203 1.8

A10 DRUGS USED IN DIABETES 187 1.6

P01 ANTIPROTOZOALS 176 1.5

C08 CALCIUM CHANNEL BLOCKERS 169 1.5

S02 OTOLOGICALS 137 1.2

A07ANTIDIARRHEALS, INTESTINAL

ANTIINFLAMMATORY/ANTIINFECTIVE 102 0.9

B01 ANTITHROMBOTIC AGENTS 95 0.8

C01 CARDIAC THERAPY 85 0.7

N03 ANTIEPILEPTICS 79 0.7

H02 CORTICOSTEROIDS FOR SYSTEMIC USE 68 0.6

M04 ANTIGOUT PREPARATIONS 67 0.6

H03 THYROID THERAPY 66 0.6

S01 OPHTHALMOLOGICALS 66 0.6

C02 ANTIHYPERTENSIVES 56 0.5

OTHERS 206 1.8

TOTAL 11393 100.0

Brand Substitution

Interventions

ATC

Code

(L2)

DESCRIPTION


Page 175 of 361

Brand

Substitutions Prescriptions

Brand Price Premium Savings

PROMISe 11422 305519 $19,936

Extrapolation to Australia

8228563 220100000 $14,362,494

Table 5.7-10: Extrapolation of Brand Substitution Direct Savings from PROMISe Study

5.8 Clinical Intervention Data

Overall there were 2,39613

clinical interventions recorded during the PROMISe study. During this

period, 435,520 prescriptions were dispensed (a rate of approximately 0.55 interventions per 100

prescriptions) for 258,979 patients (an intervention rate of 0.92 interventions per 100 patients).

5.8.1 Frequency and Rate

The frequency of clinical interventions gradually increased during the first few days of the project.

There were two reasons for this. Firstly, software was not installed in all pharmacies simultaneously.

Software installation commenced on Thursday 21st April, but was not completed in all pharmacies until

26th April. Secondly, as pharmacies were visited and pharmacists were individually shown how to use

the documentation system, their uptake improved (see Figure 5.7-1).

0.00

0.20

0.40

0.60

0.80

1.00

1.20

21/0

4/2

005

28/0

4/2

005

5/0

5/2

005

12/0

5/2

005

19/0

5/2

005

26/0

5/2

005

2/0

6/2

005

9/0

6/2

005

16/0

6/2

005

Clinical Interventions per 100 Prescriptions

Cumulative Clinical Intervention Rate (per 100

prescriptions)Poly. (Clinical Interventions per 100 Prescriptions)

Figure 5.8-1: Cumulative And Daily Clinical Intervention Rate

13 11 of the interventions did not have a patient ID associated with them


Page 176 of 361

Also shown in

Figure 5.8-1 is the fluctuation of daily clinical intervention rate which is also tabulated in Table 5.8-1.

There are a number of “low points” on the graph, which correspond with weekends and public

holidays.

Date

Proactive

Clinical

Interventions

Reactive

Clinical

Interventions

Total Clinical

Interventions

Total

PrescriptionsPatients

Clinical

Interventions

Per 100

Prescriptions

Clinical

Interventions

Per 100

PatientsThursday, 21 April 2005 56 4 60 9032 5145 0.66 1.17

Friday, 22 April 2005 62 15 77 9127 5311 0.84 1.45

Saturday, 23 April 2005 42 3 45 5455 3297 0.82 1.36

Sunday, 24 April 2005 15 2 17 2647 1732 0.64 0.98

Monday, 25 April 2005 11 0 11 2621 1641 0.42 0.67

Tuesday, 26 April 2005 68 13 81 10217 5979 0.79 1.35

Wednesday, 27 April 2005 74 13 87 8958 5271 0.97 1.65Thursday, 28 April 2005 57 21 78 8807 5298 0.89 1.47

Friday, 29 April 2005 72 18 90 9000 5252 1.00 1.71

Saturday, 30 April 2005 27 8 35 5363 3262 0.65 1.07

Sunday, 1 May 2005 11 2 13 2688 1660 0.48 0.78

Monday, 2 May 2005 57 14 71 8924 5267 0.80 1.35Tuesday, 3 May 2005 57 14 71 8423 5013 0.84 1.42

Wednesday, 4 May 2005 76 16 92 8332 4943 1.10 1.86

Thursday, 5 May 2005 51 6 57 9203 5310 0.62 1.07

Friday, 6 May 2005 68 16 84 9026 5204 0.93 1.61

Saturday, 7 May 2005 18 4 22 5239 3193 0.42 0.69Sunday, 8 May 2005 4 0 4 2363 1521 0.17 0.26

Monday, 9 May 2005 85 20 105 9601 5698 1.09 1.84

Tuesday, 10 May 2005 54 15 69 8850 5207 0.78 1.33

Wednesday, 11 May 2005 50 11 61 8948 5263 0.68 1.16

Thursday, 12 May 2005 46 16 62 8907 5276 0.70 1.18Friday, 13 May 2005 54 13 67 9151 5389 0.73 1.24

Saturday, 14 May 2005 21 7 28 5662 3480 0.49 0.80

Sunday, 15 May 2005 0 1 1 2850 1816 0.04 0.06

Monday, 16 May 2005 39 9 48 10218 6043 0.47 0.79

Tuesday, 17 May 2005 54 14 68 9028 5369 0.75 1.27

Wednesday, 18 May 2005 42 7 49 9114 5365 0.54 0.91Thursday, 19 May 2005 46 9 55 9801 5661 0.56 0.97

Friday, 20 May 2005 42 12 54 8893 5318 0.61 1.02

Saturday, 21 May 2005 18 5 23 5713 3534 0.40 0.65

Sunday, 22 May 2005 10 5 15 2784 1802 0.54 0.83

Monday, 23 May 2005 31 17 48 9649 5783 0.50 0.83Tuesday, 24 May 2005 40 8 48 8943 5378 0.54 0.89

Wednesday, 25 May 2005 26 8 34 8351 4990 0.41 0.68

Thursday, 26 May 2005 30 3 33 8627 5140 0.38 0.64

Friday, 27 May 2005 24 8 32 8422 5045 0.38 0.63

Saturday, 28 May 2005 9 2 11 5249 3242 0.21 0.34Sunday, 29 May 2005 10 5 15 2789 1760 0.54 0.85

Monday, 30 May 2005 32 12 44 9572 5677 0.46 0.78

Tuesday, 31 May 2005 18 9 27 8972 5405 0.30 0.50

Wednesday, 1 June 2005 31 6 37 8693 5145 0.43 0.72

Thursday, 2 June 2005 20 7 27 9299 5272 0.29 0.51Friday, 3 June 2005 19 7 26 9050 5254 0.29 0.49

Saturday, 4 June 2005 9 3 12 5497 3299 0.22 0.36

Sunday, 5 June 2005 0 0 0 2804 1783 0.00 0.00

Monday, 6 June 2005 26 5 31 9512 5707 0.33 0.54

Tuesday, 7 June 2005 16 6 22 8762 5158 0.25 0.43

Wednesday, 8 June 2005 26 6 32 8958 5367 0.36 0.60Thursday, 9 June 2005 35 7 42 8952 5283 0.47 0.80

Friday, 10 June 2005 28 8 36 9174 5333 0.39 0.68

Saturday, 11 June 2005 7 4 11 5489 3373 0.20 0.33

Sunday, 12 June 2005 2 0 2 2573 1620 0.08 0.12

Monday, 13 June 2005 6 0 6 3161 2014 0.19 0.30Tuesday, 14 June 2005 27 5 32 10445 6192 0.31 0.52

Wednesday, 15 June 2005 16 4 20 9405 5648 0.21 0.35

Thursday, 16 June 2005 19 5 24 9141 5355 0.26 0.45

Friday, 17 June 2005 22 11 33 9086 5236 0.36 0.63

Total 1916 469 2385 435520 258979 0.55 0.92

(Shading = Public Holiday)

Table 5.8-1: Daily Number Of Clinical Interventions, Prescriptions And Patients


Page 177 of 361

When the daily rates of clinical interventions are grouped into weeks of the study, it can be seen that

the intervention rate gradually falls away (see Table 5.8-2 and Figure 5.8-2). There is an initial

increase (correlated with activation of the pharmacies and face to face visits); this is followed by a

short plateau when all pharmacies were connected, the intervention prompt was active and observers

and PROMISe face to face visits were taking place. There is then a distinct fall away which relates to

cessation of these support activities.

• Observer visits ceased on 11th May (middle of week 3),

• The project team left Melbourne at the end of week 4

• The intervention prompt was turned off on 31st May (end of week 5)

• Remuneration for interventions was crossed over at the end of Week 2 and Payment was either

continued or reinstituted at the end of Week 4, for the third phase of the trial.

Table 5.8-2: Weekly Clinical Intervention Rates

Week of

Study

Proactive

Clinical

Interventions

Reactive

Clinical

Interventions

Total

Clinical

Interventions

Total

PrescriptionsPatients

Clinical

Interventions

Per 100

Prescriptions

Clinical

Interventions

Per 100

Patients

1 457 89 546 65864 38926 0.83 1.40

2 347 76 423 51959 30659 0.81 1.38

3 311 79 390 53059 31547 0.74 1.24

4 244 59 303 55566 33052 0.55 0.92

5 179 54 233 52489 31672 0.44 0.74

6 139 48 187 53624 31755 0.35 0.59

7 140 35 175 53659 31930 0.33 0.55

8 99 29 128 49300 29438 0.26 0.43

Total 1916 469 2385 435520 258979 0.55 0.92


Page 178 of 361

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1 2 3 4 5 6 7 8

Week of Study

Clinical Interventions Per 100Prescriptions

Clinical Interventions Per 100Patients


Figure 5.8-2: Weekly Clinical Intervention Rates Per 100 Prescriptions

The decline in documentation rate of clinical interventions was evident within the first two weeks of the

study. The raw number of clinical interventions recorded declined and the decline was not related to a

reduction in number of prescriptions, which continued to accumulate at a relatively constant rate (see

Figure 5.8-3).


Page 179 of 361

Cumulative Clinical Interventions and Prescriptions

2396

4355.2

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Thurs

day, 2

1 A

pril 2

005

Satu

rday, 2

3 A

pril 2

005

Monday, 2

5 A

pril 2

005

Wednesday, 2

7 A

pril 2

005

Frid

ay, 2

9 A

pril 2

005

Sunday, 1

May 2

005

Tuesday, 3

Ma

y 2

005

Thurs

day, 5

Ma

y 2

005

Satu

rday, 7

Ma

y 2

005

Monday, 9

May 2

005

Wednesday, 1

1 M

ay 2

005

Frid

ay, 1

3 M

ay 2

005

Sunday, 1

5 M

ay 2

005

Tuesday, 1

7 M

ay 2

005

Thurs

day, 1

9 M

ay 2

005

Satu

rday, 2

1 M

ay 2

005

Monday, 2

3 M

ay 2

005

Wednesday, 2

5 M

ay 2

005

Frid

ay, 2

7 M

ay 2

005

Sunday, 2

9 M

ay 2

005

Tuesday, 3

1 M

ay 2

005

Thurs

day, 2

Ju

ne 2

005

Satu

rday, 4

Jun

e 2

005

Monday, 6

Jun

e 2

005

Wednesday, 8

June 2

005

Frid

ay, 1

0 J

une

2005

Sunday, 1

2 J

un

e 2

005

Tuesday, 1

4 J

une 2

005

Thurs

day, 1

6 J

une 2

005

Cumulative Clinical Interventions

Cumulative Prescriptions (x100)

Figure 5.8-3: Daily Accumulation Of Clinical Interventions And Prescriptions

In order to further examine the intervention rate, a key table of parameters specific for each unique

Pharmacist and Day was constructed. Rates of clinical interventions per 100 prescriptions and per 100

patients were calculated. Days where an individual pharmacist dispensed less than 20 prescriptions

were excluded as these gave unrepresentatively high intervention rates (see more detailed analysis in

section 6).

There were significant differences between the rates of interventions between each of the different

phases of the study, with the rate during Phase one being highest, and the rate during Phase 3 being

lowest (see Table 5.8-3)


Page 180 of 361

Phase 1

Mean (95% CI) [Median]

Phase 2 Mean (95% CI)

[Median]


[Median] Intervention Rate per 100 Prescriptions

0.89 (0.78-0.99)1,2

[ 0 ] 0.73 (0.63-0.83)

1,3

[ 0 ] 0.39 (0.33-0.46)

2,3

[ 0 ]

Intervention Rate per 100 Patients

1.50 (1.33-1.68)4,5

[ 0 ] 1.22 (1.06-1.39)

4,6

[ 0 ] 0.69 (0.55-0.84)

5,6

[ 0 ]

Mann-Whitney Test

1Z = 3.854, p <0.001

2Z = 12.02, p <0.001

4Z = 3.837, p <0.001

5Z = 12.035, p <0.001

3Z = 7.155, p <0.001

6Z = 7.169, p <0.001

Table 5.8-3: Intervention Rates During Different Phases of the Study (Based on Pharmacist-Days)

There were a large number of days where no interventions and only very few prescriptions were

dispensed. As a result the median intervention rates for all groups is zero. To overcome the paucity of

data when considered at the level of pharmacist day, the data was re-constructed into weekly blocks

and a Pharmacist-Week of Study construct was prepared. As well as reducing the number of records

with zero interventions in a week (compared to a day), this process also reduced the number of

records where rates of interventions were unrealistically high (due to one intervention and only very

few prescriptions). On examining the information, it seems that some pharmacists were present at the

pharmacy on a day when they were not actively dispensing, yet they recorded interventions. It is likely

that these were interventions from a previous day. By collating a week of prescriptions and

interventions, a more representative rate of interventions for each pharmacist could be constructed.

The rates of interventions in Table 5.8-4 are higher than those above, as they include the previously

excluded cases of interventions.

Phase 1

Mean (95% CI) [Median]


[Median]


[Median] Intervention Rate per 100 Prescriptions

1.03 (0.83-1.23)1,2

[ 0.72 ]

0.86 (0.63-1.09)1,3

[ 0.40 ]

0.44 (0.34-0.55)2,3

[ 0.20 ]


1.78 (1.44-2.13)4,5

[ 1.24 ]

1.48 (1.09-1.86)4,6

[ 0.59 ]

0.75 (0.59-0.92)5,6

[ 0.31 ]

Mann-Whitney Test

1Z = 2.953 p = 0.003

2Z = 8.298, p <0.001

4Z = 2.959, p =0.003

5Z = 8.360, p <0.001

3Z = 5.118, p <0.001

6Z = 5.110, p <0.001

Table 5.8-4: Intervention Rates During Different Phases of the Study (Based on Pharmacist-Weeks)

The magnitude of the reduction in intervention rate is significant with the intervention rate during phase

three being ~42% of the rates in phase one. The reduced rate is likely to be due to a combination of a

reduced rate of recording of interventions and a reduced rate of performing interventions (see section

5.8.11).


Page 181 of 361

5.8.2 Categories and Subcategories of Interventions

The categories and subcategories of clinical interventions are shown in Table 5.8-5. The majority of

clinical interventions were one of three categories:

• drug selection problems (22.7%),

• dosage problems (19.4%) or

• education or information problems (17.4%).

A significant number of the clinical interventions were within the category of untreated indications. The

use of the automated intervention prompt (see Section 5.8.14) directly stimulated 202 aspirin

prophylaxis interventions in this category. If these aspirin interventions are not considered, the

individual subcategories that were associated with large numbers of interventions were: non-specific

drug selection problems (186 interventions) and problems with the dosage of medication being either

too high (178 interventions) or too low (169 interventions).

The range of categories and subcategories which were documented for the interventions was in

keeping with the pilot study, and with our understanding of the types of drug-related problems

identified in routine community pharmacy practice.


Page 182 of 361

Subcategory Number

Drug selection Duplication 83 3.46%

Drug selection Drug interaction 58 2.42%

Drug selection Wrong drug 120 5.01%

Drug selection Wrong dosage form 98 4.09%

Drug selection Other drug selection problem 186 7.76%

Over or underdose prescribed Dose too high 178 7.43%

Over or underdose prescribed Dose too low 169 7.05%

Over or underdose prescribed Other Dose Problem 118 4.92%

Compliance Taking too little 117 4.88%

Compliance Taking too much 48 2.00%

Compliance Intentional drug misuse 12 0.50%

Compliance Difficulty using dosage form 44 1.84%

Compliance Other Compliance Problem 54 2.25%

Untreated indications Condition not adequately treated 97 4.05%

Untreated indications Preventive therapy required 266 11.10%

Untreated indications Other Untreated indication Problem 19 0.79%

Monitoring Laboratory Monitoring 15 0.63%

Monitoring Non-Laboratory monitoring 23 0.96%

Monitoring Other Monitoring Problem 9 0.38%

Education or Information Patient drug information request 87 3.63%

Education or Information Confusion about therapy 120 5.01%

Education or Information Demonstration of device 62 2.59%

Education or Information Disease management or advice 89 3.71%

Education or Information Other Education or Information Problem 60 2.50%

N

Toxicity or Adverse reaction Toxicity caused by dose 17 0.71%

Toxicity or Adverse reaction Toxicity caused by drug interaction 87 3.63%

Toxicity or Adverse reaction Toxicity evident 129 5.38%

Toxicity or Adverse reaction Other Toxicity/Adverse Effect problem 31 1.29%

2396 100% 100%

15.9%

2.0%

17.4%

11.0%

% of Total

22.7%

19.4%

11.5%

T

Category

Total

Non Clinical (see elsewhere)

D

O

C

U

M

E

Table 5.8-5: Categories and Subcategories of Clinical Interventions

An example of one of the common types of interventions is shown in Figure 5.8-4.


Page 183 of 361

Untreated indications - preventative therapy required Summary Problem: at risk diabetic patient not on antiplatelet therapy. Male (66+ yo) diabetic patient with hypercholesterolemia presented a prescription for gliclazide. During the dispensing process the aspirin pop-up alerted the pharmacist to enquire about antiplatelet therapy. The pharmacist discussed low dose aspirin therapy with the patient and identified that it would be of benefit as it provides some cardiovascular protection and the patient had no contraindications to aspirin use. Written material was provided and the patient will discuss this with their doctor at the next visit. Outcome: the risk of a cardiovascular event reduced Category Untreated indications Subcategory Preventative therapy required Actions Investigation: Patient History Discussion with patient or carer Recommendations Refer to prescriber Education/counselling session Outcome Accepted Significance Moderate

Figure 5.8-4: Example of Intervention in Untreated Indication Category

5.8.3 Actions

In almost 90% of cases (1,924 of 2,186 cases), the pharmacist investigated the drug-related problem

by discussing the issue with the patient or the carer. In 33.8% (738) of cases, the pharmacist

contacted the prescriber in order to clarify the problem and in 44.6% (976) of cases, the pharmacist

checked the dispensing history in order to investigate the problem. Multiple actions were frequent, and

the average number of actions per intervention was 1.87.

Table 5.8-6: Actions Taken To Investigate Clinical Interventions

As would be expected, the prescriber was more likely to be contacted if the problem was related to

dosage of the medication or to a drug selection issue (see Table 5.8-7). In these circumstances, the

prescriber would be required to modify the drug dose, or change the prescription.


Page 184 of 361

INTERVENTION CATEGORYPrescriber not

contacted

Prescriber

Contacted% Contacted

DRUG SELECTION 222 291 56.7

OVER OR UNDERDOSE PRESCRIBED 149 279 65.2

COMPLIANCE 180 69 27.7

UNTREATED INDICATIONS 325 24 6.9

MONITORING 33 6 15.4

EDUCATION OR INFORMATION 337 23 6.4

TOXICITY OR ADVERSE REACTION 201 46 18.6

TOTAL 1447 738 33.8

Table 5.8-7: Occasions Where Prescriber Was Contacted Compared To Category Of Intervention

5.8.4 Recommendations

In almost one-quarter of the clinical interventions, the pharmacist recommended referral to the

prescriber to resolve the problem (521 occasions, 23.8%; see Table 5.8-8).

In 38.1% (834 occasions) of the clinical interventions, a counselling and education session was

provided to the patient to resolve the problem. These two recommendations (referral to the prescriber

and an education or counselling session) accounted for over 50% of the recommendations made by

pharmacists to resolve the drug-related problems identified. Again, this is consistent with our

understanding of community pharmacy practice, where potential problems are resolved by discussion

with the patient, their prescriber or both.

An average of 1.5 recommendations were made for each intervention, indicating that multiple

recommendations were common.

Table 5.8-8: Recommendations Made To Resolve Clinical Interventions


Page 185 of 361

When the recommendations were examined in terms of their major type, the most common type of

recommendation related to a change in therapy, with 44.4% (1,477 occasions) of clinical interventions

receiving these types of recommendations (see Table 5.8-9). Over half of these changes in therapy

recommendations (56.7%) were related to dosage or drug changes.

Provision of information was the next most common type of recommendation, with 33.2% (1,105) of

clinical interventions receiving recommendations of this type. Within this type, 75.5% (834 occasions)

of recommendations related to presumably verbal provision of information in the form of a counselling

or education session.

When a referral recommendation was made, this was almost uniformly to the prescriber (96.1%; 521

occasions).

Recommendation Code Number% within

category

% of

Total

Dose Change R1 418 28.3%

Drug Change R2 419 28.4%

Drug Formulation Change R3 112 7.6%

Drug Brand Change R4 52 3.5%

Dose Frequency or Schedule Change R5 200 13.5%

Prescription Not Dispensed R6 105 7.1%

Other Changes in Therapy R7 171 11.6%

1477 100.0%

Refer to Prescriber R8 521 96.1%

Refer to Hospital R9 5 0.9%

Refer for Medication Review R10 6 1.1%

Other Referral Required R11 10 1.8%

542 100.0%

Education or Counselling Session R12 834 75.5%

Written Summary of Medications R13 56 5.1%

Commence Dose Administration Aid R14 40 3.6%

Other Written Information R15 175 15.8%

1105 100.0%

Non-Laboratory Monitoring R16 105 63.3%

Laboratory Monitoring R17 61 36.7%

166 100.0%

No Recommendation Required R0 36 100.0% 1.1%

3326 100.0%Total

44.4%

A Change in Therapy

A Referral Required

Provision of Information

Monitoring

16.3%

33.2%

5.0%

Change in Therapy Subtotal

Referral Subtotal

Information Subtotal

Monitoring Subtotal

Table 5.8-9: Recommendations Made To Resolve Clinical Interventions (Grouped)

An example of an intervention with multiple recommendations is shown in Figure 5.8-5.


Page 186 of 361

Compliance - taking too little Summary Problem: lack of knowledge about asthma preventative. Male (21-65 yo) asthmatic presents with a prescription for salmeterol/fluticasone (Seretide) Accuhaler. Having dispensed the Seretide 500/50 2 puffs bd and noticing that the patient had not had it dispensed for 2 months; the pharmacist discussed the use of the patient’s inhalers which also included salbutamol (Ventolin). The patient mentioned that he had been using Ventolin every 2-3 hours due to worsening asthma, that woke up twice last night to use the Ventolin and was becoming concerned by potential side effects or dependence. He also mentioned that he used one puff of the Seretide every now and then only. The pharmacist educated the patient on the use of the Seretide and encouraged regular use. The patient was also advised to see the doctor if symptoms did not improve. Outcome: potentially prevented a severe asthma attack Category Compliance Subcategory taking too little Actions Education/ counselling session Recommendations Dose frequency/schedule change Refer to prescriber Discussion with patient Outcome Accepted Significance Moderate

Figure 5.8-5: Example of Intervention with Multiple Recommendations

When the types of recommendations were compared to the initial categories of interventions, a

number of relationships were identified. Interventions where the recommendation was for a change in

therapy were more likely to be either drug selection problems or dosage problems. Interventions

where a referral was required were more likely to involve a drug-related problem associated with

toxicity or an untreated indication requiring addition of therapy. Recommendations associated with

provision of information were more likely to be associated with education or compliance issues (see

shaded portions of Table 5.8-10).

Of interest, a significant number of information provision recommendations resulted from the untreated

indication category of interventions. These were related to the automated intervention prompted, which

included a printed handout which was intended to be provided to the patient.


Page 187 of 361

Table 5.8-10: Type Of Recommendation Compared To Category Of Intervention

5.8.5 Outcomes

Over 80% of the recommendations made by the pharmacists were indicated as accepted (see Table

5.8-11), and over 90% of the recommendations were indicated as either accepted or partially

accepted. In only 1.7% of the interventions (40 occasions) did the pharmacist indicate that the

recommendation(s) were not accepted.

Outcome Number% of

Total

Accepted 1967 82.1%

Not accepted 40 1.7%

Partially Accepted 235 9.8%

Unknown 154 6.4%

Total 2396 100.0%

Table 5.8-11: Outcomes Of Clinical Interventions

Commonly accepted recommendations included those related to problems that were of the drug selection, dose or education category (see shaded portions in

Table 5.8-12).

#% of

Column#

% of

Column#

% of

Column#

% of

Column#

% of

Column# %

DRUG SELECTION 490 33.18% 51 9.41% 142 12.85% 19 11.45% 5 13.89% 707 21.3%

OVER OR

UNDERDOSE

PRESCRIBED

397 26.88% 32 5.90% 75 6.79% 10 6.02% 16 44.44% 530 15.9%

COMPLIANCE 139 9.41% 56 10.33% 183 16.56% 15 9.04% 3 8.33% 396 11.9%

UNTREATED

INDICATIONS235 15.91% 218 40.22% 221 20.00% 20 12.05% 0 0.00% 694 20.9%

MONITORING 5 0.34% 13 2.40% 21 1.90% 27 16.27% 0 0.00% 66 2.0%

EDUCATION OR

INFORMATION94 6.36% 60 11.07% 343 31.04% 10 6.02% 12 33.33% 519 15.6%

TOXICITY OR

ADVERSE

REACTION

117 7.92% 112 20.66% 120 10.86% 65 39.16% 0 0.00% 414 12.4%

1477 100.00% 542 100.00% 1105 100.00% 166 100.00% 36 100.00% 3326 100.0%

Change in

TherapyTotal

TOTAL44.4% 16.3% 33.2% 5.0% 1.1% 100.0%

INTERVENTION

CATEGORY

No Recomm'n

MadeMonitoringInformationReferral


Page 188 of 361

#% of

Category#

% of

Category#

% of

Category#

% of

Category#

% of

Category

DRUG SELECTION 489 89.9% 14 2.6% 19 3.5% 22 4.0% 544 22.7%

OVER OR UNDERDOSE

PRESCRIBED432 92.9% 7 1.5% 21 4.5% 5 1.1% 465 19.4%

COMPLIANCE 209 75.7% 12 4.3% 41 14.9% 14 5.1% 276 11.5%

UNTREATED INDICATIONS 211 55.1% 1 0.3% 84 21.9% 87 22.7% 383 16.0%

MONITORING 35 74.5% 10 21.3% 2 4.3% 47 2.0%

EDUCATION OR

INFORMATION374 89.5% 2 0.5% 33 7.9% 9 2.2% 418 17.4%

TOXICITY OR ADVERSE

REACTION217 82.5% 4 1.5% 27 10.3% 15 5.7% 263 11.0%

1967 560.1% 40 10.7% 235 84.2% 154 100.0% 2396 100.0%

INTERVENTION

CATEGORY

TOTAL82.1% 1.7%

Not acceptedAccepted

9.8% 6.4% 100.0%

TotalUnknownPartially

Accepted

Table 5.8-12: Acceptance Of Recommendation Compared To Category Of Intervention

There was a high level of acceptance for recommendations concerning changes in therapy and provision of information (see shaded portions in

Table 5.8-13).

# % # % # % # % # % # %

ACCEPTED 1245 84.3% 315 58.1% 867 78.5% 121 72.9% 28 77.8% 2576 77.5%

NOT ACCEPTED 28 1.9% 6 1.1% 20 1.8% 2 1.2% 1 2.8% 57 1.7%

PARTIALLY

ACCEPTED114 7.7% 113 20.8% 130 11.8% 35 21.1% 3 8.3% 395 11.9%

UNKNOWN 90 6.1% 108 19.9% 88 8.0% 8 4.8% 4 11.1% 298 9.0%

1477 100.0% 542 100.0% 1105 100.0% 166 100.0% 36 100.0% 3326 100.0%

No

Recommendation

Made

TotalOutcome

TOTAL44.4% 16.3% 33.2% 5.0% 1.1% 100.0%

Change in Therapy

Recommendation

Referral

Recommendation

Information

Recommendation

Monitoring

Recommendation

Table 5.8-13: Acceptance Of Recommendation Compared To Type Of Recommendation

5.8.6 Clinical Significance

Almost one third of the clinical interventions (31.6%; 758 occasions) were classified as either of

moderate or severe level of clinical significance by the recording pharmacist. Moderate clinical

significance interventions were those that were likely to require medical intervention to resolve, and

severe clinical significance interventions were those that were likely to require hospitalisation to

resolve (see Table 5.8-14).

Pharmacist

Assigned

Clinical

Significance

Code Number % of Total

Nil S0 69 2.9%

Low S1 378 15.8%

Mild S2 1191 49.7%

Moderate S3 676 28.2%

Severe S4 82 3.4%

2396 100.0%Total

Table 5.8-14: Pharmacist Assigned Significance Of Clinical Interventions


Page 189 of 361

These more highly significant interventions (ie moderate or severe) were more likely to be drug

selection problems or toxicity problems (see shaded portion of Table 5.8-15).

Total

#% of

Category#

% of

Category#

% of

Category#

% of

Category#

% of

Category#

DRUG SELECTION 23 4.2% 100 18.4% 246 45.2% 148 27.2% 27 5.0% 544

OVER OR

UNDERDOSE

PRESCRIBED

15 3.2% 47 10.1% 237 51.0% 148 31.8% 18 3.9% 465

COMPLIANCE 3 1.1% 24 8.7% 146 52.9% 87 31.5% 16 5.8% 276

UNTREATED

INDICATIONS4 1.0% 13 3.4% 232 60.6% 127 33.2% 7 1.8% 383

MONITORING 3 6.4% 9 19.1% 16 34.0% 18 38.3% 1 2.1% 47

EDUCATION OR

INFORMATION10 2.4% 149 35.6% 198 47.4% 58 13.9% 3 0.7% 418

TOXICITY OR

ADVERSE

REACTION

11 4.2% 36 13.7% 116 44.1% 90 34.2% 10 3.8% 263

TOTAL 69 2.9% 378 15.8% 1191 49.7% 676 28.2% 82 3.4% 2396

INT CATEGORY

SevereModerateMildLowNil

Table 5.8-15: Clinical Significance Of Interventions Compared To Category

As would be expected, the type of action taken to investigate the problem was not specifically related

to the clinical significance of the intervention (see Table 5.8-17)

#% of

Significance#

% of

Significance#

% of

Significance#

% of

Significance#

% of

Significance

Nil 32 2.4% 13 1.7% 46 2.3% 11 13.4% 102 2.5%

Low 170 12.6% 79 10.6% 279 14.2% 20 24.4% 548 13.2%

Mild 653 48.5% 339 45.3% 980 49.9% 28 34.1% 2000 48.3%

Moderate 430 31.9% 266 35.6% 591 30.1% 19 23.2% 1306 31.5%Severe 61 4.5% 51 6.8% 68 3.5% 4 4.9% 184 4.4%

1346 100.0% 748 100.0% 1964 100.0% 82 100.0% 4140 100.0%Total

Investigation

ActionContact Prescriber

Discussion with

PatientSIGNIFICANCE

Other Actions Total Actions

32.5% 18.1% 47.4% 2.0% 100.0%

Table 5.8-16: Clinical Significance Of Interventions Compared To Investigative Actions

Interventions where a referral was recommended and interventions where monitoring was

recommended were also associated with moderate or severe level clinical significance interventions

(see shaded portion of Table 5.8-17)


Page 190 of 361

Table 5.8-17: Clinical Significance Of Interventions Compared To Recommendations Made

There was no relationship between the clinical significance of the intervention and whether or not the

recommendations made were accepted (see

Table 5.8-18).

Table 5.8-18: Clinical Significance Of Interventions Compared To Acceptance Of Recommendation (Outcome)

5.8.7 Proactive vs Reactive Clinical Interventions

Overall, 1917 (80%) of the interventions were classified as proactive by the pharmacist recording the

intervention (see Table 5.8-19).

Proactive or Reactive Number % of Total

Proactive 1917 80.0%

Reactive 468 19.5%

Not Specified 11 0.5%

Total 2396 100.0%

Table 5.8-19: Number Of Proactive And Reactive Clinical Interventions

# % # % # % # % # %Nil 63 3.2% 4 1.7% 2 1.3% 69 2.9%

Low 331 16.8% 25 10.6% 5 12.5% 17 11.0% 378 15.8%

Mild 966 49.1% 125 53.2% 21 52.5% 79 51.3% 1191 49.7%

Moderate 538 27.4% 73 31.1% 12 30.0% 53 34.4% 676 28.2%

High 69 3.5% 8 3.4% 2 5.0% 3 1.9% 82 3.4%

1967 100.0% 235 100.0% 40 100.0% 154 100.0% 2396 100.0%Total

Total

82.1% 9.8% 1.7% 6.4% 100.0%

SIGNIFICANCEUnknownNot Accepted

Partially

AcceptedAccepted

#% of

Recs#

% of

Recs#

% of

Recs#

% of

Recs#

% of

Recs#

% of

Recs

Nil 39 2.6% 4 0.7% 24 2.2% 6 3.6% 4 11.1% 77 2.3%

Low 155 10.5% 36 6.6% 187 16.9% 24 14.5% 18 50.0% 420 12.6%

Mild 716 48.5% 268 49.4% 578 52.3% 74 44.6% 10 27.8% 1646 49.5%

Moderate 507 34.3% 206 38.0% 287 26.0% 49 29.5% 3 8.3% 1052 31.6%

Severe 60 4.1% 28 5.2% 29 2.6% 13 7.8% 1 2.8% 131 3.9%

1477 100.0% 542 100.0% 1105 100.0% 166 100.0% 36 100.0% 3326 100.0%

33.2% 5.0% 1.1% 100.0%

Referral Rec'nChange in

Therapy Rec'nSIGNIFICANCE

Total44.4% 16.3%

TotalNo Rec'n

Made

Monitoring

Rec'n

Information

Rec'n


Page 191 of 361

The proportion of proactive and reactive interventions for each of the categories and subcategories of

clinical interventions are shown in

Table 5.8-20.

As can be seen, drug selection and dosage problems were more likely to be proactive, and

compliance and education or information type interventions were less likely to be proactive. This is

consistent with the fact that a number of patients directly request information or indicate compliance

issues. Within the intervention category of toxicity, there was a variation in the proportion of

proactiveness. Interventions where a drug interaction prompted the detection of the toxicity problem

were almost universally proactive, while those interventions where the toxicity was evident were

commonly reactive, as the patient may have brought the symptoms to the pharmacist’s attention.

Type Description Subtype Description Reactive Proactive%

Proactive

%

Proactive

Drug selection Duplication 13 70 84.3%

Drug selection Drug interaction 6 52 89.7%

Drug selection Wrong drug 14 105 88.2%

Drug selection Wrong dosage form 11 87 88.8%

Drug selection Other drug selection problem 10 176 94.6%

Over or underdose prescribed Dose too high 13 165 92.7%

Over or underdose prescribed Dose too low 15 154 91.1%

Over or underdose prescribed Other Dose Problem 11 107 90.7%

Compliance Taking too little 19 98 83.8%

Compliance Taking too much 10 37 78.7%

Compliance Intentional drug misuse 2 10 83.3%

Compliance Difficulty using dosage form 17 27 61.4%

Compliance Other Compliance Problem 19 35 64.8%

Untreated indicationsCondition not adequately

treated22 75 77.3%

Untreated indications Preventive therapy required 8 255 97.0%

Untreated indicationsOther Untreated indication

Problem5 14 73.7%

Monitoring Laboratory Monitoring 1 14 93.3%

Monitoring Non-Laboratory monitoring 5 18 78.3%

Monitoring Other Monitoring Problem 2 7 77.8%

Education or Information Patient drug information

request65 22 25.3%

Education or Information Confusion about therapy 42 77 64.7%

Education or Information Demonstration of device 16 46 74.2%

Education or Information Disease management or

advice29 60 67.4%

Education or Information Other Education or

Information Problem21 36 63.2%

Toxicity or Adverse reaction Toxicity caused by dose 7 10 58.8%

Toxicity or Adverse reaction Toxicity caused by drug

interaction1 86 98.9%

Toxicity or Adverse reaction Toxicity evident 74 54 42.2%

Toxicity or Adverse reaction Other Toxicity/Adverse Effect

problem11 20 64.5%

469 1917

90.1%

91.6%

75.5%

90.8%

83.0%

58.2%

64.6%

Total 80.3%

Table 5.8-20: Proactiveness Of Intervention Compared To Category And Subcategory


Page 192 of 361

There did not seem to be any relationship between proactiveness of intervention and the action that

was taken by the pharmacist to investigate the problem (see Table 5.8-21).

# % # % # % # % # %

Reactive 246 18.4% 76 10.2% 396 20.2% 19 23.5% 737 17.9%

Proactive 1091 81.6% 670 89.8% 1563 79.8% 62 76.5% 3386 82.1%

1337 100.0% 746 100.0% 1959 100.0% 81 100.0% 4123 100.0%Total

Investigation

Action

Contact

Prescriber

Discuss with

PatientProactivenessOther Actions Total

32.4% 18.1% 47.5% 2.0% 100.0%

Table 5.8-21: Proactiveness Of Intervention Compared To Actions To Investigate The Problem

There also did not seem to be any relationship between the the proactiveness of the intervention and

its significance as rated by the pharmacist who was documenting the event (see Table 5.8-22).

SIGNIFICANCE Reactive Proactive%

ProactiveNil 8 61 88.4%

Low 110 264 70.6%

Mild 229 957 80.7%

Moderate 109 566 83.9%

Severe 13 69 84.1%

Total 469 1917 80.3%

Table 5.8-22: Proactiveness Of Intervention Compared To Clinical Significance

5.8.8 Drugs Involved

There were 2,021 of the different individual drugs involved with the 2,396 clinical interventions (see

Table 5.8-23). A large number of the interventions were related to drugs for diabetes, many of these

interventions were prompted by the automated reminder discussed in section 5.8.14.

A wide range of drugs involved indicates that at least some types of interventions being performed

related to many different groups of drugs. It should be noted at this stage, however, that each

intervention is listed in the database as being associated with the dispensed drug, although other

drugs may be associated with the intervention. An example of an intervention where multiple drugs are

involved is shown in Figure 5.8-6.


Page 193 of 361

Drug selection - drug interaction Summary Problem: patient signs of toxicity whilst taking paroxetine, tramadol and methadone. Female patient (21-65 yo) currently taking paroxetine, tramadol and methadone. Reports symptoms of sweating and anxiety to their pharmacist. The patient had noticed these symptoms after commencing the paroxetine. The pharmacist referred the patient to their prescriber. The pharmacist was unsure if current symptoms were a sign of increased serotonin or under dose of methadone. The pharmacist recommended that the patient stop taking the tramadol and see their prescriber to discuss the current symptoms. Outcome: cause of symptoms in this patient to be fully explored by their prescriber. Category Drug selection Subcategory Drug interaction Actions Investigation: Patient History Investigation: Software Discussion with patient or carer Recommendations Refer to prescriber Drug change Education/counselling session Outcome Accepted Significance High

Figure 5.8-6: Example of Clinical Intervention Involving Multiple Drugs

5.8.8.1 Number of Clinical Interventions

When the drugs involved in the interventions are considered by generic drug name (see

Table 5.8-24), the most common drug involved was the antidiabetic agent, metformin (associated with

153 of the interventions). Again, many of these are related to the aspirin prophylaxis prompt (see

section 5.8.14). Notwithstanding the frequency of metformin and other antidiabetic agents, other

medications such as amoxicillin and salbutamol were associated with a significant number of

interventions.


Page 194 of 361

Drug Brand and Strength

Number of

Clinical

Interventions

% of Total

DIAFORMIN TAB 500mg 38 1.88%

DIABEX 1000 TAB 1000mg 90 29 1.43%

VENTOLIN CFC FREE MET-AERO 200 Dose 27 1.34%

METFORMIN TAB 500mg 24 1.19%

DIABEX TAB 500mg 23 1.14%

FOSAMAX ONCE WEEKLY TAB 70mg 21 1.04%

DIAMICRON TAB 80mg 20 0.99%

PANAMAX TAB 500mg 100 16 0.79%

DIAMICRON MR TAB 30mg 16 0.79%

NEXIUM EC-TABS 20mg 16 0.79%

NEXIUM EC-TABS 40mg 14 0.69%

SERETIDE 250/25 CFC FREE MET-AERO 250mcg/2 13 0.64%

LIPITOR TAB 20mg 13 0.64%

DAONIL TAB 5mg 13 0.64%

MOBIC TAB 15mg 13 0.64%

SPIRIVA INH-CAP 18mcg 13 0.64%

PREDNISOLONE TAB 25mg 12 0.59%

SOMAC EC-TABS 40mg 12 0.59%

COVERSYL TAB 4mg 11 0.54%

COVERSYL TAB 2mg 11 0.54%

COVERSYL PLUS TAB 4mg/1.25mg 11 0.54%

SERETIDE 250/50 A-HALER 250mcg/50mcg 60d 11 0.54%

VOLTAREN EC-TABS 50mg 11 0.54%

AVAPRO HCT TAB 300mg/12.5mg 11 0.54%

TEMAZE TAB 10mg 10 0.49%

EFEXOR XR SR-CAP 75mg 10 0.49%

BRICANYL T-HALER 500mcg 10 0.49%

PREDNISOLONE TAB 5mg 9 0.45%

GLICLAZIDE TAB 80mg 9 0.45%

PANADEINE FORTE TAB 500-30mg 20 9 0.45%

CEPHALEXIN CAP 500mg 9 0.45%

SALBUTAMOL CFC FREE MET-AERO 200 Dose 8 0.40%

PANAMAX 240 O-LIQ 240mg/5mL 200mL 8 0.40%

NORVASC TAB 5mg 8 0.40%

SYMBICORT TURBHAL 200mcg/6mcg 120d 8 0.40%

LOSEC TAB 20mg (base) 8 0.40%

ASTRIX TAB 100mg 112 8 0.40%

AMOXYCILLIN CAP 500mg 8 0.40%

PARIET EC-TABS 20mg 8 0.40%

OTHERS 1492 73.82%

TOTAL 2021 100.00%

Table 5.8-23: Specific Drugs Involved In Clinical Interventions

The medications involved in clinical interventions can be grouped using a multilevel anatomical

therapeutic category (ATC) classification code. The groupings of the drugs involved at four different

levels are shown in

Table 5.8-25 to Table 5.8-28.


Page 195 of 361

Number Percent

A10BA02 Metformin 153 6.7

J01CA04 Amoxicillin 73 3.2

A10BB09 Gliclazide 53 2.3

R03AC02 Salbutamol 53 2.3

B01AC06 Acetylsalicylic acid 47 2.1

H02AB06 Prednisolone 44 1.9

A02BC05 Esomeprazole 40 1.8

R03AK06 Salmeterol and other drugs for obstructive airway diseases 38 1.7

C10AA01 Simvastatin 37 1.6

N02BE01 Paracetamol 37 1.6

M01AB05 Diclofenac 37 1.6

C10AA05 Atorvastatin 35 1.5

C09AA04 Perindopril 30 1.3

N02AX02 Tramadol 28 1.2

J01DB01 Cefalexin 28 1.2

M05BA04 Alendronic acid 24 1.1

M01AC06 Meloxicam 24 1.1

J01FA01 Erythromycin 24 1.1

J01FA06 Roxithromycin 23 1.0

J01CR02 Amoxicillin and enzyme inhibitor 23 1.0

N02AA59 Codeine, combinations excl. psycholeptics 23 1.0

C09DA04 Irbesartan and diuretics 22 1.0

N06AX16 Venlafaxine 22 1.0

G03AA07 Levonorgestrel and estrogen 20 0.9

A10BB01 Glibenclamide 20 0.9

C03CA01 Furosemide 19 0.8

A02BC02 Pantoprazole 19 0.8

N05CD07 Temazepam 19 0.8

B01AA03 Warfarin 19 0.8

C09AA05 Ramipril 19 0.8

P01AB01 Metronidazole 18 0.8

C08CA01 Amlodipine 18 0.8

A02BC01 Omeprazole 18 0.8

C09CA04 Irbesartan 17 0.7

J01DC04 Cefaclor 16 0.7

R03BA05 Fluticasone 16 0.7

N06AB06 Sertraline 16 0.7

J01AA02 Doxycycline 15 0.7

R03BB04 Tiotropium bromide 15 0.7

J01EA01 Trimethoprim 15 0.7

Others 1068 46.7

Total 2285 100.0

ATC Code

(L5)DESCRIPTION

Clinical Interventions

Table 5.8-24: Generic Drugs Involved In Clinical Interventions (By WHO ATC Codes Level 5)


Page 196 of 361

Number Percent

A10BA Biguanides 153 6.7

A02BC Proton pump inhibitors 91 4.0

A10BB Sulfonamides, urea derivatives 91 4.0

C10AA HMG CoA reductase inhibitors 83 3.6

J01CA Penicillins with extended spectrum 73 3.2

R03AC Selective beta-2-adrenoreceptor agonists 69 3.0

C09AA ACE inhibitors, plain 63 2.8

J01FA Macrolides 59 2.6

B01AC Platelet aggregation inhibitors excl. heparin 58 2.5

N06AB Selective serotonin reuptake inhibitors 54 2.4

R03AK Adrenergics and other drugs for obstructive airway diseases 51 2.2

H02AB Glucocorticoids 48 2.1

N02AA Natural opium alkaloids 45 2.0

M01AB Acetic acid derivatives and related substances 44 1.9

C08CA Dihydropyridine derivatives 40 1.8

N02BE Anilides 38 1.7

N06AX Other antidepressants 34 1.5

C09CA Angiotensin II antagonists, plain 34 1.5

C09DA Angiotensin II antagonists and diuretics 31 1.4

M01AC Oxicams 29 1.3

M05BA Bisphosphonates 29 1.3

J01DB First-generation cephalosporins 28 1.2

N02AX Other opioids 28 1.2

R03BA Glucocorticoids 27 1.2

C07AB Beta blocking agents, selective 25 1.1

N05BA Benzodiazepine derivatives 24 1.1

N06AA Non-selective monoamine reuptake inhibitors 23 1.0

J01CR Combinations of penicillins, incl. beta-lactamase inhibitors 23 1.0

G03AA Progestogens and estrogens, fixed combinations 23 1.0

D07AC Corticosteroids, potent (group III) 23 1.0

J01AA Tetracyclines 22 1.0

P01AB Nitroimidazole derivatives 21 0.9

C01DA Organic nitrates 21 0.9

N05CD Benzodiazepine derivatives 21 0.9

G03CA Natural and semisynthetic estrogens, plain 20 0.9

R03BB Anticholinergics 20 0.9

C09BA ACE inhibitors and diuretics 20 0.9

B01AA Vitamin K antagonists 19 0.8

C03CA Sulfonamides, plain 19 0.8

A02BA H2-receptor antagonists 19 0.8

S01ED Beta blocking agents 17 0.7

Others 625 27.4

Total 2285 100.0

ATC Code

(L4)DESCRIPTION

Clinical



Page 197 of 361

Number PercentA10B ORAL BLOOD GLUCOSE LOWERING DRUGS 261 11.4R03A ADRENERGICS, INHALANTS 120 5.3J01C BETA-LACTAM ANTIBACTERIALS, PENICILLINS 119 5.2


OESOPHAGEAL REFLUX DISEASE (GORD)112 4.9

N06A ANTIDEPRESSANTS 112 4.9

M01AANTIINFLAMMATORY AND ANTIRHEUMATIC

PRODUCTS, NON-STEROIDS98 4.3

C10A CHOLESTEROL AND TRIGLYCERIDE REDUCERS 94 4.1B01A ANTITHROMBOTIC AGENTS 78 3.4N02A OPIOIDS 73 3.2C09A ACE INHIBITORS, PLAIN 63 2.8

J01FMACROLIDES, LINCOSAMIDES AND

STREPTOGRAMINS61 2.7

H02A CORTICOSTEROIDS FOR SYSTEMIC USE, PLAIN 48 2.1

R03BOTHER DRUGS FOR OBSTRUCTIVE AIRWAY

DISEASES, INHALANTS47 2.1

N02B OTHER ANALGESICS AND ANTIPYRETICS 46 2.0J01D OTHER BETA-LACTAM ANTIBACTERIALS 44 1.9

C08CSELECTIVE CALCIUM CHANNEL BLOCKERS WITH

MAINLY VASCULAR EFFECTS40 1.8

C07A BETA BLOCKING AGENTS 40 1.8C09C ANGIOTENSIN II ANTAGONISTS, PLAIN 34 1.5G03A HORMONAL CONTRACEPTIVES FOR SYSTEMIC USE 34 1.5C09D ANGIOTENSIN II ANTAGONISTS, COMBINATIONS 31 1.4

M05BDRUGS AFFECTING BONE STRUCTURE AND

MINERALIZATION29 1.3

N03A ANTIEPILEPTICS 28 1.2D07A CORTICOSTEROIDS, PLAIN 28 1.2N05C HYPNOTICS AND SEDATIVES 26 1.1S01E ANTIGLAUCOMA PREPARATIONS AND MIOTICS1) 26 1.1J01E SULFONAMIDES AND TRIMETHOPRIM 25 1.1N05B ANXIOLYTICS 24 1.1J01A TETRACYCLINES 22 1.0

P01AAGENTS AGAINST AMOEBIASIS AND OTHER

PROTOZOAL DISEASES21 0.9

C01D VASODILATORS USED IN CARDIAC DISEASES 21 0.9N05A ANTIPSYCHOTICS 21 0.9G03C ESTROGENS 20 0.9C09B ACE INHIBITORS, COMBINATIONS 20 0.9C03C HIGH-CEILING DIURETICS 19 0.8

C08DSELECTIVE CALCIUM CHANNEL BLOCKERS WITH

DIRECT CARDIAC EFFECTS18 0.8

J07B VIRAL VACCINES 16 0.7

OTHER 366 16.0

TOTAL 2285 100.0

ATC Code

(L3)DESCRIPTION

Clinical

Interventions



Page 198 of 361

Number PercentJ01 ANTIBACTERIALS FOR SYSTEMIC USE 288 12.6A10 DRUGS USED IN DIABETES 276 12.1R03 DRUGS FOR OBSTRUCTIVE AIRWAY DISEASES 169 7.4C09 AGENTS ACTING ON THE RENIN-ANGIOTENSIN SYSTEM 148 6.5N02 ANALGESICS 124 5.4N06 PSYCHOANALEPTICS 123 5.4A02 DRUGS FOR ACID RELATED DISORDERS 112 4.9M01 ANTIINFLAMMATORY AND ANTIRHEUMATIC PRODUCTS 98 4.3C10 SERUM LIPID REDUCING AGENTS 94 4.1B01 ANTITHROMBOTIC AGENTS 78 3.4

G03 SEX HORMONES AND MODULATORS OF THE GENITAL SYSTEM 74 3.2

N05 PSYCHOLEPTICS 71 3.1C08 CALCIUM CHANNEL BLOCKERS 58 2.5S01 OPHTHALMOLOGICALS 57 2.5H02 CORTICOSTEROIDS FOR SYSTEMIC USE 48 2.1C01 CARDIAC THERAPY 40 1.8C07 BETA BLOCKING AGENTS 40 1.8C03 DIURETICS 36 1.6P01 ANTIPROTOZOALS 32 1.4D07 CORTICOSTEROIDS, DERMATOLOGICAL PREPARATIONS 30 1.3M05 DRUGS FOR TREATMENT OF BONE DISEASES 29 1.3N03 ANTIEPILEPTICS 28 1.2J07 VACCINES 19 0.8

OTHERS 213 9.3TOTAL 2285 100.0

ATC Code

(L2)DESCRIPTION

Clinical

Interventions


Number Percent

A ALIMENTARY TRACT AND METABOLISM 443 19.4

C CARDIOVASCULAR SYSTEM 424 18.6

N NERVOUS SYSTEM 368 16.1

J ANTIINFECTIVES FOR SYSTEMIC USE 315 13.8

R RESPIRATORY SYSTEM 175 7.7

M MUSCULO-SKELETAL SYSTEM 145 6.3

B BLOOD AND BLOOD FORMING ORGANS 93 4.1

G GENITO URINARY SYSTEM AND SEX HORMONES 85 3.7

S SENSORY ORGANS 67 2.9

H SYSTEMIC HORMONAL PREPARATIONS, EXCL. 59 2.6

D DERMATOLOGICALS 41 1.8

P ANTIPARASITIC PRODUCTS, INSECTICIDES AND REPELLENTS 32 1.4

NO ATC CODE 16 0.7

L ANTINEOPLASTIC AND IMMUNOMODULATING AGENTS 12 0.5

V VARIOUS 10 0.4

TOTAL 2285 100.0

DESCRIPTIONATC Code

(L1)

Clinical

Interventions



Page 199 of 361

As can be seen from the tables above, drugs commonly associated with clinical interventions were in

the groups of antibiotics, drugs for diabetes, cardiovascular drugs and drugs for respiratory disorders.

5.8.8.2 Rate of Clinical Interventions

Although some conclusions can be drawn from the frequency of interventions associated with different

generic drugs and drug groups, it is more appropriate to consider the frequency of interventions in

relation to the frequency of prescriptions for those drugs.

Table 5.8-29 shows the rate (number of clinical interventions per 100 prescriptions) of clinical

interventions for individual generic drugs where more than 10 interventions were recorded against that

particular generic drug. As can be seen, individual generic drugs with a high frequency of interventions

included glibenclamide, gliclazide, aspirin, metformin and metronidazole. The presence of aspirin and

the antidiabetic agents in this list indicates the impact of the aspirin prophylaxis prompt.


Page 200 of 361

ATC Code

(L5)DESCRIPTION

Clinical

InterventionsPrescriptions

Clinical

Intervention

RateA10BB01 Glibenclamide 20 878 2.28

A10BB09 Gliclazide 53 2391 2.22

B01AC06 Acetylsalicylic acid 47 2546 1.85

A10BA02 Metformin 153 8556 1.79

P01AB01 Metronidazole 18 1027 1.75

J01FA01 Erythromycin 24 1444 1.66

H02AB06 Prednisolone 44 3428 1.28

R03BA05 Fluticasone 16 1294 1.24

J01EA01 Trimethoprim 15 1218 1.23

M01AB05 Diclofenac 37 3358 1.10

R03BB04 Tiotropium bromide 15 1822 0.82

R03AK06Salmeterol and other drugs for obstructive

airway diseases38 5171 0.73

C01DA02 Glyceryl trinitrate 14 2028 0.69

M05BA04 Alendronic acid 24 3523 0.68

N06AX16 Venlafaxine 22 3245 0.68

C03CA01 Furosemide 19 2832 0.67

N02AX02 Tramadol 28 4174 0.67

M01AC06 Meloxicam 24 3684 0.65

R03AC02 Salbutamol 53 8522 0.62

J01CA04 Amoxicillin 73 11816 0.62

A02BC05 Esomeprazole 40 6535 0.61

G03AA07 Levonorgestrel and estrogen 20 3546 0.56

J01DC04 Cefaclor 16 2929 0.55

J01AA02 Doxycycline 15 2857 0.53

A02BC02 Pantoprazole 19 3708 0.51

D07AC01 Betamethasone 14 2740 0.51

J01FA06 Roxithromycin 23 4641 0.50

C09AA04 Perindopril 30 6183 0.49

J01CR02 Amoxicillin and enzyme inhibitor 23 4802 0.48

J01DB01 Cefalexin 28 6383 0.44

N02BE01 Paracetamol 37 9151 0.40

B01AA03 Warfarin 19 4907 0.39

A02BC01 Omeprazole 18 4679 0.38

C09DA04 Irbesartan and diuretics 22 6116 0.36

C08CA01 Amlodipine 18 5218 0.34

C09AA05 Ramipril 19 5590 0.34

N05CD07 Temazepam 19 5852 0.32

N06AB06 Sertraline 16 5063 0.32

C10AA01 Simvastatin 37 11845 0.31

N02AA59 Codeine, combinations excl. psycholeptics 23 7657 0.30

C10AA05 Atorvastatin 35 14446 0.24

C09CA04 Irbesartan 17 7052 0.24

Others 1027 179137 0.57

Total 2272 383994 0.59

Table 5.8-29: Rate Of Clinical Interventions For Individual Generic Drugs (ATC Level 5)


Page 201 of 361

Once again, it is possible to group the generic drugs into different levels of anatomical and therapeutic

categories in order to obtain information regarding groups of agents, rather than individual generic

drugs.

The information on frequency of clinical interventions compared to number of prescriptions for

particular groups of drugs are shown from Table 5.8-30 to Table 5.8-32. As before, the information in

these tables only includes those groups of drugs where more than 10 interventions were recorded

against the particular drug group.

ATC Code

(L4)DESCRIPTION

Clinical

InterventionsPrescriptions Rate

A10BG Thiazolidinediones 14 588 2.38A10BB Sulfonamides, urea derivatives 91 4537 2.01A10BA Biguanides 153 8556 1.79C01AA Digitalis glycosides 13 775 1.68P01AB Nitroimidazole derivatives 21 1295 1.62M01AE Propionic acid derivatives 16 1164 1.37R03BA Glucocorticoids 27 2112 1.28J01EA Trimethoprim and derivatives 15 1218 1.23N03AG Fatty acid derivatives 12 986 1.22S01BA Corticosteroids, plain 11 934 1.18J01CF Beta-lactamase resistant penicillins 15 1284 1.17M01AB Acetic acid derivatives and related substances 44 4222 1.04J01MA Fluoroquinolones 11 1179 0.93H02AB Glucocorticoids 48 5149 0.93B01AC Platelet aggregation inhibitors excl. heparin 58 6627 0.88

R03AKAdrenergics and other drugs for obstructive airway

diseases51 5983 0.85

R03BB Anticholinergics 20 2426 0.82J01FA Macrolides 59 7205 0.82H03AA Thyroid hormones 11 1369 0.80S01ED Beta blocking agents 17 2361 0.72R03AC Selective beta-2-adrenoreceptor agonists 69 9683 0.71N06AX Other antidepressants 34 4997 0.68N02AX Other opioids 28 4174 0.67M01AC Oxicams 29 4326 0.67C03CA Sulfonamides, plain 19 2856 0.67C01DA Organic nitrates 21 3237 0.65M05BA Bisphosphonates 29 4633 0.63J01CA Penicillins with extended spectrum 73 11819 0.62C07AA Beta blocking agents, non-selective 12 1943 0.62N06AA Non-selective monoamine reuptake inhibitors 23 3910 0.59G03CA Natural and semisynthetic estrogens, plain 20 3449 0.58J01AA Tetracyclines 22 3922 0.56J01DC Second-generation cephalosporins 16 2929 0.55A02BA H2-receptor antagonists 19 3740 0.51

Others* 696 192273 0.36

1817 317861 0.57*Based on only drug groups with >10 interventions recorded

Total*

Table 5.8-30: Rate Of Clinical Interventions For Drug Groups (ATC Level 4)


Page 202 of 361

ATC

Code (L3)DESCRIPTION

Clinical


B03A IRON PREPARATIONS 11 257 4.28A10B ORAL BLOOD GLUCOSE LOWERING DRUGS 261 13846 1.89A10A INSULINS AND ANALOGUES 15 884 1.70C01A CARDIAC GLYCOSIDES 13 775 1.68

P01AAGENTS AGAINST AMOEBIASIS AND OTHER

PROTOZOAL DISEASES21 1295 1.62

J01E SULFONAMIDES AND TRIMETHOPRIM 25 2006 1.25S01B ANTIINFLAMMATORY AGENTS 11 995 1.11N04B DOPAMINERGIC AGENTS 13 1235 1.05

R03BOTHER DRUGS FOR OBSTRUCTIVE AIRWAY

DISEASES, INHALANTS47 4701 1.00

P01B ANTIMALARIALS 11 1169 0.94J01M QUINOLONE ANTIBACTERIALS 11 1179 0.93

H02A CORTICOSTEROIDS FOR SYSTEMIC USE, PLAIN 48 5178 0.93

J01FMACROLIDES, LINCOSAMIDES AND

STREPTOGRAMINS61 7282 0.84

H03A THYROID PREPARATIONS 11 1369 0.80N03A ANTIEPILEPTICS 28 3585 0.78R03A ADRENERGICS, INHALANTS 120 15666 0.77

M01AANTIINFLAMMATORY AND ANTIRHEUMATIC

PRODUCTS, NON-STEROIDS98 13159 0.74

B01A ANTITHROMBOTIC AGENTS 78 11585 0.67C03C HIGH-CEILING DIURETICS 19 2857 0.67C01D VASODILATORS USED IN CARDIAC DISEASES 21 3246 0.65J01C BETA-LACTAM ANTIBACTERIALS, PENICILLINS 119 18746 0.63

M05BDRUGS AFFECTING BONE STRUCTURE AND

MINERALIZATION29 4655 0.62

M04A ANTIGOUT PREPARATIONS 13 2116 0.61G03C ESTROGENS 20 3449 0.58J01A TETRACYCLINES 22 3922 0.56N06A ANTIDEPRESSANTS 112 22066 0.51


OESOPHAGEAL REFLUX DISEASE (GORD)112 22177 0.51

N02A OPIOIDS 73 14657 0.50J01D OTHER BETA-LACTAM ANTIBACTERIALS 44 9312 0.47S01A ANTIINFECTIVES 11 2330 0.47

G03AHORMONAL CONTRACEPTIVES FOR SYSTEMIC

USE34 7266 0.47

N02B OTHER ANALGESICS AND ANTIPYRETICS 46 11055 0.42

C08DSELECTIVE CALCIUM CHANNEL BLOCKERS

WITH DIRECT CARDIAC EFFECTS18 4513 0.40

C09B ACE INHIBITORS, COMBINATIONS 20 5038 0.40

S01EANTIGLAUCOMA PREPARATIONS AND

MIOTICS1)26 7005 0.37

C09D ANGIOTENSIN II ANTAGONISTS, COMBINATIONS 31 8615 0.36

N05A ANTIPSYCHOTICS 21 5851 0.36J07B VIRAL VACCINES 16 4465 0.36C07A BETA BLOCKING AGENTS 40 11228 0.36C09A ACE INHIBITORS, PLAIN 63 17915 0.35

C08CSELECTIVE CALCIUM CHANNEL BLOCKERS

WITH MAINLY VASCULAR EFFECTS40 11614 0.34

D07A CORTICOSTEROIDS, PLAIN 28 8622 0.32

C10A CHOLESTEROL AND TRIGLYCERIDE REDUCERS 94 31183 0.30

N05C HYPNOTICS AND SEDATIVES 26 8763 0.30C09C ANGIOTENSIN II ANTAGONISTS, PLAIN 34 11822 0.29N05B ANXIOLYTICS 24 8716 0.28

Total* 2039 359370 0.57*Based on only drug groups with >10 interventions recorded



Page 203 of 361

ATC

Code

(L2)DESCRIPTION

Clinical


A10 DRUGS USED IN DIABETES 276 14730 1.87

A07ANTIDIARRHEALS, INTESTINAL

ANTIINFLAMMATORY/ANTIINFECTIVE 16 1096 1.46

P01 ANTIPROTOZOALS 32 2464 1.30B03 ANTIANEMIC PREPARATIONS 15 1218 1.23H02 CORTICOSTEROIDS FOR SYSTEMIC USE 48 5178 0.93N04 ANTI-PARKINSON DRUGS 13 1409 0.92R03 DRUGS FOR OBSTRUCTIVE AIRWAY DISEASES 169 20949 0.81N03 ANTIEPILEPTICS 28 3585 0.78C01 CARDIAC THERAPY 40 5198 0.77

M01ANTIINFLAMMATORY AND ANTIRHEUMATIC

PRODUCTS98 13171 0.74

H03 THYROID THERAPY 11 1552 0.71B01 ANTITHROMBOTIC AGENTS 78 11585 0.67J01 ANTIBACTERIALS FOR SYSTEMIC USE 288 42911 0.67M05 DRUGS FOR TREATMENT OF BONE DISEASES 29 4655 0.62C03 DIURETICS 36 5828 0.62M04 ANTIGOUT PREPARATIONS 13 2116 0.61N06 PSYCHOANALEPTICS 123 23139 0.53A02 DRUGS FOR ACID RELATED DISORDERS 112 22304 0.50

G03SEX HORMONES AND MODULATORS OF THE

GENITAL SYSTEM74 15201 0.49

N02 ANALGESICS 124 26981 0.46J07 VACCINES 19 4817 0.39S01 OPHTHALMOLOGICALS 57 14723 0.39C08 CALCIUM CHANNEL BLOCKERS 58 16127 0.36C07 BETA BLOCKING AGENTS 40 11228 0.36

C09AGENTS ACTING ON THE RENIN-ANGIOTENSIN

SYSTEM148 43390 0.34

D07CORTICOSTEROIDS, DERMATOLOGICAL

PREPARATIONS30 8875 0.34

N05 PSYCHOLEPTICS 71 23330 0.30C10 SERUM LIPID REDUCING AGENTS 94 31183 0.30

TOTAL* 2140 378943 0.56* Based on drug groups with >10 clinical Interventions


Again, the impact of the aspirin prophylaxis prompt is evident in the higher rates of clinical intervention

within the group of oral antidiabetic agents (see Table 5.8-31). However, this “assisted” intervention

was not directly prompted in patients receiving insulin, for which there is also a high rate of

interventions.

Other groups of drugs that seem associated with a higher rate of clinical interventions include

antiinfective agents, iron preparations, corticosteroids and drugs associated with respiratory diseases

(see Table 5.8-32).


Page 204 of 361

5.8.8.3 Nature of Clinical Interventions for Specific Groups of Drugs

Groups of drugs of interest (for example, groups with either a high number of interventions, a high rate

of interventions or drug groups that are known from the literature to have a high rate of drug-related

problems) can be examined for different characteristics. We examined the following drug groups in

terms of the category of intervention and the clinical significance as assigned by the recording

pharmacist:

• Oral Antidiabetic Agents

• Antibacterial Agents

• Agents for Obstructive Airways Diseases

• Systemic Corticosteroids

• Agents for Cardiac Therapy

• Antiinflammatory Agents

• Antithrombotic Agents

5.8.8.3.1 Oral Antidiabetic Agents

In

Table 5.8-33, the interventions relating to oral antidiabetic agents (ATC L2 Code A10) are examined.

Although a significant number (170) were associated with the aspirin prophylaxis prompt (and are

therefore coded as preventive therapy required), 35% of the interventions associated with this class of

medication were of other types. Problems with education or information, compliance or drug selection

occurred frequently in association with this group of medications.

If the prompted interventions are ignored, problems concerning patients taking too little antidiabetic

therapy, and problems relating to patients requiring disease management advice accounted for almost

30% of the interventions in this group of medications.


Page 205 of 361

INT

CODEINT CATEGORY INT SUB CATEGORY Number % %

D1 DUPLICATION 4 1.4%D2 DRUG INTERACTION 2 0.7%D3 WRONG DRUG 5 1.8%D4 WRONG DOSAGE FORM 11 4.0%

D0 OTHER DRUG SELECTION PROBLEM 1 0.4%

O1 DOSE TOO HIGH 2 0.7%

O2 DOSE TOO LOW 7 2.5%

O0 OTHER DOSE PROBLEM 6 2.2%

C1 TAKING TOO LITTLE 16 5.8%

C2 TAKING TOO MUCH 3 1.1%

C0 OTHER COMPLIANCE PROBLEM 4 1.4%U1 CONDITION NOT ADEQUATELY TREATED 5 1.8%U2 PREVENTIVE THERAPY REQUIRED 170 61.6%U0 OTHER UNTREATED INDICATION PROBLEM 5 1.8%

M3 MONITORING NON-LABORATORY MONITORING 1 0.4% 0.4%

E1 PATIENT DRUG INFORMATION REQUEST 1 0.4%

E2 CONFUSION ABOUT THERAPY 5 1.8%

E3 DEMONSTRATION OF DEVICE 3 1.1%

E4 DISEASE MANAGEMENT OR ADVICE 13 4.7%

E0 OTHER EDUCATION OR INFORMATION PROBLEM 8 2.9%

T1 DOSE RELATED 2 0.7%T2 CAUSED BY DRUG INTERACTION 1 0.4%T3 TOXICITY/ ADVERSE REACTION EVIDENT 1 0.4%

276 100.0% 100.0%

8.3%

5.4%

8.3%

65.2%

10.9%

1.4%

TOTAL

EDUCATION OR

INFORMATION

TOXICITY OR ADVERSE

REACTION

DRUG SELECTION

OVER OR UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

Table 5.8-33: Categories And Subcategories Of Interventions Associated With Oral Antidiabetic Agents

Clinical interventions associated with the oral antidiabetic agents did not differ from the overall dataset

of clinical interventions in terms of clinical significance (see Table 5.8-34 below and Figure 5.8-8 on

page 212)

SIGNIFICANCE Number %

Nil 5 1.8%

Low 16 5.8%

Mild 159 57.6%

Moderate 93 33.7%

Severe 3 1.1%Total 276 100.0%

Table 5.8-34: Clinical Significance Of Interventions Associated With Oral Antidiabetic Agents

5.8.8.3.2 Antibacterial Agents

Almost 300 interventions were recorded as associated with antibacterial agents (ATC L2 Code J01;

Table 5.8-32). These interventions are examined further in Table 5.8-35 and Table 5.8-36.

Interventions associated with antibacterial agents were represented in each category of interventions

available. Frequent interventions associated with antibacterial agents included drug selection

problems and dosage problems, particularly underdosage of these agents (see Table 5.8-35). Of the

39 interventions associated with under dosage of antibacterial agents, 15 were associated with

paediatric prescriptions.


Page 206 of 361

INT CATEGORY INT SUB CATEGORY Number % %DUPLICATION 4 1.4%

DRUG INTERACTION 13 4.5%

WRONG DRUG 12 4.2%

WRONG DOSAGE FORM 14 4.9%

OTHER DRUG SELECTION PROBLEM 48 16.7%

DOSE TOO HIGH 19 6.6%

DOSE TOO LOW 39 13.5%

OTHER DOSE PROBLEM 14 4.9%

TAKING TOO LITTLE 4 1.4%

TAKING TOO MUCH 1 0.3%

DIFFICULTY USING DOSAGE FORM 10 3.5%

OTHER COMPLIANCE PROBLEM 10 3.5%

CONDITION NOT ADEQUATELY TREATED 13 4.5%

PREVENTIVE THERAPY REQUIRED 8 2.8%

OTHER UNTREATED INDICATION PROBLEM 1 0.3%

LABORATORY MONITORING 2 0.7%

NON-LABORATORY MONITORING 1 0.3%

OTHER MONITORING PROBLEM 1 0.3%

PATIENT DRUG INFORMATION REQUEST 10 3.5%

CONFUSION ABOUT THERAPY 14 4.9%

DEMONSTRATION OF DEVICE 1 0.3%

DISEASE MANAGEMENT OR ADVICE 6 2.1%

OTHER EDUCATION OR INFORMATION PROBLEM 7 2.4%

CAUSED BY DRUG INTERACTION 8 2.8%

TOXICITY/ ADVERSE REACTION EVIDENT 14 4.9%

OTHER TOXICITY/ADVERSE EFFECT PROBLEM 14 4.9%

288 100.0% 100.0%

1.4%

13.2%

12.5%

DRUG SELECTION

OVER OR

UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

31.6%

25.0%

8.7%

7.6%

TOTAL

MONITORING

EDUCATION OR

INFORMATION

TOXICITY OR

ADVERSE

REACTION

Table 5.8-35: Categories And Subcategories Of Interventions Associated With Antibiotic Agents

A higher proportion of clinical interventions associated with antibiotic agents was rated as of potentially

severe clinical significance by the documenting pharmacist, compared with the proportion in the entire

PROMISe dataset (see Table 5.8-36 and Figure 5.8-8 on page 212).


Nil 16 5.6%

Low 52 18.1%

Mild 132 45.8%

Moderate 72 25.0%

Severe 16 5.6%Total 288 100.0%

Table 5.8-36: Clinical Significance Of Interventions Associated With Antibiotic Agents

5.8.8.3.3 Drugs for Obstructive Airways Disease

Drugs for obstructive airways disease were involved with 169 interventions (ATC L2 Code R03, Table

5.8-32). Interventions associated with these drugs were most commonly of the education and

information or compliance categories. One particularly frequent intervention was the demonstration of


Page 207 of 361

a device (43 interventions, 25.4% of interventions in this drug group), which is a common education

issue given the range of formulations available within this drug group.

INT CATEGORY INT SUB CATEGORY Number % %DUPLICATION 5 3.0%

WRONG DRUG 4 2.4%




DOSE TOO LOW 6 3.6%








OTHER UNTREATED INDICATION

PROBLEM1 0.6%





OTHER EDUCATION OR INFORMATION

PROBLEM2 1.2%

TOXICITY OR

ADVERSE

REACTION

TOXICITY/ ADVERSE REACTION EVIDENT 3 1.8% 1.8%

169 100.0% 100.0%

UNTREATED

INDICATIONS

EDUCATION OR

INFORMATION

TOTAL

14.2%

13.6%

22.5%

13.0%

34.9%

DRUG

SELECTION

OVER OR

UNDERDOSE

PRESCRIBED

COMPLIANCE

Table 5.8-37: Categories And Subcategories Of Interventions Associated With Drugs For Respiratory Disease

The overall significance of interventions associated with drugs for respiratory disease was lower than

the dataset average, as a result of many “mild” level significance interventions associated with

demonstration of dosage devices (see Table 5.8-37 and Figure 5.8-8 on page 212).


Nil 1 0.6%

Low 19 11.2%

Mild 100 59.2%

Moderate 47 27.8%

Severe 2 1.2%Total 169 100.0%

Table 5.8-38: Clinical Significance Of Interventions Associated With Drugs For Respiratory Disease


Page 208 of 361

5.8.8.3.4 Systemic Corticosteroids

There was a higher than average rate of clinical interventions associated with corticosteroids for

systemic use (ATC L2 Code H02, Table 5.8-32). These interventions are examined in more detail in

Table 5.8-39 and Table 5.8-40. The most common category of intervention associated with systemic

corticosteroids was dose problems (18 interventions, 37.5% of the total), with three of these

interventions being rated as highly clinically significant. Almost half (48%) of the clinical interventions

associated with corticosteroids were rated as either moderate or severe clinical significance, a rate

higher than that in the entire intervention dataset (see Figure 5.8-8). This may be due to pharmacists

considering that long-term preventive recommendations (common in this drug group) are highly

significant.

INT CATEGORY INT SUB CATEGORY Number % %WRONG DRUG 1 2.1%

WRONG DOSAGE FORM 1 2.1%OTHER DRUG SELECTION PROBLEM 2 4.2%

DOSE TOO HIGH 4 8.3%OTHER DOSE PROBLEM 14 29.2%

TAKING TOO LITTLE 3 6.3%INTENTIONAL DRUG MISUSE 1 2.1%

DIFFICULTY USING DOSAGE FORM 1 2.1%OTHER COMPLIANCE PROBLEM 1 2.1%

CONDITION NOT ADEQUATELY TREATED 1 2.1%PREVENTIVE THERAPY REQUIRED 5 10.4%

OTHER UNTREATED INDICATION PROBLEM 1 2.1%CONFUSION ABOUT THERAPY 5 10.4%

DISEASE MANAGEMENT OR ADVICE 1 2.1%OTHER EDUCATION OR INFORMATION

PROBLEM1 2.1%



48 100.0% 100.0%

OVER OR UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

TOTAL

EDUCATION OR

INFORMATION

TOXICITY OR ADVERSE

REACTION

8.3%

37.5%

12.5%

14.6%

14.6%

12.5%

DRUG SELECTION

Table 5.8-39: Categories And Subcategories Of Interventions Associated With Systemic Corticosteroids


Nil 0 0.0%

Low 4 8.3%

Mild 21 43.8%

Moderate 20 41.7%

Severe 3 6.3%

Total 48 100.0%

Table 5.8-40: Clinical Significance Of Interventions Associated With Systemic Corticosteroids

An example of a clinical intervention relating to this class of drugs is shown in Figure 5.8-7.


Page 209 of 361

Over or underdose prescribed - dose too high Summary Problem: high dose of prednisolone prescribed for 16kg child Prescription for prednisolone mixture 5mL three times a day for a 16kg child. The pharmacist contacted the prescriber and recommended reducing the dose of prednisolone to 1mL three times a day. The prescriber accepted the change as they were not aware that the prednisolone mixture was 5mg/mL and had assumed that it was 1mg/mL. Outcome: overdose of prednisolone avoided Category Over or underdose prescribed Subcategory Dose too high Actions Contacted prescriber Recommendations Dose change Outcome Accepted Significance High

Figure 5.8-7: Example of Clinical Intervention For Systemic Corticosteroids

5.8.8.3.5 Cardiac Therapy Drugs

Clinical interventions associated with drugs for cardiac therapy occurred at a higher than average rate

(ATC L2 Code C01, Table 5.8-32). Problems with dosage were common amongst this particular group

of medication is (11 interventions, 27.5% of total). A large proportion (over 50%) of the interventions in

this group of medications were rated as of either high or moderate clinical significance by the

recording pharmacists (see Table 5.8-42 and Figure 5.8-8).

INT CATEGORY INT SUB CATEGORY Number % %

DUPLICATION 3 7.5%

WRONG DRUG 2 5.0%




DOSE TOO LOW 2 5.0%






OTHER UNTREATED INDICATION PROBLEM 1 2.5%



OTHER MONITORING PROBLEM 1 2.5%






40 100.0% 100.0%

7.5%

12.5%

15.0%

DRUG SELECTION

OVER OR

UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

17.5%

27.5%

15.0%

5.0%

TOTAL

MONITORING

EDUCATION OR

INFORMATION

TOXICITY OR

ADVERSE REACTION

Table 5.8-41: Categories And Subcategories Of Interventions Associated With Drugs For Cardiac Therapy


Page 210 of 361


Nil 2 5.0%

Low 6 15.0%

Mild 9 22.5%

Moderate 17 42.5%

High 6 15.0%Total 40 100.0%

Table 5.8-42: Clinical Significance Of Interventions Associated With Drugs For Cardiac Therapy

5.8.8.3.6 Anti-inflammatory Agents

There were 98 clinical interventions recorded for drugs in the anti-inflammatory agent group (ATC L2

Code M01, Table 5.8-29). It was common for duplication of these agents to occur (14 interventions,

14.3% of total), or for toxicity relating to these agents to be evident (15 interventions, 15.3% of total).


DUPLICATION 14 14.3%


WRONG DRUG 1 1.0%




DOSE TOO LOW 2 2.0%







MONITORING NON-LABORATORY MONITORING 1 1.0% 1.0%





DOSE RELATED 3 3.1%




98 100.0% 100.0%

DRUG SELECTION

OVER OR

UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

EDUCATION OR

INFORMATION

TOXICITY OR

ADVERSE

REACTION

TOTAL

25.5%

12.2%

7.1%

7.1%

11.2%

35.7%

Table 5.8-43: Categories And Subcategories Of Interventions Associated With Anti-inflammatory Agents

The clinical significance of interventions associated with the anti-inflammatory agents is outlined in

Table 5.8-44 and compared to other drug groups in Figure 5.8-8.


Page 211 of 361


Nil 3 3.1%

Low 10 10.2%

Mild 49 50.0%

Moderate 32 32.7%

High 4 4.1%Total 98 100.0%

Table 5.8-44: Clinical Significance Of Interventions Associated With Anti-inflammatory Agents

5.8.8.3.7 Antithrombotic Agents

There were 78 interventions associated with antithrombotic agents (ATC L2 Code B01, Table 5.8-29).

Many of these interventions were related to the addition of prophylactic aspirin in diabetic patients at

high risk of cardiovascular events, an intervention highlighted by the automated prompt (23

interventions, 29.5% of total). Many of the other interventions in this category related to the use of

warfarin, an antithrombotic agent with a much narrower therapeutic margin than aspirin. As a result,

the clinical significance of interventions associated with drugs of this class was rated higher than that

of the entire PROMISe data set (see Table 5.8-46 and Figure 5.8-8 )


DUPLICATION 1 1.3%


WRONG DRUG 3 3.8%




DOSE TOO LOW 2 2.6%












TOXICITY/ ADVERSE REACTION EVIDENT 2 2.6%78 100.0% 100.0%

10.3%

9.0%

6.4%

DRUG SELECTION

OVER OR UNDERDOSE

PRESCRIBED

COMPLIANCE

UNTREATED

INDICATIONS

23.1%

7.7%

12.8%

30.8%

TOTAL

MONITORING

EDUCATION OR

INFORMATION

TOXICITY OR ADVERSE

REACTION

Table 5.8-45: Categories And Subcategories Of Interventions Associated With Antithrombotic Agents


Nil 2 2.6%

Low 9 11.5%

Mild 29 37.2%

Moderate 36 46.2%

High 2 2.6%Total 78 100.0%

Table 5.8-46: Clinical Significance Of Interventions Associated With Antithrombotic Agents


Page 212 of 361

5.8.8.3.8 Clinical Significance of Interventions Associated with Different Drug Groups

The recorded clinical significance of interventions associated with each of the seven groups of

medications outlined above is shown in Table 5.8-47 and Figure 5.8-8. The proportion of interventions

that were rated either moderate or severe level clinical significance was highest for drugs in the

cardiac therapy group (23 of 40; 57.5%).

# % # %C01 Cardiac Therapy 17 42.5% 23 57.5% 40B01 Antithrombotic Agents 40 51.3% 38 48.7% 78H01 Systemic Corticosteroids 25 52.1% 23 47.9% 48M01 Antiinflammatory Agents 62 63.3% 36 36.7% 98A10 Antidiabetic Agents 200 69.4% 88 30.6% 288R01 Respiratory Drugs 120 71.0% 49 29.0% 169

Entire PROMISe Dataset 1638 68.4% 758 31.6% 2396

Total

Lower (Nil. Low and

Mild)

Higher (Moderate

and Severe)Drug GroupATC Code

(L3)

Table 5.8-47: Clinical Significance Of Interventions Associated With Different Drug Groups

Entire PROMISeDataset

31.6% Higher Significance

Figure 5.8-8: Clinical Significance Of Interventions Associated With Different Drug Groups

The ability to selectively examine the nature and frequency of clinical interventions by drug groups can

be invaluable in determining educational priorities as well as targeting patient groups for clinical

activities.


Page 213 of 361

5.8.9 Effect of Remuneration (univariate analysis)

Remuneration was randomly assigned to each pharmacy for either the first or second phase of the

study, while remuneration for interventions was provided for all pharmacies during the third phase of

the study (see Figure 5.8-9).

PaymentPayment

PaymentPayment


PaymentPayment

Group A

Group B

Payment, $15 for each clinical intervention



N = 29

N = 23

N = 29

N = 23

N = 52

Figure 5.8-9: Remuneration Crossover for Different Phases of the Study

Using univariate analysis, remuneration had a significant effect on intervention rates (see Table

5.8-48) during Phase 1 of the study, but not during Phase 2. Interestingly, those pharmacies that had

the higher rate (possibly due to remuneration or other factors) during Phase 1 had a higher rate during

Phase 2 of the study. Although there may be multiple factors leading to this, one possibility is that

once a pharmacist gets into the routine of documenting, their rate of decline may be attenuated (see

Figure 5.8-10).



Paid Not Paid Mann-Whitney Paid Not Paid Mann-Whitney

Intervention Rate per 100 Prescriptions

1.09 (0.91-1.27)

0.71 (0.60-0.82)

Z= 2.153 p= 0.031

0.65 (0.52-0.79)

0.81 (0.65-0.97)

Z= 1.008 P= 0.313


1.83 (1.53-2.13)

1.22 (1.02-1.41)

Z= 2.159 p= 0.031

1.12 (0.90-1.34)

1.34 (1.09-1.60)

Z= 1.020 P= 0.308

Table 5.8-48: Effect of Remuneration in Different Phases of the Study


Page 214 of 361

Figure 5.8-10: Attenuation of reduction of intervention rate by Remuneration

5.8.10 Effect of Aspirin Intervention Prompt on Overall Clinical Intervention Rate (univariate analysis)

The aspirin intervention prompt was installed in the 31 of the 52 pharmacies for a period of

approximately 4 weeks during the study. There were 202 intervention prompted by the software tool

during the study, 201 of which where in pharmacies with the software prompt installed and during the

time the installation was active (see further discussion in section 5.8.14). The comparisons made in

this section relate to the overall clinical intervention rate (including prompted interventions) over the

entire eight weeks of the study.

The effect of the aspirin intervention prompt on the overall intervention rates, as measured per 100

prescriptions or per 100 patients, is shown in Figure 5.8-11. When considering the intervention rate

over the entire data collection period, there was no difference in the overall intervention rate (either as

per 100 prescriptions or per 100 patients) (see Figure 5.8-12 and Figure 5.8-13).


Page 215 of 361

.00 1.00

Aspirin

0.00

2.00

4.00

6.00

8.00

10

1

10

3

1

3


Clinical Interventions per 100 Patients

Figure 5.8-11: Effect of Aspirin Intervention Prompt on Overall Intervention Rates (1 = Aspirin Popup Installed, 0 = no Aspirin Popup)

No Statistically Significant Difference: Mann-Whitney U statistic = 243, p = 0.124

0 1 2 3 4 5


0

2

4

6

8

10

Fre

qu

en

cy

Mean = 0.456Std . D ev. = 0.44443N = 21

No Aspirin Intervention Prompt

0.0 0 1.00 2.00 3.00 4.00 5.00


0

2

4

6

8

10

12

14

Fre

qu

en

cy

Mean = 0.8635Std. D ev. = 0.99118N = 31

Aspirin Intervention Prompt Installed

Figure 5.8-12: Effect of Aspirin Intervention Prompt on Overall Intervention Rate Per 100 Prescriptions


Page 216 of 361

0.00 2.00 4.00 6.00 8.00


0

5

10

15

20

Fre

qu

en

cy

Mean = 1.467 6Std. D ev. = 1.58756N = 31

Aspirin Intervention Prompt Installed

0 1 2 3 4 5


0

2

4

6

8

10

Fre

qu

en

cy

Mean = 0.456Std. D ev. = 0.44443

N = 21

No Aspirin Intervention Prompt

No Statistically Significant Difference: Mann-Whitney U statistic = 253, p = 0.176

Figure 5.8-13 : Effect of Aspirin Intervention Prompt on Overall Intervention Rate Per 100 Patients

However, if the period of time when the aspirin intervention prompt was active is considered, there

was a significant impact of the aspirin prompt on the overall clinical intervention rate.

First Half (Phase 1 and 2) Mean (95% CI)

Second Half (Phase 3) Mean (95% CI)

Aspirin Pop-up Present

No Aspirin Pop-up

Mann-Whitney


No Aspirin Pop-up

Mann-Whitney


1.04 (0.93-1.15)

0.53 (0.45-0.60)

Z= 6.727 P<0.001

0.45 (0.35-0.55)

0.31 (0.25-0.38)

Z= 0.004 P= 0.997


1.74 (1.55-1.93)

0.91 (0.77-1.05)

Z= 6.700 P<0.001

0.80 (0.56-1.04)

0.55 (0.43-0.66)

Z= 0.029 P= 0.977

Table 5.8-49: Effect of Aspirin Intervention Prompt on Overall Intervention Rate in Different Phases of the Study

It is interesting to note that there were only 201 interventions in the first phase of the study that were

prompted by the electronic reminder (of 1949 interventions in these first two phases of the study).

Thus, the impact of the aspirin popup seems to increase the rate of interventions above that caused by

aspirin interventions alone (see analysis later in section 6 ). Again, although the reasons for this may

be multifactorial, one possibility is that the aspirin prompt reminds the dispensing pharmacist to

consider the possibility of an intervention, and this flows on to other prescriptions being dispensed.

Non-aspirin related interventions were separately examined using the unique unit of pharmacist-day,

with days where less than 20 prescriptions were dispensed by a particular pharmacist excluded. There

was a statistically significant difference in non-aspirin clinical intervention rates in the first half of the

study, when the aspirin popup was installed and active (see Table 5.8-50).


Page 217 of 361

First Half (Phase 1 and 2) Mean (95% CI)

Second Half (Phase 3) Mean (95% CI)


No Aspirin Pop-up

Mann-Whitney


No Aspirin Pop-up

Mann-Whitney


0.90 (0.80-1.00)

0.52 (0.44-0.60)

Z= 5.102 P<0.001

0.43 (0.33-0.53)

0.31 (0.24-0.38)

Z= 0.476 P= 0.632


1.49 (1.33-1.49)

0.90 (0.76-1.04)

Z= 5.061 P<0.001

0.77 (0.53-1.01)

0.55 (0.43-0.66)

Z= 0.493 P= 0.622

Table 5.8-50: Effect of Aspirin Intervention Prompt on Non-Aspirin Intervention Rate in Different Phases of the Study

5.8.11 Effect of Observation (univariate analysis)

The rate of clinical interventions per 100 prescriptions and per 100 patients was higher in the

pharmacies that ever had regular observation visits (see Figure 5.8-14). It should be noted that this

comparison is of clinical intervention rates between those pharmacies that were ever observed and

those that were not. The actual observation period was only for the first three weeks of the trial, and

the effect of observation in a particular pharmacy was not long-lasting. There was an even more

marked difference in the intervention rate on the actual day of observation, even within the same

pharmacy (see Section 6).


Page 218 of 361

.00 1.00

Ever Observed

0.00

2.00

4.00

6.00

8.00

10

10

1

4

2

1

42



Figure 5.8-14: Effect of Observation on Clinical Intervention Rate (1 = Observed, 0 = Not Observed)

The difference between the pharmacies that were observed for part of their recording time and those

that were never observed was statistically significant, whether the rate was measured in terms of per

100 prescriptions or per 100 patients (see Figure 5.8-15and Figure 5.8-16).

Statistically Significant Difference: Mann-Whitney U statistic = 175, p = 0.04

0 1 2 3 4 5 6


0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 0.4606Std. D ev. = 0.50327N = 30

Never Observed

0.00 1.00 2.00 3.00 4.00 5.00 6.00


0

2

4

6

8

10

12

Fre

qu

en

cy

Mean = 1.0241Std . D ev. = 1.07186N = 22

Observed for First 3 Weeks

Figure 5.8-15: Effect of Observation on Intervention Rate per 100 Prescriptions


Page 219 of 361

0.00 2.00 4.00 6.00 8.00


0

2

4

6

8

10

Fre

qu

en

cy

Mean = 1.7576Std . D ev. = 1.68255N = 22

Observed for First 3 Weeks

0 2 4 6 8


0

3

6

9

12

15F

req

uen

cy

Mean = 0.793Std . D ev. = 0.88954N = 30

Never Observed

Statistically Significant Difference: Mann-Whitney U statistic = 176, p = 0.04

Figure 5.8-16: Effect of Observation on Intervention Rate per 100 Patients

Ten of the 13 pharmacies in Group 1 (Above average uptake) were in the observed group of

pharmacies. Conversely, 12 of the 15 pharmacies with very low uptake (Group 4) were not observed

(see Table 5.8-51).

Not Observed Observed Total Intervention Uptake Group

# % # %

Group 1: Above Average Uptake 3 23.1% 10 76.9% 13

Group 2: Average Uptake 5 55.6% 4 44.4% 9

Group 3: Low Uptake 10 66.7% 5 33.3% 15

Group 4: Very Low Uptake 12 80.0% 3 20.0% 15

Total 30 57.7% 22 42.3% 52

*Significantly higher than expected; Pearson Chi Square Statistic 9.95, p = 0.019

Table 5.8-51: Effect Of Observation On Clinical Intervention Rate

There are two possible reasons for the increased intervention rate in the presence of an observer. The

first is that pharmacists increase the frequency of the documentation of interventions while the

observer is present, and the second is that they actually increase their frequency of performing

interventions while the observer is present. While it is likely that there is an element of each of these

effects, it would be useful to consider which of these effects is dominant, as this impacts on the scope

for increasing intervention rate through educational means.

In the questionnaire given to each pharmacist after completion of the study (see section 7.1), a

question was asked to obtain the pharmacist’s perception of what proportion of clinical interventions

they documented during the trial. Sixty-eight of 114 respondents (59%) indicated that they

documented 75% or more of the interventions they performed (see Figure 5.8-17). Thus, if

pharmacists in the non-observed arm were documenting a significant number of their interventions (at


Page 220 of 361

least 50% by their own estimate), the implication is that the reduced number of interventions in the

non-observed arm was due to reduced performance of interventions.

9

14

23

51

17

0

10

20

30

40

50

60

10% or less 25% 50% 75% 100%

Figure 5.8-17: Proportion of Clinical Interventions Documented by Pharmacists in the PROMISe Trial

5.8.12 Combined Effects of Remuneration, Intervention Prompt and Observation (multivariate analysis)

Using the unique record of Pharmacy-Day, we constructed a table of intervention rate and the factors

of remuneration, the aspirin prompt and observation, in order to perform ANOVA.

Univariate general linear model analysis was conducted for these factors against intervention rate for

each different phase of the study. A summary of the mean intervention rates for each of the subgroups

analysed is shown in Figure 5.8-18 and the detailed ANOVA results for each study phase are shown

in Table 5.8-52, Table 5.8-53 and Table 5.8-54.

In Phase 1, there was a significant effect of observation (F= 26.422, p<0.001). The mean difference

was an increase of 1.23 interventions per 100 prescriptions , or 154% compared to the unobserved

group. There was also a significant difference between remuneration and no remuneration in this

phase, but only in the group that were not observed.In the observed pharmacies, there was no

difference in overall intervention rate between pharmacies with and without the aspirin intervention

prompt. However, in pharmacies that were not observed, a difference was shown in the aspirin

intervention prompt group when remuneration was in place (F= 8.867, p=0.003).


Page 221 of 361

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 218.011(a) 7 31.144 10.682 .000

Intercept 474.712 1 474.712 162.821 .000

MaxOfPaid 7.019 1 7.019 2.407 .121

MaxOfAspirin 25.853 1 25.853 8.867 .003

MaxOfObserved_Day 77.035 1 77.035 26.422 .000

MaxOfPaid * MaxOfAspirin 4.652 1 4.652 1.596 .207

MaxOfPaid * MaxOfObserved_Day .311 1 .311 .107 .744

MaxOfAspirin * MaxOfObserved_Day 2.065 1 2.065 .708 .400

MaxOfPaid * MaxOfAspirin * MaxOfObserved_Day 1.459 1 1.459 .500 .480

Error 1900.932 652 2.916

Total 2786.877 660

Corrected Total 2118.943 659

Table 5.8-52: Multivariate ANOVA for Phase 1



Intercept 207.998 1 207.998 76.530 .000

MaxOfPaid .152 1 .152 .056 .813

MaxOfAspirin 39.817 1 39.817 14.650 .000

MaxOfObserved_Day 50.066 1 50.066 18.421 .000

MaxOfPaid * MaxOfAspirin 2.007 1 2.007 .738 .391

MaxOfPaid * MaxOfObserved_Day .044 1 .044 .016 .898

MaxOfAspirin * MaxOfObserved_Day 10.299 1 10.299 3.789 .052

MaxOfPaid * MaxOfAspirin * MaxOfObserved_Day .542 1 .542 .200 .655

Error 1589.949 585 2.718

Total 2234.009 593



In Phase 2, the effect of observation was maintained (even though observation only took place for one

of the two weeks in Phase 2), with a mean difference in intervention rate of 1.62 interventions per 100

prescriptions (208%). The effect shown with the Aspirin intervention prompt was also maintained, but

this time in the unpaid pharmacy group. It should be noted that the unpaid group in phase 2 are the

same pharmacies that were paid in Phase 1, so the maintenance of an effect of aspirin is most likely to

be related to other factors present in these pharmacies.


Page 222 of 361



Intercept 317.139 1 317.139 102.706 .000

MaxOfPaid .000 0 . . .

MaxOfAspirin 31.327 1 31.327 10.145 .001

MaxOfObserved_Day .000 0 . . .

MaxOfPaid * MaxOfAspirin .000 0 . . .

MaxOfPaid * MaxOfObserved_Day .000 0 . . .

MaxOfAspirin * MaxOfObserved_Day .000 0 . . .

MaxOfPaid * MaxOfAspirin * MaxOfObserved_Day .000 0 . . .

Error 3486.163 1129 3.088

Total 3819.415 1131



The effect of the aspirin prompt was continued in the third phase of the study, despite the fact that the

aspirin prompt was only installed for one week of the 4 weeks of this Phase. There were on average

0.37 more interventions per 100 prescriptions in those pharmacies that had the aspirin intervention

prompt installed (a difference of 88%).


Page 223 of 361

2,384 Pharmacy Days where more than 20 prescriptions dispensedMean Intervention Rate 0.74 (Standard Deviation 1.78)

Phase 1660 Pharmacy Days

1.01 (1.79)

Phase 2593 Pharmacy Days

0.88 (1.73)

Phase 31,131 Pharmacy Days

0.52 (1.76)

Observed112

2.02 (2.26)

Observed36

2.40 (2.87)

Unobserved557

0.78 (1.59)

Observed0

Unobserved1131

0.52 (1.76)

Unobserved548

0.80 (1.61)

Paidn=63

2.29 (2.43)

Unpaidn=49

1.68 (2.00)

Paid243

0.98 (1.94)

Unpaid305

0.65 (1.27)

Paid16

2.05 (2.29)

Unpaid20

2.67 (3.29)

Paid299

0.66 (1.37)

Unpaid258

0.91 (1.80)

Paid11310.52

(1.76)

Unpaid0

Aspirin52

No Aspirin

11Aspirin

29

No Aspirin

20Aspirin

147

No Aspirin

96Aspirin

172

No Aspirin

133Aspirin

8

No Aspirin

8Aspirin

16

No Aspirin

4Aspirin

171

No Aspirin

128Aspirin

167

No Aspirin

91Aspirin

278

No Aspirin

853

2.45 (2.59)

1.56 (1.27)

1.95 (2.17)

1.29 (1.69)

1.31 (2.29)

0.48 (1.05)

0.66 (1.32)

0.63 (1.18)

2.90 (2.51)

1.20(1.81)

3.07 (3.58)

1.10 (0.70)

0.74 (1.40)

0.56 (1.33)

1.27(2.12)

0.26 (0.52)

0.81(2.98)

0.42 (1.10)

NS; p=0.198 0.36 [0.19-0.53];

p<0.001

1.23 [0.87-1.58]; p<0.001

1.62 [1.05-2.19]; p<0.001

NS; p=0.159 NS; p=0.525 NS; p=0.062 0.33 [0.06-0.60];

p=0.016

0.83 [0.34-1.32]; p=0.001

NS; p=0.2551.01 [0.57-

1.46]; p<0.001NS; p=0.819NS; p=0.259NS; p=0.275 NS; p=0.143 NS; p=0.298

0.37 [0.15-0.62]; p=0.001

(Mean Difference (95% CI])

Figure 5.8-18: Subgroups Used for Univariate Analysis


Page 224 of 361

5.8.13 Pharmacy Specific Information

The overall clinical intervention rate was 0.55 clinical interventions per 100 prescriptions dispensed

(approximately one intervention for every 200 prescriptions). There were, however, a number of

pharmacies, whose intervention rates were significantly higher than this (see Table 5.8-55 and Figure

5.8-19).


Page 225 of 361

Pharmacy

Number

Clinical

InterventionsPrescriptions Patients

Clinical

Interventions

per 100

Prescriptions

Clinical

Interventions

per 100

Patients35 186 3622 2411 5.14 7.71

49 79 3420 2152 2.31 3.67

7 131 6504 3713 2.01 3.53

13 71 4027 2096 1.76 3.39

31 96 6137 3139 1.56 3.06

21 119 7327 3914 1.62 3.04

16 61 4097 2109 1.49 2.89

52 20 1554 785 1.29 2.55

51 63 4490 2492 1.40 2.53

18 225 15927 9883 1.41 2.28

14 71 6206 3294 1.14 2.16

29 52 5363 2865 0.97 1.82

17 74 7364 4465 1.00 1.66

3 24 3300 1831 0.73 1.31

4 42 6925 3725 0.61 1.13

30 63 8935 5686 0.71 1.11

5 38 7686 3477 0.49 1.09

38 46 7254 4494 0.63 1.02

8 37 7447 3727 0.50 0.99

40 42 6883 4255 0.61 0.99

22 119 20007 12082 0.59 0.98

10 46 8437 5191 0.55 0.89

42 10 1797 1271 0.56 0.79

1 62 13329 8062 0.47 0.77

32 21 5216 3007 0.40 0.70

2 20 4976 3011 0.40 0.66

24 19 3796 2914 0.50 0.65

48 27 7577 4397 0.36 0.61

44 10 3430 1687 0.29 0.59

36 14 3829 2531 0.37 0.55

45 13 4178 2471 0.31 0.53

19 42 13365 8601 0.31 0.49

41 58 19579 11968 0.30 0.48

11 8 3086 1681 0.26 0.48

50 63 19831 13351 0.32 0.47

9 38 14903 8464 0.25 0.45

20 25 9428 5744 0.27 0.44

6 12 4926 2956 0.24 0.41

23 21 9201 5557 0.23 0.38

39 24 10273 6382 0.23 0.38

27 28 12323 7495 0.23 0.37

43 30 13372 8090 0.22 0.37

37 28 13628 7943 0.21 0.35

15 13 8269 4530 0.16 0.29

26 30 16715 10671 0.18 0.28

34 16 11840 6787 0.14 0.24

33 13 10378 6306 0.13 0.21

46 4 2738 2045 0.15 0.20

28 5 4660 2764 0.11 0.18

47 12 14032 7637 0.09 0.16

25 10 12521 7610 0.08 0.13

12 4 9412 5260 0.04 0.08

Total 2385 435520 258979 0.55 0.92

Table 5.8-55: Clinical Intervention Rate For Different Pharmacies


Page 226 of 361

0

15 15

9

12

4

6

0

2

4

6

8

10

12

14

16

18

20

0 0.25 0.5 0.75 1 1.25 1.5 More

Clinical Intervention Rate

Group 4 (Very low uptake)

Group 3 (Low

uptake)

Group 2 (Average uptake)

Group 1 (Above average uptake)

0

15 15

9

12

4

6

0

2

4

6

8

10

12

14

16

18

20

0 0.25 0.5 0.75 1 1.25 1.5 More

Clinical Intervention Rate

Group 4 (Very low uptake)

Group 3 (Low

uptake)

Group 2 (Average uptake)

Group 1 (Above average uptake)

Figure 5.8-19: Number Of Pharmacies With Different Clinical Intervention Rates

As can be seen,

• 15 (29%) of the pharmacies had an overall clinical intervention rate significantly lower than the

mean (i.e. rates of 0.25 interventions per 100 prescriptions or less);

o Group 4:Very low uptake pharmacies

• another 15 (~29%) had intervention rates of between 0.25 and 0.5 per 100 prescriptions;

o Group 3: Low uptake pharmacies

• Nine (17%) had interventions around the average rate (between 0.5 and 0.75 per 100

prescriptions);

o Group 2: Average uptake pharmacies,

and

• The remaining 13 (25%) had intervention rates above the average (0.75 intervention per 100

prescriptions or more);

o Group 1: Above average uptake pharmacies.


Page 227 of 361

Summaries of the pharmacy factors in this section will relate to these four groups of intervention

uptake. More detailed analysis taking into account exact days of trial participation, excluding days

where there were no prescriptions dispensed and correcting for some of the key factors will be

undertaken in the economic evaluation section (see section 6).

5.8.13.1 Entrepreneurial Orientation

There was no simple relationship between overall clinical intervention rate and the entrepreneurial

orientation score (see Figure 5.8-20). The effects of observation and the intervention prompt may be

impacting on the detection of any relationship. In addition, the entrepreneurial orientation score is

more likely to increase in pharmacies with a more business orientated style, that may not relate to the

degree of professional integrity and clinical skills of the pharmacists working within the pharmacy.

Above Average Uptake

Average Uptake Low Uptake Very Low Uptake

Group

12.50

15.00

17.50

20.00

22.50

25.00

27.50

En

trep

ren

ori

en

tati

on

Figure 5.8-20: Relationship between Entrepreneurial Orientation and Intervention Rate

5.8.13.2 QCPP Adopter Status

As outlined in section 5.3.1.6, the QCPP adopter status of the pharmacies that participated in the

PROMISe study was determined from their date of accreditation. This status was adjusted for the

number of pharmacies with very short ownership (1 year or less) as they would not have had time to

become accredited yet. This status was compared to the intervention rate using the pharmacy-day


Page 228 of 361

construct previously described. Days where less than 20 prescriptions were dispensed have been

excluded from the analysis as they resulted in unrepresentatively high intervention rates.

The results of this comparison are shown in Table 5.8-56 and Figure 5.8-21. There was a significant

difference between the intervention rate for early adopters compared to the late majority and laggard

category of adopters.

Mean

Standard

Error of

Mean

Standard

DeviationMinimum Maximum Median

Tota l N

(Pharmacy-

Days)

innovator 0.61 0.07 1.00 0.00 5.97 0.00 236

early adopter 1.08 0.21 3.15 0.00 42.86 0.00 220

early majority 0.82 0.05 1.80 0.00 14.81 0.00 1121

late majority 0.43 0.05 1.11 0.00 10.56 0.00 473

laggards 0.42 0.09 1.00 0.00 6.59 0.00 116

Interventions per 100 Prescriptions

QCPP adopter

category

Table 5.8-56: QCPP Adopter Status (adjusted for short ownership) and Intervention Rate


QCPP adopter category

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

95%

CI In

terv

en

tio

ns p

er

100 P

rescri

pti

on

s

Figure 5.8-21: QCPP Adopter Status (adjusted for short ownership) and Intervention Rate


Page 229 of 361

5.8.14 Pharmacist Specific Factors

As outlined in section 5.8.1, a construct of Pharmacist-Week was used for the overall analysis of

pharmacist clinical intervention rates. When examined in this way, the mean rate of clinical

interventions per 100 prescriptions was 0.91, with individual pharmacists having rates ranging from 0

to 6.99 interventions per 100 prescriptions (see Figure 5.8-22 and Table 5.8-57).

7

41

24

11

76

4

0

3

0

5

10

15

20

25

30

35

40

45

0 0.5 1 1.5 2 2.5 3 3.5 More


Nu

mb

er

of

Ph

arm

ac

ists

Figure 5.8-22: Number of Pharmacists with Different Clinical Intervention Rates


Page 230 of 361

By examining the top 20 pharmacists (by average intervention rate over the three phases), we can see

that there are differences between pharmacists rates during the different phases of the trial. The

overall intervention rate fell from 1.03 interventions per 100 prescriptions to 0.44 from phase 1 to

phase 3 (see section 5.8.1). In the “top 20” pharmacists, some pharmacists increased their rate of

interventions (against the trend in the overall results) and others’ rates did not decline to the same

extent. There are likely to be a number of individual characteristics of these pharmacists that will

influence their intervention rates, and this is an area for further research.

Phase 1 Phase 2 Phase 3 Mean1 5.71 8.54 6.72 6.99

2 4.95 4.90 5.95 5.273 1.21 5.72 3.72 3.55

4 2.93 3.83 2.02 2.93

5 3.12 3.23 2.11 2.82

6 2.72 3.47 1.59 2.59

7 0.00 7.78 0.00 2.598 3.64 2.45 1.33 2.47

9 2.56 3.09 1.54 2.40

10 1.75 2.92 0.00 1.56

11 0.00 0.89 5.91 2.27

12 5.26 1.22 0.00 2.1613 4.49 0.75 0.89 2.04

14 6.00 0.00 0.00 2.00

15 1.45 3.14 1.31 1.97

16 2.36 2.12 1.35 1.94

17 4.37 0.00 1.08 1.82

18 1.65 3.32 0.48 1.8219 1.89 2.61 0.85 1.78

20 1.95 2.09 0.59 1.54

Pharmacist IDInterventions per 100 Scripts (mean)

Table 5.8-57: Pharmacist Specific Intervention Rates by Phase of Study (top 20 Pharmacists only)


Page 231 of 361

5.8.14.1 Pharmacist Factors That May Affect Clinical Intervention Rate

5.8.14.1.1 Clinical Problem Solving Skills Score

One factor that may be related to the rate of clinical interventions is an individual pharmacist’s clinical

skill level. The clinical problem solving skills score calculated for each pharmacist (see section 5.4.6)

was examined in relation to the intervention rate per 100 prescriptions (see Figure 5.8-23; extreme

values have been excluded).

The distribution of clinical skills score was not directly related to the intervention rate, implying that

other factors influence intervention rate more than the clinical skills score. The data examined in this

section do not take into account the effects of observation, the intervention prompt and remuneration,

and these may be masking the underlying relationship. There are likely to be other pharmacist factors

that influence the intervention rate.

0.00 2.00 4.00 6.00 8.00 10.00

Interventions per 100 Scripts

0.00

100.00

200.00

300.00

400.00

Clin

icalS

kill

Figure 5.8-23: Clinical Skill Score versus Intervention Rate


Page 232 of 361

5.8.14.1.2 Job Satisfaction

The Job satisfaction score calculated previously (see section 4.3.3.2) was plotted against the

intervention rate and the result is shown in Figure 5.8-24. There is some indication of a relationship

and this would require further exploration.

2.00 3.00 4.00 5.00 6.00 7.00

JobSatisfaction

0.00

2.00

4.00

6.00

8.00

10.00

Inte

rven

tio

ns

per

100 S

cri

pts

Figure 5.8-24: Pharmacists Job Satisfaction Score Compared to Clinical Intervention Rate


Page 233 of 361

5.8.14.1.3 Change Preparedness

The Change Preparedness score calculated for the PROMISe pharmacists (see section 4.3.3.2) was

plotted against the intervention rate and the result is shown in Figure 5.8-25. Again, there is some

indication of a relationship and this would require further exploration.

2.00 3.00 4.00 5.00 6.00 7.00

ChangePreparadness

0.00

2.00

4.00

6.00

8.00

10.00

Inte

rven

tio

ns

per

100 S

cri

pts

Figure 5.8-25: Change Preparadness Score Compared to Clinical Intervention Rate

5.9 Overall Impact of Automated Intervention Prompt

The effect of the Aspirin Intervention alert on pharmacist behaviour during the trial period was

significant. Pharmacists in the Aspirin Intervention Alert arm of the PROMISe trial documented 202

aspirin-related interventions over the trial period; it is estimated that all but one of these occurred as a

direct result of the alert. No aspirin-related interventions of this type were documented by pharmacists

not exposed to the alert.

Interventions resulting from the Aspirin Intervention Alert accounted for 77% of all interventions

associated with oral anti-diabetic drugs during the trial period. The intervention rate while the Aspirin

Intervention Alert arm of the PROMISe trial was active was 1.04 (per 100 prescriptions dispensed);

markedly higher than the intervention rates for the non-Aspirin arm which was 0.53 (see section

5.8.10).


Page 234 of 361

5.9.1 Cumulative interventions as a result of the Aspirin Intervention Alert

Figure 5.9-1 shows cumulative aspirin interventions over the trial period. The rate of interventions until

the 31st of May is fairly constant; but begins to level out after 4 weeks as pharmacists become

accustomed to the alert. This suggests some level of fatigue with the alert. An analysis of the dates for

which the interventions were recorded showed that only 7 aspirin interventions occurred after the alert

had been turned off (i.e. the alert was turned off on 31st of May, 2005 but the recording of interventions

continued until the 17th June 2005).

0

50

100

150

200

250

21-Apr-0

5

28-Apr-0

5

05-May-

05

12-May-

05

19-May-

05

26-May-

05

02-Jun-0

5

09-Jun-0

5

16-Jun-0

5

Date

Inte

rven

tio

ns

Aspirin Intervention Prompt Software

Removed31st May 2005

Figure 5.9-1: Cumulative Aspirin Interventions Over The PROMISe Trial Period (20th April To 17th June, 2005)


Page 235 of 361

5.9.2 Time taken to complete an aspirin intervention

Each Aspirin Intervention Alert intervention took approximately 4 minutes to complete (mean= 4.15,

SD = 2.77), see Figure 5.9-2.

5 10 15 20

TIME CODE

0

20

40

60

80

100F

req

uen

cy

Mean = 4.15Std. Dev. = 2.777N = 202

Figure 5.9-2: Time Taken To Complete An Intervention Related To The Aspirin Intervention Alert

5.9.3 Aspirin interventions by pharmacy

Thirty-one pharmacies had the Aspirin Intervention Alert installed. The mean number of aspirin

interventions per pharmacy during the trial period in this arm was 6.48 (range 0 to 33). Ten

pharmacies in the Aspirin Intervention Alert arm did not document any aspirin interventions, despite

having the alert installed (see Table 5.9-1)

The average rate of aspirin interventions for the 21 pharmacies that did document interventions was

approximately 10 per 100 oral hypoglycaemic prescriptions dispensed (see Table 5.9-2). Eight out of

the 10 pharmacies that did not document any aspirin interventions did not have an observer present.

The total number of aspirin interventions during the trial period in the observed arm was four times the

number in the non-observed arm (160 versus 42).


Page 236 of 361

Pharmacy Observed Y/N Number of interventions

1 Y 33

2 Y 26

3 Y 24

4 N 13

5 Y 13

6 Y 11

7 Y 10

8 Y 9

9 Y 9

10 N 9

11 Y 8

12 N 7

13 Y 7

14 Y 5

15 N 3

16 N 3

17 N 3

18 Y 3

19 Y 2

20 N 2

21 N 1

22 N 0

23 N 0

24 Y 0

25 N 0

26 N 0

27 N 0

28 N 0

29 N 0

30 Y 0

31 N 0

Average 6.48

Total interventions 207.481

Table 5.9-1: Number Of Interventions Recorded By Pharmacies With The Aspirin Intervention Alert

In the aspirin arm of the trial, the average intervention rate per pharmacy was 3.9 interventions every

100 anti-diabetic drugs dispensed,14

with a range of 0 (10 pharmacies) to 16.1. Thirteen of the 31

pharmacies had intervention rates above the average.

14 A single patient may have more than one anti-diabetic medication dispensed and may visited a

pharmacy on more than one occasion during the trial.


Page 237 of 361

Pharmacy No.

Aspirin Interventions

Oral hypoglycaemic Prescriptions

Rate of interventions per 100 oral hypoglycaemic

dispensed

1 9 56 16.1

2 24 155 15.5

3 13 98 13.3

4 33 274 12.0

5 13 128 10.2

6 7 74 9.5

7 26 306 8.5

8 9 131 6.9

9 10 149 6.7

10 8 136 5.9

11 5 102 4.9

12 7 143 4.9

13 11 228 4.8

14 3 97 3.1

15 3 99 3.0

16 3 113 2.7

17 9 391 2.3

18 2 135 1.5

19 3 232 1.3

20 1 107 0.9

21 2 260 0.8

22 0 111 0.0

23 0 74 0.0

24 0 96 0.0

25 0 142 0.0

26 0 384 0.0

27 0 185 0.0

28 0 221 0.0

29 0 179 0.0

30 0 42 0.0

31 0 331 0.0

Totals 201 5179 3.9

Table 5.9-2: Rate Of Interventions Per Oral Hypoglycaemic Prescription Dispensed


Page 238 of 361

Table 5.9-3: Aspirin Interventions Per Oral Antidiabetic Agents Dispensed

5.9.4 Potential contraindications to aspirin and subsequent recommendation by the pharmacist

Potential contraindications to aspirin use include active peptic ulcer or already taking antiplatelet

therapy (i.e. ticlodipine, clopidogrel or dipyridamole) or warfarin. Aspirin should not be recommended

in patients taking celecoxib or meloxicam due to increased risk of bleeding.17

We attempted to assess

if pharmacists took note of these contraindications prior to recommending aspirin by analysing the

dispensing history of patients.

Of the 202 aspirin interventions, 197 were able to be linked back to the dispensed medication history

for the patient. The types of drugs identified in the patient dispensing history are shown in Table 5.9-4

(includes current oral antidiabetic medication; excludes duplicate instances of the drug in the patient’s

history). Three-quarters of all patients (75%) were dispensed metformin, and one-quarter were

15 A single patient may be taking more than one oral antidiabetic agent.

16 Acarbose was not one of the drugs associated with the Aspirin Intervention Alert. There were only 3

interventions for acarbose over the trial period (all arms), and these all involved recommending aspirin

prophylaxis.

17 A link from the alert box to a Pharmacist Information Leaflet (PDF) detailed contraindications to

recommending aspirin.

Totals for trial period for aspirin intervention arm for 20th

April to 31st

May, 2005

Drug15

Prescriptions

All Clinical Ints

Aspirin Ints Intervention - % due to alert

Intervention rate

- aspirin per 100

prescriptions

acarbose16 60 3 3 100 5 glipizide 76 5 3 60 4 gliclazide 867 53 41 77 5

glibenclamide 290 20 13 65 5 rosiglitazone 136 9 4 44 3

metformin 3264 153 106 69 3 glimepiride 388 13 13 100 3 pioglitazone 97 5 2 40 2

Diabetes managed by

diet alone - 1 -

Cardiovascular drug

- 10 -

Unknown or misclassified

- 5 -

Total 202


Page 239 of 361

dispensed gliclazide (note that each patient may be taking more the one drug). The range of different

drugs in the dispensing history was 1 to 34.

C10AA05 Atorvastatin 41 20.8%C10AA01 Simvastatin 34 17.3%B01AC06 Acetylsalicylic acid 27 13.7%A10BB12 Glimepiride 26 13.2%A10BB01 Glibenclamide 24 12.2%C09AA04 Perindopril 22 11.2%C09AA05 Ramipril 18 9.1%C09CA04 Irbesartan 17 8.6%C08CA01 Amlodipine 17 8.6%C09DA04 Irbesartan and diuretics 16 8.1%J01CA04 Amoxicillin 15 7.6%C09BA04 Perindopril and diuretics 14 7.1%V04CA02 Glucose 14 7.1%M01AC06 Meloxicam 12 6.1%C10AB04 Gemfibrozil 12 6.1%C07AB03 Atenolol 12 6.1%M01AH01 Celecoxib 11 5.6%N02AA59 Codeine, combinations excl. psycholeptics 11 5.6%C08CA02 Felodipine 10 5.1%R03AC02 Salbutamol 10 5.1%A02BC05 Esomeprazole 10 5.1%

R03AK06Salmeterol and other drugs for obstructive airway

diseases10 5.1%

M05BA04 Alendronic acid 9 4.6%A02BA02 Ranitidine 9 4.6%A10BG02 Rosiglitazone 9 4.6%S01XA20 Artificial tears and other indifferent preparations 9 4.6%C08DB01 Diltiazem 9 4.6%R03BB04 Tiotropium bromide 9 4.6%

Others (including 5 warfarin) 381

Total 1022

Table 5.9-4: Other Medications Being Taken By The Patients Involved In The Aspirin Interventions

Drugs in the dispensing history may indicate potential contraindications to aspirin. For example, proton

pump inhibitors and H2 receptor antagonists may indicate peptic ulcer disease in some cases; NSAIDs

such as coxibs and oxicams in combination with aspirin may increase the risk of gastrointestinal

bleeding; and other antiplatelet agents or warfarin can increase bleeding risk. These drug groups and

the frequency with which they appeared in the dispensing history of patients recommended aspirin are

highlighted in Table 5.9-5.


Page 240 of 361

ATC Code

(L4)DESCRIPTION Number

% of 197

patients

A10BA Biguanides 148 75.1%A10BB Sulfonamides, urea derivatives 110 55.8%C10AA HMG CoA reductase inhibitors 81 41.1%C09AA ACE inhibitors, plain 59 29.9%C08CA Dihydropyridine derivatives 35 17.8%A02BC Proton pump inhibitors 30 15.2%B01AC Platelet aggregation inhibitors excl. heparin 29 14.7%C09CA Angiotensin II antagonists, plain 25 12.7%C09BA ACE inhibitors and diuretics 22 11.2%C09DA Angiotensin II antagonists and diuretics 20 10.2%C07AB Beta blocking agents, selective 16 8.1%M01AC Oxicams 15 7.6%A10BG Thiazolidinediones 15 7.6%J01CA Penicillins with extended spectrum 15 7.6%C10AB Fibrates 14 7.1%N05BA Benzodiazepine derivatives 14 7.1%V04CA Tests for diabetes 14 7.1%R03AC Selective beta-2-adrenoreceptor agonists 13 6.6%N02AA Natural opium alkaloids 12 6.1%

R03AKAdrenergics and other drugs for obstructive airway

diseases12 6.1%

M01AH Coxibs 11 5.6%R03BB Anticholinergics 11 5.6%A02BA H2-receptor antagonists 10 5.1%C08DB Benzothiazepine derivatives 9 4.6%M05BA Bisphosphonates 9 4.6%N06AB Selective serotonin reuptake inhibitors 9 4.6%D07AC Corticosteroids, potent (group III) 9 4.6%S01XA Other ophthalmologicals 9 4.6%

Others 242

Total 1018

Table 5.9-5: Grouping Of Medications From The Aspirin Interventions

An analysis of the notes made by the pharmacists in the comments fields suggests that some

pharmacists were aware of the potential contraindications and had elicited further information from the

patient. For example, for one patient taking a proton pump inhibitor, the pharmacist had noted the

indication as gastric reflux rather than peptic ulcer disease before proceeding with the intervention.

Furthermore for the Platelet aggregation inhibitor category, 27 of the 29 interventions were for aspirin,

and for most of these interventions the pharmacist had been alerted to a compliance problem with

previously prescribed aspirin or they were initiating aspirin therapy (according to an analysis of the

classification code and notes made in the comment field).

However, a significant number of interventions appeared to not take into account contraindicated

drugs in the dispensing history. Of particular concern, were the 5 interventions where warfarin

appeared in the dispensing history and the 2 interventions where clopidogrel was present. However,

further analysis of the event notes documented by the pharmacist sometimes revealed a fuller story:


Page 241 of 361

“The patient (had) stopped taking Plavix (clopidogrel) in April 2004 for no reason. (I) noticed this

when intervention alert for aspirin (with) diabetic came up. Dr will now re-initiate Plavix

(coplidogrel).”

Recommendation — contacted prescriber; Outcome — accepted.

These explanatory notes show where the documenting pharmacist had considered contraindications

and reveal a different story to that indicated by the coding system alone. The following notes recorded

by the pharmacist show several instances where potential contraindicated drugs in the dispensing

history were investigated by the pharmacist:

“(Patient had a) history of leg pain (arthritis?),been on many different NSAIDs, recently

complaining of paraesthesia of her feet. Suggest commencing again on aspirin — gave her

information sheet.”

Recommendation — refer to prescriber; Outcome — accepted.

“Diabetic (patient) over 50 (is) not taking aspirin. No contra-indications although (she was taking

an H2-antagonist) for reflux (which was) not well controlled.”

Recommendation — refer to prescriber; Outcome — unknown.

“Diabetic, hypertension, 56yrs, but stomach ulcer. (Intervention recorded but not recommended

aspirin)”

Recommendation — refer to prescriber; Outcome — partially accepted.

“Commenced Amaryl, no antiplatelet agent, ? contraindications (on Pariet - details of indication

unknown). Information provided to patient.”

Recommendation — refer to prescriber; Outcome — partially accepted.

For one case where clopidogrel appeared in the dispensing history and one case where warfarin

appeared in the dispensing history the outcome code was flagged as ‘accepted’ while the

recommendation code was flagged as counseled patient (no discussion with prescriber). These cases

are of particular concern.

It appears that in some instances pharmacists may have responded to the alert by recommending

aspirin with little or no consideration for contraindications. However, sometimes the event notes

revealed a different story.


Page 242 of 361

6 Results and Discussion Part 2: Actual and Potential Value of Interventions

Sections 6.1 and 6.3 outline the key methodological issues and the summary results, and sections 6.4

to 6.6 look in more detail at the results and allow for some interactive manipulation of the assumptions

that determine the estimates.

6.1 Key Methodological Issues

Clinical and economic valuations of health care interventions are common practice in health care

research. The evidence resulting from such research is routinely used in a wide range of policy

development and decision making. Such valuations tend to be in relation to a specific therapy,

procedure or public health program, for example, a program to improve patient compliance with

diabetes therapies. The clinical and economic valuation of new medical technologies, particularly

pharmaceuticals, in the form of cost effectiveness analyses, is perhaps the most widely recognized

and commonly performed. These studies are used throughout Europe, North America, and Oceania

to inform reimbursement decisions by National Health Insurers by allowing decision makers to

compare the likely costs and benefits of alternative strategies, typically adoption vs non-adoption of a

new technology.

Commercial, regulatory and accreditation decisions made by pharmacy owners, professional

associations and national and regional health funders will often impact, intentionally or unintentionally,

on the rate and/or quality of clinical interventions performed by community pharmacists. If the rate of

clinical interventions changes, so will patients’ health and wellbeing, and resource use in the health

care system. Decisions from amongst alternative strategies that may impact on pharmacist

intervention rates could be improved by routine consideration of the evidence of the clinical and or

economic valuation of community pharmacist interventions.

The intervention process begins with the detection and recognition of a drug related problem by the

pharmacist, followed by a recommendation for the resolution of the problem. In considering the value

of the intervention process, a number of issues are raised at each step of the process (see Figure

6.1-1).


Page 243 of 361

Routine Assessment and

Investigation

No Opportunity for Intervention

Opportunity for Intervention

Pharmacist Doesn’t Intervene

Pharmacist Intervenes

No Intervention Occurs

Someone Else Intervenes Later

Recommendation Made

Recommendation Made

Recommendation Accepted

Recommendation Accepted

Consequences of No Intervention

Consequences of Late Intervention

Consequences of PharmacistIntervention

Prescription

• Workload will impact on frequency of this activity

• Observation and study participation may influence frequency of this activity

Intervention Process Issues To Be Considered

• Prompts and educational techniques will influence detection of opportunity for intervention

• Value cannot be attributed to the pharmacist if someone else carries out the intervention

• Compliance with the recommendation will affect actual but not potential value of the intervention

• Value of the pharmacists intervention is based on the difference between the value of consequences with and without the intervention

• Multiple consequences are possible for each situation• Each consequence may have a different level of severity• Each consequence may have a different probability of

occurring• Impact of the particular consequence (at a particular level of

severity) may be in terms of general health, and in health resource use (and therefore financial impact)

Opportunity Recognised by

Pharmacist

Opportunity Not Recognised by

Pharmacist

• Nature of workload will influence opportunities for intervention (e.g. Number of scripts per patient)

• Professional ethics, remuneration perceived consequences and workload may influence whether pharmacist intervenes

Figure 6.1-1: Issues to be considered in Determining Value of The Intervention Process


Page 244 of 361

Published studies of the clinical and economic value of pharmacist interventions are comparatively

rare.18,19,20,21,22,23

Existing studies are primarily observational studies of the value of community pharmacists’ activity at

existing rates of intervention. Their primary intention appears to be to validate the existing and

expanded role of a pharmacist within the primary care setting. The reasons that published studies on

the value of pharmacist interventions were performed include, to assess:

1) the contribution of community pharmacists in detecting and reducing the impact of drug related

problems23

(Buurma);

2) the importance of the pharmacist contribution to the wellbeing of a patient in the hospital setting21,22

(Dooley, Nesbitt) and community setting 18

(Rupp);

3) the clinical and economic value of a range of strategies to increase the rate of intervention by

pharmacists compared to the cost of achieving these changes 21,20

(Nesbitt, Benrimoj); and

4) the potential to strengthen the role of community pharmacists and determine whether remuneration

based on pharmaceutical care rather than script volume is more likely to be effective 19

(Hawksworth)

5) the economic value of pharmacists in a climate of spiraling costs and the view of some policy

makers that “today, virtually all prescriptions can be filled by anyone who can read and count.” 18

(Rupp)

Existing studies of the value of community pharmacist interventions raise distinct methodological

issues that, unlike the case of more common areas of evaluation such as health technologies, are not

well documented in the literature. For studies of improved intervention rates to have value and

credibility in the decision making processes, the strength and limitations of alternative valuation

methods needs to be documented and understood. We reviewed these studies and summarised their

objectives, methods and results (see Table 6.1-4 a to d)

18 Rupp MT. Value of community pharmacists’ interventions to correct prescribing errors. Ann

Pharmacother 1992;26:1580-1584.

19 Hawksworth GM, Corlett AJ, Wright DJ, Chrystyn H. Clinical pharmacy interventions by community

pharmacists during the dispensing process. Br J Clin Pharmacol 1999; 47:695-700.

20 Benrimoj SI, Langford JH, Berry G, Collins D, Laughlan R, Stewart K, Aristides M, Dobson M.

Economic impact of increased clinical intervention rates in community pharmacy. Pharmacoeconomics

2000; 18(5): 459-68.

21 Nesbitt TD, Shermock KM, Bobeck MB et al. Implementation and pharmacoeconomic analysis of a

clinical staff pharmacy practice model. Am J Health-Syst Pharmacists 2001; 58: 784-90.

22 Dooley MJ, Allen KM, Doecke CJ, et al. A prospective multicentre study of pharmacist initiated

changes to drug therapy and patient management in acute care government funded hospitals. Br J Clin Pharmacol 2003; 57(4): 513-21.

23 Buurma H, De Smet PA, Leufkens HG, Egberts AC. Evaluation of the clinical value of pharmacists’

modicifations of prescription errors. Br J Clin Pharmacol 2004; 58(5):503-511.


Page 245 of 361

Rupp (1992) The objective of the study was: “estimate the impact on patient health status of the interventions that pharmacists to perform to correct the prescribing problems they identified” Value of pharmacist intervention Expected Medical costs avoided = P (most likely harmful outcome would have occurred) X the cost of medical care associate with that outcome Counterfactual: State of the most likely harmful outcome weighted by the probability that this state would otherwise have occurred. Factual: State of absence of that harmful outcome. The result was expressed as :

• The rate of intervention per script. • The percentage of interventions where there was a potential for harm, had the pharmacist not intervened. • For these cases, the mean probability of harm in absence of intervention. • The expected medical costs avoided per intervention.

Hawksworth 1999 The objective of the study was: To evaluate the professional contact between the community pharmacists and prescriber . Value of pharmacist intervention: A rating from 0 to 10 of the confidence that the intervention : 1) improved the efficacy of the patients therapeutic management; 2) prevented harm; 3) (version 1) preventing a hospital admission or (version 2) Preventing the likelihood of a hospital admission Counterfactual: State of lessened efficacy, harm and/or hospitalization or likely hospitalization, adjusted by the confidence that these states would have occurred. Factual: State of absence of the counterfactual hospital admission and harm and wit the presence of improved efficacy. The result was expressed as :

• Clinical interventions per script dispensed. • % of clinical interventions where there was at least one evaluator who was at least 10% confidence that harm avoided or efficacy improved. • The percent of scripts where there was at least 10% confidence by assessors that harm was avoided.

Table 6.1-1 a: Summary of Methodologies for Existing Studies of Value of Pharmacist's Clinical Interventions


Page 246 of 361

Benrimoj 2000 The objective of the study was: To determine the economic impact of an interventional program aimed to increase the rate of clinical interventions undertaken in community pharmacy. Value of pharmacist intervention: Economic impact = Health care costs avoided minus (health care costs incurred plus changes in medication costs plus pharmacy time plus telephone calls); where Health care costs avoided: = (version 1) P(intervention would prevent an adverse outcome) or (version 2) P(event occurring) X the costs of the probable course of treatment had the pharmacist not intervened. Counterfactual: The state of the adverse event and associated course of treatment weighted by one or other of the probability that 1) it would occur or 2) that it was avoided. The pre-intervention course of medications. Factual: The absence of the adverse outcome and corresponding course of treatment, changed medication and the additional phonecalls and pharmacist time. The result was expressed as: For each arm of this trial that had varying levels of education and financial incentives, the result was expressed in terms of the dollar savings from each arm of each trial, on a per script and per intervention basis

Nesbit 2001 The objective of the study was: The economic analysis was performed to estimate the value of the CSP practice model to the institution. Value of pharmacist intervention: Net economic value to institution = costs savings plus cost avoidance related to averted AEs less the costs of the CSP model Cost savings = Cost of recommended drug therapy less cost of previous drug therapy Cost avoidance = P(ADE in absence of intervention) X Average cost of AE Counterfactual: An adverse event weighted by its probability in absence of intervention. Continued pre intervention drug therapy. Factual: The recommended course of medication and the absence of the adverse state The CSP model is implemented. The result was expressed as :

• Cost avoidance and cost savings per intervention, per intervention type per 12 months • Distribution of interventions by probability of adverse event in absence of intervention (eg 53% had probability of 0.1) • Net economic value to institution

Table 6.1-2 b: Summary of Methodologies for Existing Studies of Value of Pharmacist's Clinical Interventions


Page 247 of 361

Dooley 2003 The objective of the study was: To determine the cost savings of pharmacist initiated changes to hospitalized patients’ drug therapy at hospitals Value of pharmacist intervention: Financial value = $ value (readmissions and LOS prevented + difference in medical procedures and pharmacotherapy costs) • Expected value (EV) of readmission prevented = (P(readmission w/o intervention) – P(readmission w intervention)) X cost per relevant DRG • EV of LOS= (actual LOS + expected additional days of current admission had intervention not occurred) X cost/day • EV of medical procedures = (1- P(medical procedures changed as a consequence of intervention)) X the cost of the medical procedures that actually

occurred • EV of pharmacotherapy = the cost of the pharmacotherapy that was recorded as occurring prior to the intervention Counterfactual: A DRG specified readmission without intervention (weighted by probability it would occur), actual LOS plus expected additional days in absence of intervention; pre intervention drug therapy continued until discharge; the actual medical procures and tests less the expected change Factual: A DRG specified readmission with intervention (weighted by probability it would occur in presence of intervention– not follow-up observational) actual LOS, actual medical procedures, actual pharmacotherapy and diagnostics that actually occurred. The result was expressed as : The value by type of intervention and the annualized costs savings

Buurma 2004 The objective of the study was: To examine the clinical value of pharmacists interventions to correct prescription errors Value of pharmacist intervention: “What is contribution of the intervention to the pharmacotherapy of this patient?” Each intervention that had a potential positive contribution was rated on either or both of the following: Avoided ADE plotted as (x, y) where: X = version 1 P(there was a contribution); version 2 P (that the ADE would have otherwise occurred); Y = severity of ADE. Improved therapy plotted as (x, y) where: X = as for ADE, Y = importance of therapy improvement Counterfactual: w/o improved therapy and with ADE Factual: state with no ADE and with improved therapy It is unclear whether the confidence weighting related to the factual or counterfactual – is it the confidence it made a contribution or the confidence that it would have otherwise occurred

The result was expressed as :

• A quadrant diagram that plotted each of the one or two points derived as above for each intervention judged to make a positive contribution.

• A statement on the % of interventions that made a positive contribution.

Table 6.1-3 c: Summary of Methodologies for Existing Studies of Value of Pharmacist's Clinical Interventions


Page 248 of 361

Promise 2006 The objective of the study was: To estimate the clinical and economic value of clinical interventions performed in community pharmacies, under a range of prompts (pop-ups and/or observers or neither) and incentives (payment or no payment) Value of pharmacist intervention: The value of intervention = (Expected economic and clinical value of the patients state over the next 12 months had no intervention occurred Less the expected value of health state as a consequence of the intervention) X P(that intervention would not have occurred had the pharmacist not performed it) (attribution) Assume compliance Economic value of a state has the following characteristics: 1) Health loss – days of health loss and severity of that health loss; 2) hospital admission - probability of admissions and days in hospital; 3) Resource use: consultations and diagnostics tests. Counterfactual: The state where no intervention, from any professional occurs; with a complete range of possible consequences identified and weighted by severity and probability. Factual: The state where the intervention occurs with each consequence weighted by severity and probability of occurring. The result was expressed as: The prevalence of clinical interventions per 1000 scripts The following per intervention and per 1000 scripts • The Dollar savings to health sector • The days of health loss prevented • The number of days in hospital prevented • The number of admissions prevented. • Consultations prevented • For sample and extrapolated to annual for Australia

Table 6.1-4 d: Summary of Methodologies for Existing Studies of Value of Pharmacist's Clinical Interventions


Page 249 of 361

We identified four key methodological issues that were resolved in distinctly different ways in each

published study and may in most cases have resulted in significant under or over estimates of the

clinical and economic value of community pharmacist interventions. A summary of these four

methodological issues follows.

6.1.1 How do we define the value of pharmacist activity?

While all the studies reviewed had the broad objective of validating the role of community pharmacists

in primary health care there were differences across studies in how this objective was interpreted as

the actual method and analysis. Does value of pharmacist activity require us to compare current

practice to what would happen if scripts were dispensed without a clinical context? Do we assume that

other health professionals intervene opportunistically or not at all? Do we assume current pharmacist

activity is at the highest possible incidence and quality? Do we consider the value when compliance by

prescribers and patients is complete or partial? There is a need for clarity in the definition of the object

of the evaluation. The main issues to be considered in defining the value of a pharmacist’s activitiy

are summarised in Table 6.1-5.

1. There are a number of alternative ways to define the value of community pharmacist activity, and the choice from amongst these depends upon the objective of the study, which, broadly speaking is either symbolic or purposeful.

2. A symbolic study – analogous to burden of disease studies, generally either explicitly or implicitly asks:

• What would the costs (health and financial) of removing the pharmacist’s clinical input in the dispensing process, assuming no one substituted for these.

3. A purposeful study – analogous to the trial of a new pharmacotherapy, usually designed to understand the implications of a decision in terms of at least one of the following:

• What is the value of increasing (or costs of decreasing) the rate of interventions?

• What is the value of increasing (or costs of decreasing) compliance with interventions by patients and prescribers?

• What is the value of improving (or costs of reducing) the quality of interventions?

4. Regardless of the objective of the study, there are five estimates that need to be made – either empirically or by assumptions:

• the time taken to investigate all scripts and to perform interventions when required;

• the incidence of interventions;

• value of the advice, assuming full compliance;

• compliance, by prescribers and patients, with the pharmacist advice;

• Attribution - the probability that someone else would have provided that advice or acted consistently with that advice had the pharmacist not provided it.

Table 6.1-5: Issues Regarding the Assessment of Value of Pharmacists' Activities


Page 250 of 361

All published studies sought at the broadest level to validate the role of the pharmacist as a contributor

to patient’s health and wellbeing through part of the primary health care sector, or the hospital setting.

Two published studies had an experimental component: (Nesbitt21

and Benrimoj20

). Nesbitt analyzed

the value of the pharmacist activity in terms of the value of the activity performed in an alternative

model of service provision. The unstated assumption was that none of these interventions would have

occurred had this model of service provision not been in place. Benrimoj assessed the value of the

additional interventions that occurred under a range of strategies, including basic or advanced

education and remuneration. These two studies measured the additional value of pharmacist

interventions as a result of alterative strategies to improve the rate of interventions. The remaining four

published studies sought to measure the value of what pharmacists were currently doing, essentially

by asking what would be the cost in terms of financial or health loss had the dispensing occurred at a

purely administrative level - by a technician. The question of whether the rate could be increased was

not considered.

All studies estimated the incidence of interventions, and the type of interventions included in the

valuation varied across studies.

Compliance with pharmacists’ advice in Buurma23

, Nesbitt21

and Dooley22

was 100%. In Buurma, this

was due to the subset of interventions included (prescriber was phoned to correct a script prior to

dispensing). The other two studies were in the hospital and the data was available to confirm whether

the advice was complied with. In other studies it was assumed, implicitly to be 100%. In all published

studies it was assumed that had the pharmacist not performed the intervention, no other health

professional would have intervened – that is the pharmacist was attributed the full value of the

intervention.

Studies that included an estimate of pharmacist time did not include the costs of the additional time

taken to investigate a script, regardless of whether an intervention was actually required. It is difficult

to collect this data on a per script basis as the recording systems tend to be on a per intervention

basis. Ideally a survey of scripts could be performed, as it is likely that many scripts are investigated

for each one that results in an intervention.

In reviewing these studies it is clear that, unlike a straightforward clinical trial, there is a possibility that

there will be lack of consistency across studies is what is meant by “value of pharmacists’

interventions or activity”. In general, studies that measure the effect of an intervention designed to

change the rate or value of interventions, or compliance with pharmacist advice will be more useful to

decision makers than studies designed to estimate the prevailing incidence of interventions and their

value.

The decision as to whether to consider the attribution or compliance is less straight forward. Using one

of either is straightforward as the final estimate is simply weighted by the probability of either. Ideally

we would estimate both, however the mathematics of incorporating both may be complex as the

simulations discussed in the following section would need to be adjusted by further scenarios to

address the interactions between, for example, non compliance by the patient and then a second

health professional intervening.


Page 251 of 361

6.1.2 Defining the Factual and the Counterfactual States and the Effect of the Recommendation Made.

Studies of community pharmacist interventions tend to be, for practical and ethical reasons,

observational, focus on pharmacist rather than prescriber and patients activity and do not have patient

outcome or resource use data. Common to all published studies is that all simulate or hypothesise

rather than test the effect of absence of pharmacist interventions on patient outcome and resource use

– they use simulation rather than observation.

Such methods require careful specification and valuation by panels of assessors of the effect being

simulated or hypothesised, typically of the form: “What adverse outcome did the pharmacists’

intervention prevent?” or “What improvement in therapy resulted from the pharmacists’ intervention?”.

The lack of agreement on this type of valuation processes, contrasts to the well documented protocols

for identifying and measuring relevant outcomes and comparing these across the arms of a controlled

clinical trial (seeTable 6.1-6).

1. Specify the hypothesised counterfactual and factual relating to the recommendation being carried.

• The counterfactual is the hypothesised possible patient states should the recommendation not be carried out, by any provider.

• The factual is the hypothesised possible patient states had the recommendation been carried out, by any provider

• The effect of the clinical recommendation is the difference in the expected value of the patient’s factual and counterfactual states.

2. Adjusted the expected value of advice by estimates of attribution and/or compliance.

Table 6.1-6: Issues to be Considered in Determining Effect of Recommendation Made by Pharmacist

In a controlled trial we are able to compare the case of the intervention, in this case the actions of the

pharmacist, with a case in which the pharmacist does not act, the counterfactual. The implications of

not allowing a pharmacist to intervene, once the need for an intervention is established, are not

acceptable. It is necessary then to hypothesise as to the state or states the patient would have been

in had no intervention been performed.

No published study explicitly defined both a counterfactual and factual state and most studies sought

the assessors to value an effect of the intervention rather than to value to the two possible patient

states and then estimate effect as the difference in value between these two states. This leads to a

number of points of possible ambiguity. Consider the example of a patient who is being prescribed two

NSAIDs and the pharmacist intervenes to stop one of them (see examples later). Are we asking

assessors: 1) the probability that a GI event would occur in the absence of the pharmacist’s

intervention; 2) the increase in baseline risk of a GI event occurring: or 3) the probability that the

intervention prevented a GI event. In this example, there is a baseline risk of a GI event when a

patient is using one NSAID and it is clear that 2) rather than 1) will result in a lower and more accurate

estimate of the relevant risk. The problem with asking assessors to look at change is that they may

refer to a reduction in relative risk (for example 50%) rather than the correct response, the absolute


Page 252 of 361

risk reduction. The interpretation of 3) by the assessor is difficult to anticipate. Does this relate to the

probability that the intervention resulted in the event being prevented (eg 100%)?

Our suggestion is that to separately value the factual and counterfactual and finding the difference

results in a more accurate estimate of effect by removing ambiguity and allowing for the possibility that

a risk of an event will not be reduced to 0 if the advice is complied with, due to either the partial

effectiveness of the intervention, the baseline risk of that event or the possibility that the risk will not

reduce to baseline immediately.

6.1.3 Accuracy of Specification of Factual and Counterfactual States

The interventions performed by community pharmacists reduce the risk of a significant range of

possible consequences for patients with a range of conditions and characteristics. There is a

requirement to obtain reasonable estimates of consequences for a large range of interventions in a

variety of patients, rather than highly accurate estimates for a few consequences in a strictly defined

patient group, as is typically the case for a clinical trial. These consequences need to be weighted for

both their probability and severity. These wide ranging estimates of effect also need to be aggregated

using common indicators of patients’ health and wellbeing and resource use. Again, there are no

established methods for addressing this complex methodological requirement (see Table 6.1-7)

1. Specify each of the clinical pathways possible, for the counterfactual and the factual.

2. Fully specify the probabilities of each clinical pathway for both of the states.

3. Include in the outcomes both health related and resource use: days of health loss and severity of the health loss, whether there was an admission, days in hospital, consultations with GPs and specialists and diagnostic tests.

4. Separate the resource use associated with the states the patient is in as a result of the advice being carried out (eg the patient goes to their GP)

Table 6.1-7: Issues to Be Considered in Determining Factual and Counterfactual States

If we agree that we need to specify both the counterfactual and factual states, then the next

methodological decision is how many of the possible clinical pathways in these states and how much

detail in terms of health and resource related outcomes should be specified.

While a simple approach, for example selecting the most significant adverse events seems appealing,

and could be the only computationally feasible option for one or two of the earlier studies, it can lead

to ambiguities such as is this the most likely adverse event or the most severe adverse event.


Page 253 of 361

There are four main methodological issues:

1) the narrative provided to the assessors;

2) the number of possible consequences and clinical pathways (or levels of severity) that assessors

can characterise the interventions by;

3) the number of outcomes that can be considered in relation to each of the interventions; and

4) compliance with the recommendations by the patient and/or the prescriber

6.1.3.1 The Narrative Provided to the Assessors

More complete specification of possible pathways requires more complete patient information, which is

difficult to achieve in the community compared to the hospital setting, unless individual patient consent

is sought and given. The practicalities of retrospective consent from patients who are identified as

having an intervention, are limited in the community setting and such a method was not used in any of

the published community studies. In most community pharmacy studies, the researchers do have

available dispensing information including the other scripts dispensed to that patient from that

pharmacy on the same occasion. In our study we also had dispensing data from that pharmacy for up

to three months previous to the study and for the weeks of the study. The notes pertaining to each

intervention were recorded in various forms in most studies. All published community studies made

the best use of available information to develop a description of the intervention and to provide the

assessor with enough information to determine possible clinical pathways. The hospital studies,

typically involved less pharmacists (less sites and longer time periods), involved the dispensing

pharmacist in the development of narratives.

6.1.3.2 The Number of Possible Consequences of the Intervention

The second problem is how to specify the possible clinical pathways for the patients in a way that

allows consistency across assessors, and how fully they should be specified. The majority of the

published methods required the assessor to select either the most significant or most likely, harm

prevented and in some cases the most significant therapeutic benefit. No published study either

required or allowed the assessors to specify either more than one possible set of consequences

(gastrointestinal or cardiovascular) or more than one level of severity (bleeding, moderate pain or mild

discomfort) within these consequences. This could have resulted in three sources of bias in the

estimates of the value of the pharmacist’s interventions. First, by requiring assessors to identify the

most significant harm prevented, it is unclear whether assessors would have chosen the low

probability severe impact event or a high probability moderate impact event. Given the symbolic

imperative to demonstrate that a hospital admission was prevented, there is likely to have been a bias

towards the former. The combined effect of choosing the low risk high impact event and only one type

of consequence is likely to be an underestimate of the true value of clinical interventions. Second,

most studies looked at types of harms in a general outcome sense, for example, probability of hospital

admission or poor disease control, without reference to the underlying clinical pathway. There could be

a number of clinical pathways by which one type of harm, eg poor disease control, could have been

reached. It is unclear whether the assessors consider the harm relating to one clinical pathway or the


Page 254 of 361

expected outcomes of multiple possible clinical pathways, each weighted by their probability. Again

this would lead to an underestimate of the clinical and economic value of interventions. Third, a

specific clinical pathway could have had a number of types of harms associated with it.

6.1.3.3 The Range of Outcomes from the Intervention

The third problem is how many outcomes to include. All studies that quantified outcomes such as

readmission probability, but no published study had used an aggregately indicator of health status.

Hence the only indicator by which aggregation could occur tended to be the financial outcomes. This

could lead to underestimate of the full value of the interventions and tend to focus any discussions on

the potential financial savings resulting from pharmacists’ clinical interventions and hospital

admissions avoided rather than the more frequent occasions on which a modest health loss was

prevented.

We used days of health status lost and whether this loss was severe, moderate or mild as a way of

aggregating the clinical benefit across many different types of clinical pathways across individuals and

for multiple possible pathways for one individual. Similarly we looked at each of four possible areas of

resource impact – pharmaceuticals, consultations (GP and specialist), days in hospital and diagnostic

tests.

6.1.3.4 Compliance with the Suggestions

The fourth problem is how to adjust the value of the advice to consider the possibility that another

health professional would have otherwise performed the intervention and the possibility that neither

the patient nor the prescriber complies with the advice. The pragmatic adjustment for the former is a

weighting of the difference between the factual and counterfactual to allow for the possible that

another health professional would have performed the intervention had the pharmacist not. The

pragmatic adjustment for the latter is an equivalent weighting to the effect. Ideal adjustments would

require that the counterfactual and factual states have separate weightings applied to them. It is

difficult to adjust for both without extensive modification to the scenarios presented to the assessors.

6.1.4 Consideration of Separate Consequences and Outcomes for One Individual and for the Population.

There is a risk in the collection, analysis and reporting of information as complex as that proposed

above becomes too difficult. It is also possible that this potential difficulty was a constraint on the

published studies, none of which sought to elicit as complex information from assessors or consider

clinical value of pharmacist interventions using an indicator of health impact (see Table 6.1-8).


Page 255 of 361

1. Data collection from the pharmacies using appropriate software allows the

minimum level of data required to develop narrative for each of the selected

interventions for review but maximises the rate of recording by pharmacists.

2. Distinguish between incidence of opportunities for, and performance and

recording of, interventions.

3. Elicit responses from assessors using software available online, menus to

access types of consequences and levels of severity of these consequences,

the separate valuation of factual and counterfactual and

4. Allow the health and resource use tables to be updated using local and current

data without requiring the probabilities of consequences to be estimated.

5. Data collection from the pharmacies using appropriate software allows the

minimum level of data required to develop narrative for each of the selected

interventions for review but maximizes the rate of recoding by pharmacists.

6. Distinguish between incidence of opportunities for, and performance and

recording of, interventions.

7. Elicit responses from assessors using software available online, menus to

access types of consequences and levels of severity of these consequences,

the separate valuation of factual and counterfactual and

8. Allow the health and resource use tables to be updated using local and current

data without requiring the probabilities of consequences to be estimated.

Table 6.1-8: Issues to be Considered Relating to Quality of Information Regarding the Intervention

For example, Buurma used a method that considered both the severity and probability of adverse

events prevented, and the severity and importance of effectiveness improvements, however, while

these results were presented graphically in a quadrant diagram, there was no aggregation across the

interventions. Similarly Hawksworth did look at the different level of confidence associate with the four

possible consequences but it was not possible to aggregate them to a series of indicators of clinical

and economic value of pharmacist interventions. All studies that included financial indicators

aggregated them across interventions, but there is limited value in these as they do not cover all

consequences and may not be appropriately weighted for risk and severity. Furthermore they may

under estimate resource use in the factual state (typically assumed to be zero).


Page 256 of 361

Three key methodological innovations in the Promise study that made this a manageable task were in

relation to: the collection of data from the pharmacies; the collection of data from assessors; and the

choice of clinical and economic summary indicators of outcomes.

1. The PROMISe intervention database and the associated prescription information allowed an accurate narrative to be reconstructed of the intervention.

2. The software used to illicit responses from assessors had three parts:

a. the narratives and dispensing information for each intervention;

b. an online tool that allowed each assessor for each intervention to select one or more possible consequences from a range of different consequences and to assign probabilities to each of three possible levels of severity of that consequence, and also to the probability of no effect;

c. the separate specification of these consequences and probabilities of the consequences for the with intervention and without intervention state; and the opportunity to observe their input graphically, for their assessments alone or for all assessors assessment of that intervention.

3. Each of the consequences that could be selected by the assessors had the clinical and resource use outcomes assigned separately.

In summary, the PROMISe approach is more likely to lead to an accurate estimate of the value

of the pharmacist intervention, and it may be lower or higher than the alternative methods. Its

comprehensive approach to the identification of possible consequences leads to a more

accurate and potentially higher estimate, however, the methods of weighting, including

allowing for multiple states of severity and the possibility of no effect occurring, make it also

possible that the estimate is lower. There are two factors that will definitely lower the estimate

developed using the Promise method: the inclusion of a weighting for attribution and the use

the difference between the expected value of the factual and counterfactual states as a way of

measuring the effect, rather than the possibility of preventing the counterfactual state.

6.1.5 Comparison of Different Methodologies

We identified three interventions which occurred frequently during the PROMISe Intervention Study

and attempted to allocate a value to each of them using the six methods reviewed (PROMISe

method and five others). The three example interventions are shown in Table 6.1-9.


Page 257 of 361

Intervention Example 1: Aspirin Prophylaxis

A 58 year old female patient with type 2 diabetes with a history of hypertension and mild ischaemic heart disease is advised to commence an anti platelet agent (aspirin) to reduce her cardiovascular risk.

Intervention Example 2: NSAID Duplication

A 65 year old male with osteoarthritis who is already taking celecoxib is prescribed meloxicam (in addition to the celecoxib). The pharmacist suggests changing this to regular paracetamol.

Intervention Example 3: Paediatric Dose Increase

A 5-year-old, 23kg child is prescribed amoxicillin suspension at a dose of 100mg three times daily. The pharmacist checks the paediatric dosing schedule and suggests an increased in dose to 250mg three times daily

Table 6.1-9 : Sample Interventions Used to Campare Value Methods

Where inadequate information was available in the published method, assumptions were made and

these are included in each of the analyses. Comments on any difficulties applying each method are

also made.

6.1.5.1 Rupp Method

Rupp used a series of questions to assess the potential value of the intervention. The questions are

shown, summarised, in Table 6.1-10 (note that Rupp gives a list of options as responses to questions

2 and 4). The first question relates to the possibility that an adverse health consequence could have

occurred if the pharmacist did not intervene. This question does not specify a timeframe, and it does

not specify a lower limit of possibility. Based on this, the answer to “could this have resulted in…?” is

almost always yes. The second question then asks what is the most likely consequence. The

assessor, in considering the first point, would be considering any adverse health impact, regardless of

likelihood, whereas in the second question, they are directed to the most probable consequence. In

many cases, the most probable consequence is not the consequence that contributes the most to the

overall health impact of the intervention. Indeed, in the third question, the assessor is asked to

estimate the probability that the consequence would occur. Note that the probability estimates are set,

and there is no scope for a consequence that occurs at less than 10% frequency (such as a stroke or

heart attack in sample intervention 1).


Page 258 of 361

Aspirin

Prophylaxis NSAID

Duplication

Paediatric Antibiotic

Underdose 1. Could this event have

resulted in adverse health consequence to the patient if the pharmacist had not intervened?

Yes1 Yes Yes

2. What adverse health consequence do you consider most likely to have resulted from this event if the pharmacist had not intervened?

• Toxic effects of drug(s) involved

• Inadequate control of patient’s condition

• Allergy/hypersensitivity reaction

• Other

Angina Gastrointestinal Discomfort3

Continuation of underlying Infection

3. Based on the available information, what is your estimate of the probability that this event would have resulted in adverse health consequence specified above? (0.1,0.3,0.5,0.7,0.9)

0.32 0.7 0.5

4. What intensity of healthcare would be needed to treat the adverse health consequence specified above, assuming that it did occur?

• Medical attention (hospitalisation) US$1891

• Medical Attention (no hospitalisation) US$110

• Unscheduled physician contact US$60

• Scheduled physician contact US$40

• Self Care

Medical Attention (no hospitalisation) US$110

Unscheduled physician contact US$60

Unscheduled physician contact US$60

Value of Intervention 0.3 X $110= US$33 (1992)

0.7 X $60 = US$42 (1992)

0.5 X $60 = US$30 (1992)

Comments and Assumptions

1 the likelihood of a consequence is low

2probability depends on the timeframe (assumed 1 year)

3may result in multiple consequences and the most likely

may not be the most significant Table 6.1-10: Assessment of Sample Interventions Using Rupp's Method

The fourth question asks the assessor to estimate a level of treatment required for the consequence,

assuming that it did occur. The resultant value is based on the probability of the consequence and the

cost of the medical treatment for it.


Page 259 of 361

6.1.5.2 Hawksworth Method

Scoring in this method was on a scale of 0 (definitely not) through to 10 (100% confident) for each of

three areas, improved efficacy, preventing harm and preventing hospitalisation. There is a difficulty in

interpretation of this scale, as shown in the sample interventions. Is the score allocated related to the

likelihood of improved efficacy, or is it the degree of improved efficacy? For example, in the aspirin

prophylaxis sample intervention, there is definitely some improved efficacy (reduced risk of cardiac

events), so the score would be high. However, the degree of improved efficacy (the reduction in

cardiac events) may result in a significantly lower score being applied. A similar interpretation can be

made for all three sections of the evaluation using the Hawkeworth method (see Table 6.1-11).

The results were not converted to dollar values and are presented as the proportion of prescriptions

where a 10% or more risk of harm or improves afficacy (ie a score of 1 or more) was allocated.

Aspirin Prophylaxis NSAID Duplication Paediatric Antibiotic

Underdose 1. Improved efficacy

91 0

3 9

2. Prevented harm 9

8 4,5

06

3. Prevented hospitalisation

12

25 0

7

Value of Intervention

Individual interventions were not valued, an overall value was expressed as a proportion of prescriptions that had a score of 1 or more in categories 1 or 2


1 confidence in improved efficacy (of any level) or the level of improved

efficacy? 2probability depends on the timeframe (assumed 1 year)

3 possibility of decreased efficacy of analgesia, which may have occurred from

this intervention. 4may have multiple consequences that are harmful (e.g. gastric, cardiac or

renal) 5risk of hospitalisation or harm exists even with the intervention

6low risk of drug related harm, but reasonable risk of disease (infection) related

harm 7risk of hospitalisation likely to be less than 10%, but no scope for entering a

number between 0 and 1 Table 6.1-11: Assessment of Sample Interventions Using Hawksworth’s Method


Page 260 of 361

6.1.5.3 Benrimoj Method

The economic analysis as described in this paper includes the sum of healthcare costs avoided and

the healthcare costs incurred. Avoided costs were based on a 5 member clinical panel’s opinion on the

probable course of treatment and the probability that the intervention would prevent an adverse

outcome. Incurred costs (GP or emergency department visits, changes in medication costs, pharmacy

time and telephone calls) were estimated for each intervention assessed by the panel. The economic

impact of the intervention was then used to calculate a cost saving per 1000 prescriptions.


Underdose Healthcare costs avoided (HCCA)= Probability

1 X cost

of probable course of treatment

0.32 x ED visit cardiovascular ($32)3

= $9.60

0.7 x ED visit digestive ($49)3=

$34.30

0.5 x ED visit infection ($37)3 =$18.50

Economic Impact = HCCA – (Healthcare costs incurred + Changes in medication cost + pharmacy time + telephone calls)

$ 9.60 – (cost of drug added $6.36+ 4

minutes of pharmacists time

$2.42 +2 phone calls $0.70)

$ 34.30 - (cost of drug ceased -$32.98 +

pharmacist time $4 + phone calls $0.35)

$18.50 - (cost of drug dose increase $0.00 + pharmacist time $4 +

phone calls $0.35)


$9.60 - $ 9.48 = $0.12 (1997)

$34.30 -$28.63 = $5.67 (1997)

$18.50 -$4.35 = $14.15 (1997)


1Probability is considered differently in the two Benrimoj papers, either the probability of an event occurring or the probability that the intervention would prevent an adverse outcome 2 assumed a 1 year timeframe 3 DoHA Manual of Resources and Their Associated Costs, 1993

Table 6.1-12: Assessment of Sample Interventions Using Benrimoj’s Method


Page 261 of 361

6.1.5.4 Nesbit Method

In this hospital-based study, Nesbit focused on determining value if the intervention based on the

probability of an adverse drug event (ADE) occurring (as allocated by a five member panel), and the

direct drug acquisition costs. The estimated cost avoidance was calculated as the probability

determined above by a set amount (US$5006) which was based on a previously published prospective

trial.24

There are a number of issues with this method of analysis. Firstly, the set value of the ADE is based

on the costs determined in an inpatient hospital setting in the USA, and may substantially overestimate

the ADEs being considered in community pharmacy interventions. Secondly, there are a number of

interventions (such as the Aspirin and Paediatric antibiotic examples) that do not have any probability

of preventing an ADE, and indeed may increase the likelihood of an ADE (e.g. increased bleeding or

GI effects form the addition of aspirin, or diarrhoea from the antibiotic dose increase). Thus, where an

ADE is being prevented, this method overestimates the value, and it underestimates the value of

interventions where increased efficacy is the outcome (see Table 6.1-13).


Underdose Probability of Adverse Drug Event (0,0.01, 0.1, 0.4,0.6)

01 0.6 0

Estimated Cost Avoidance

02 0.6 X $5006= $3004 02

Direct Drug Costs increased cost of

aspirin acquisition, $6.36

Saving of drug ceased, $32.98

Increased drug costs due to increased

dose,


negative (~$6) $3004 + ~$33=

US $3037 (2001) negative (~$15)


1No chance of ADE before drug was commenced 2Increased efficacy interventions are allocated zero cost avoidance

Table 6.1-13: Assessment of Sample Interventions Using Nesbit’s Method

24 Bates DW, Spell N, Cullen DJ, et. al. The costs of adverse drug events in hospitalized patients.

JAMA 1997; 277: 307-11.


Page 262 of 361

6.1.5.5 Dooley Method

This hospital based study assessed the probability of readmission before and after the intervention,

and is the only other study we could fid that addressed the “before and after” impact of the situation

separately. Being, hospital based, length of hospital stay was considered as was the costs of medical

procedures and drugs. Dooley used AR DRG values for calculation of readmission costs, but also

included medical procedures (which are included in the DRG cost weights).

We assumed the interventions that were being assessed took place in a hospital setting in order to

attribute the probabilities. Dooley’s method seems to underestimate the value of the aspirin

prophylaxis intervention (see Table 6.1-14).


Underdose

Length of Stay No change

10% increase 3.7 to 4.1 days x

cost/day 0.37 X $9912

=$366.67

0.2 day increase x cost/day

0.2 * $991= $198.20

Probability of Readmission to Hospital

10% before intervention, 8% after

intervention (for expected DRG of

angina) 0.02 X $12781


intervention (for expected DRG of gastrointestinal

bleeding) 0.15 X $11993 =

$179.85


intervention (for expected DRG of

infection, paediatric) 0.02 X $1611

5 = $32.22

Medical Procedures or Laboratory Monitoring

Nil Blood tests and

possible endoscopy4 Nil

Drug Costs Increase of $2.50 Decrease of ~$33 Increase of ~$11


$25.56-$2.50 = $23.06

$366.67 + $179.85 + $33 =$579.52

$198.20 + $32.22 - $11 =

$219.42


1ARDRG F74Z 2002/3

2General bed day as described in paper

3ARDRG G61B 2002/3

4These costs would be included in the DRG cost if an admission was to result

5ARDRG T63B 2002/3

Table 6.1-14: Assessment of Sample Interventions Using Dooley’s Method


Page 263 of 361

6.1.5.6 Buurma Method

Buurma’s method involved 20 reviewers assessing whether the intervention improved the

effectiveness of therapy and/or prevented an adverse drug reaction. A probability (scale of 1 to 5) was

assigned for the outcome and a level of severity was also applied (scale 1 to 5). Results were

presented as a quadrant diagram of probability versus seriousness and were not converted into dollar

values.

The requirement for an intervention to either improve efficacy or prevent and ADR presented some

interpretation difficulty in cases where drugs were added for untreated indications (e.g. the aspirin

prophylaxis sample intervention). If the general intention of the statement is considered, then the

addition of treatment would improve the efficacy of management of the underlying disease or

symptom.


Underdose Improved Effectiveness

Yes1 No Yes1

Probability of improved

effectiveness (1-5) 42 na 3

Importance of Improved

Effectiveness (1-5) 2 na 3

Prevention of Adverse Drug Reaction (ADR)

No Yes No

Probability of prevention of

ADR(1-5) na 3 na

Seriousness of ADR(1-5)

na 4 na


Value of interventions described as coordinates on a quadrant diagram.


1May not meet the definition as aspirin was not present, so it is not “improved” effectiveness. 2Interpretation difficult, almost 100% probability of improved efficacy, but to what degree?

Table 6.1-15: Assessment of Sample Interventions Using Buurma’s Method


Page 264 of 361

6.1.5.7 PROMISe Method

As described in section 4.5.3, the PROMISe value assessment process was comprehensive and

involved 16 assessors allocating probability estimates for particular consequences at three different

levels of severity for the situation with and without the intervention. The average assessor results for

each of the three sample interventions are shown in Table 6.1-16, Table 6.1-17 and Table 6.1-18.

Attributability of the interventions were high, with the aspirin prohylaxis intervention rating ~60% and

the other two interventions rating ~95%.

intidHealth Status

Impact

Duration of

Health Impact

(Days)

Duration of

Hospital

Days Change

Number of

GP Consults

Number of

Specialist

Consults

Total Costs

00654K-128 1 -0.40 0.00 0.00 0.00 $14.55

00654K-128 2 -2.70 -0.03 0.01 0.00 -$4.51

00654K-128 3 -0.79 0.01 0.00 0.00 $12.41

Total -3.89 -0.01 0.00 0.00 22.44

20868H-5 1 -2.73 0.00 -0.06 $2.08

20868H-5 2 -13.71 -0.20 -0.21 -0.06 -$140.09

20868H-5 3 -6.20 -0.11 -0.11 0.00 -$94.79

Total -22.64 -0.31 -0.37 -0.06 -232.80

Average -13.26 -0.16 -0.18 -0.03 -$105.18

Table 6.1-16: Aspirin Prophylaxis Interventions as assessed in PROMISe Study

intidHealth Status

Impact

Duration of

Health Impact

(Days)

Duration of

Hospital

Days Change

Number of

GP Consults

Number of

Specialist

Consults

Total Costs

00654K-128 1 -17.72 0.00 -0.11 0.00 -$60.67

00654K-128 2 -19.60 -0.04 -0.48 -0.04 -$124.09

00654K-128 3 -4.16 -0.15 -0.10 0.00 -$111.63

Total -41.47 -0.19 -0.69 -0.04 -$296.39

20868H-5 1 -9.42 0.00 -0.08 0.00 -$35.98

20868H-5 2 -13.96 -0.02 -0.32 0.00 -$70.77

20868H-5 3 -1.18 -0.06 -0.02 0.00 -$42.80

Total -24.55 -0.08 -0.43 0.00 -$149.54

Average -33.01 -0.13 -0.56 -0.02 -$222.97

Table 6.1-17: NSAID Duplication Interventions as assessed in PROMISe Study


Page 265 of 361

intidHealth Status

Impact

Duration of

Health Impact

(Days)

Duration of

Hospital

Days Change

Number of

GP Consults

Number of

Specialist

Consults

Total Costs

00654K-128 1 -3.06 0.00 -0.20 0.00 -$5.63

00654K-128 2 -1.87 0.00 -0.12 0.00 -$5.58

00654K-128 3 -0.90 -0.24 -0.06 0.00 -$182.87

Total -5.83 -0.24 -0.38 0.00 -$194.08

20868H-5 1 -1.94 0.00 -0.13 0.00 -$3.54

20868H-5 2 -5.71 0.00 -0.38 0.00 -$17.11

20868H-5 3 -0.70 -0.19 -0.05 0.00 -$145.85

Total -8.35 -0.19 -0.55 0.00 -$166.49

Average -7.09 -0.22 -0.47 0.00 -$180.28

Table 6.1-18: Paediatric Antibiotic Underdose Interventions as assessed in PROMISe Study

A comparison of the values determined by each of the methods is shown in Table 6.1-19. This shows

that the there is significant variation in the results obtained as a result of the different methodological

issues discussed. We are confident that the PROMISe method is comprehensive and takes into

account the important factors associated with determining value of interventions.


Underdose

Rupp US$33 (1992) US$42 (1992) US$30 (1992)

Hawksworth Individual interventions were not valued, an overall value was expressed as a proportion of prescriptions that had a score of 1 or more in categories 1 or 2

Benrimoj $0.12 (1997) $5.67 (1997) $14.15 (1997)

Nesbit Negative US$6 (2001) US $3037 (2001) Negative US$15

(2001)

Dooley $23.06 (2003) $579.52 (2003) $219.42 (2003)

Buurma Value of interventions described as coordinates on a quadrant diagram.

PROMISe $105.18 (2005) $222.97 (2005) $180.28 (2005) Table 6.1-19: Value of Sample Interventions as Assessed by Different Economic Methods

6.2 Overview of Economic Results

Our clinical and economic analysis suggests that the value of Australian community pharmacist

interventions related to prescription medication, in terms of financial costs to the health system

prevented, is in the order of $350M each year, or $17.50 per capita.


Page 266 of 361

In addition;

• 262,000 hospital bed days are avoided (1.3 days per 1000 population) per annum,

• 53.2M days of adverse health impact are avoided (2.66 days per capita) per annum, and

• 1.48M consultations with GPs or specialists are avoided per annum

Our estimate is that 0.7 hours of a pharmacist’s time is spent undertaking 6.9 interventions for every

1000 prescriptions (extrapolates to 154,000 hours undertaking 1.6M interventions nationally each

year).

These estimates are based on a 72% attribution of the benefit to the pharmacist. That is, the estimate

takes into account the possibility that, had a pharmacist not performed the intervention, another health

professional would have intervened (see Table 6.2-1).

This estimate of value of community pharmacy interventions is based on:

• An estimate of rates of interventions in a sample of community pharmacies, on a per 1000

prescriptions basis (the PROMISe dataset) (1);

• Multiplied by an estimate of the average value of these interventions (2); and

• Extrapolation of these results to estimates of prescriptions dispensed nationally (3). (see (1), (2)

and (3) in Figure 6.2-1).

All Australian Pharmacies

220.1m prescriptions per year

PROMISe Sample

52 Pharmacies for 8 weeks

2396 Interventions435,000 prescriptions

PROMISe Sample

52 Pharmacies for 8 weeks

2396 Interventions435,000 prescriptions

Value of Interventions in PROMISe Sample (per

1000 scripts)

• 1.45 days in hospital• 7.1 consultations• $174 in total costs• 261 days of health loss

Value of Interventions in All Australian Pharmacies

• 363,000 days in hospital• 1.78m consultations• $436m in total costs• 65.3m days of health loss

(1)

(2)

(3)

Figure 6.2-1: Overview of Economic Analysis (see text for explanation)


Page 267 of 361

Our analysis also suggests that pharmacists may identify and act upon only a third to a half of all

possible opportunities for clinical intervention, and that the costs of these omissions is likely to be

greater than the benefits of the current rate of intervention. This claim is based on information gained

relating to:

• the rates of interventions in observed and non-observed periods;

• the rates of interventions on busy and less busy days;

• the rates of interventions with and without the aspirin intervention prompt; and

• on assumptions regarding the proportion of interventions performed by pharmacists that were

recorded on the PROMISe data base (i.e. recording rate).

Methods that could be examined, that may increase the rate of intervention, include:

• reducing pharmacist workloads (increasing staff levels);

• introducing prompts such as the aspirin alert;

• introducing payments for meeting selected intervention targets; and

• practice-based payments for improved intervention rates.

This analysis only provides an indication of the potential benefit of increasing intervention rates. An

accurate estimate of the extent of investments in workforce and in education required to achieve the

change in rates are beyond the scope of this study.

The potential additional value of an educational intervention prompt (similar to the aspirin popup),

increased pharmacist hours, and an increased rate of interventions is shown in Table 6.2-1.

Our estimate of the additional annual value of increasing pharmacist’s intervention rates to those

achieved in many of the pharmacies in the PROMISe project is:

• $606M in direct health costs,

• 749,000 hospital admission days avoided,

• 2.26M GP or specialist consultations avoided, and

• 91.8M days of adverse health impact avoided.


Page 268 of 361

Interven-tions

Hours on inter-vention

s

days of loss 3 pre-

vented

days of loss 2 pre-

vented

days of loss 1 pre-

vented

Total days

health loss pre-vented

Days in hospital

pre-vented

Consultations pre-

vented

Total costs pre-

vented

Estimate of value of current rate of intervention

Per 1000 scripts 6.9 0.7 26 109 95 230 1.13 6.4 1,508$

Per pharmacist hour 0.08 0.01 0.30 1.27 1.11 2.68 0.01 0.07 $ 17.63

National annual ('000s) 1,606 154 6,030 25,189 21,959 53,179 262 1,481 349,275$

per capita annual 0.08 0.01 0.30 1.26 1.10 2.66 0.01 0.07 17.46$ Estimate of ADDITIONAL value of aspirin pop up

Per 1000 scripts 8.2 0.79 47 114 47.89 208 2.30 4.51 1,379$



per capita annual 0.10 0.01 0.54 1.32 0.55 2.41 0.03 0.05 15.98$ Estimate of ADDITIONAL value of increased pharmacist hours

Per 1000 scripts 4.19 0.40 16 66 57 139 0.68 3.86 910$

Per pharmacsit hour 0.04 0.00 0.14 0.57 0.50 1.21 0.01 0.03 $ 7.95

National annual ('000s) 970 255 3,640 15,204 13,254 32,098 158 894 210,818$

per capita annual 0.05 0.01 0.18 0.76 0.66 1.60 0.01 0.04 10.54$ Estimate of ADDITIONAL value of maximum interventions in current hours

Per 1000 scripts 13.92 1.33 68 203 126 396 3.23 9.75 2,615$



per capita annual 0.16 0.00 0.79 2.35 1.45 4.59 0.04 0.11 30.28$

Summary of value of existing and changed rate of interventions

Table 6.2-1: Summary of existing and changed rate of interventions


Page 269 of 361

The $350M dollar value of current interventions (“a million dollars a day”) and the $955M value of

potential increased rate of community pharmacist interventions is a “headline” result. It is inevitable

that expressing the result of a study in such terms will lead to some researchers and commentators

assuming that there is an unwarranted confidence in the underlying analysis. This is a valid concern,

so it is worthwhile considering both why headline results cause some discomfort amongst researchers

and how these issues have been addressed in this analysis.

1) A simple result does not reflect the context and the richness of the quantitative data, let alone the

depth of qualitative data from focus groups. However, that does not mean that the underlying

analysis does not reflect this. In the case of the analysis performed for this study, the details of

the underlying analysis are documented in the body of this section of the report, and the scope of

factors considered in generating this simple result are apparent.

2) Unlike a less reductionist result, a simple headline requires first, a substantial number of

assumptions to be made and, secondly, that these assumptions are not listed with the headline

result. This is an inevitable consequence of the method used to present the results rather than

the quality of the analysis or results. The assumptions we have made are detailed in the report,

as are their justification. The Project Team believes that these assumptions have resulted in

conservative estimates of the economic benefits of pharmacists’ interventions. Furthermore, if the

reader of the report has access to the electronic version of both the report and the spread sheet

model, they can easily change the assumptions by accessing the spread sheet through the word

document.

Table 6.2-1 presents the summary results of the study of the economic value.

The consequence of a pharmacist intervention has a number of dimensions outlined below and in

Figure 6.2-2 :

1) The days of loss in health status prevented:

a. We have used a measure that is relative (loss relative to health status that would

otherwise have occurred) rather than absolute (the absolute health status that

occurred). By using a relative health status, we account for the range of possible

initial health states of patients.

b. In the absence of the intervention, the patient could suffer a severe, moderate or mild

loss in previous health status, for a given number of days.

2) The days in hospital prevented:

a. The number of days measure was used instead of number of admissions prevented,

as we considered the possibility that an admission could occur after an intervention

but the effect of the intervention is to shorten that stay.

3) The consultations with General Practitioners and specialists prevented:

a. These are the consequence of the health outcomes being prevented (for example

preventing a stroke through use of aspirin) and do not include the outcomes of the

actual intervention (for example seeing a GP to discuss the use of aspirin).


Page 270 of 361

4) The total financial costs to MBS and hospitals prevented:

a. These include hospital costs and costs to the MBS for consultations and

investigations.

b. Costs to PBS were not considered. Unlike consultations, the impact on

pharmaceuticals would be more complex and would have required more assumptions

to be made in both the estimate of an average value of intervention and the

extrapolation to a population. Furthermore, there is substantial variation in the costs to

the PBS of medications, and unlike consultations, it is difficult to determine an

appropriate average cost.

Hospital Admissions

DurationCost

General Practitioner

ConsultsNumber

Cost

Specialist ConsultsNumber

Cost

Investi-gations

Cost

4. Direct Costs to Health System

Days of Health

Status 1 (mild)

Days of Health

Status 2 (moderate)

Days of Health

Status 3 (severe)

1.Days of Adverse

Impact on Health

2. Admission Days

3. Number of GP and/or

Specialist Consults

Figure 6.2-2: Derivation of Main Value Indicators

Costs to the patient were not considered, because of the variation in the co-payment by patients, and

the number of patients who are bulkbilled is likely to be high. In addition, we found that the technique

we used to value the consequences of the cases where pharmacists intervened is not amenable to the

inclusion of costs to patients, apart from an average co-payment for a consultation. If preferred, an

average co-payment of $10 (for example) could be applied to the figure for consultations prevented, of

1.48M, to provide an estimate of $14.8M in costs to patients prevented. However, it may be that some

of these consultations would have otherwise occurred.

We tested a series of alternative values for the key assumptions used in the analysis. These six

alternatives are descried in Table 6.2-2. This document has been designed to ensure that if preferred,


Page 271 of 361

the user can run alternative values of key hypotheses. In relation to a general sensitivity analysis on

the value we have assigned to the consequences, a proxy for reducing this value is to reduce the

attribution to the pharmacist, currently 72%.

Scenario Description of Key Assumptions Varied in the Scenario

Base Case

Estimate A Conservative Set of Assumptions as described in the text.

1

Assumes that the recording rate for interventions on observed days is 100% and that on unobserved days is 75%. This has the effect of reducing the difference between the recorded and actual rates of interventions.

2 Assumes an increased attribution rate of 95%, having the effect of increasing the average value of each intervention.

3 Assumes a decreased attribution rate of 65%, having the effect of decreasing the average value of each intervention.

4 Assumes that the Aspirin intervention prompt will only be as effective as it was on unobserved days (best estimate case uses the observed day rate for Aspirin intervention rate).

5

Uses a set of extreme assumptions:

• A recording rate in unobserved pharmacies of 75% of all interventions,

• the minimum possible value of benefit as assigned by the clinical panel and

• an attribution rate of 65% to the pharmacist.

6 Assumes a higher rate of other interventions (non-aspirin interventions) consistent with rates achieved by some pharmacies in the study.

Table 6.2-2: Alternative Assumptions Used in Determining Sensitivity Analysis

The results of these six alternative scenarios are presented in Table 6.2-3. The top half of the table

details the assumptions and the impact of these assumptions on the rate of intervention.

Scenarios 1 and 5 give the lower bound of current value of interventions, largely by assuming that the

rate of recording on unobserved days is higher than the initial assumption (75% compared to 50%).

The effect of this is to reduce the difference in the recorded rate and the actual rate of interventions.

In both of these cases, the gain in outcomes prevented remains significant, largely because these

gains are based on the rate of intervention on observed days, already assumed in the base case to

involve a high rate of recording. In all cases the gain from current to optimal recording is at least a

100% increase. The days in hospital gain is significant due to the effect of the aspirin intervention.


Page 272 of 361

Table 6.2-3: Sensitivity Analysis of Estimated Value of Interventions

Ba se 1 2 3 4 5 6

Assumptions

R e cord ing ra te

Observed days 90% 100% 90% 90% 90% 90% 90%

Non observed days 50% 75% 50% 50% 50% 75% 50%

E xtra po la tion to N a tio na l

(number of scripts)

Scripts d ispensed na tionally

(M) 232 232 244 244 244 232 232

Actua l ra te o f a sp irin

inte rve ntio ns pe r 1000 scrip ts

(recording rate ad justed)

Current 0 0 0 0 0 0 0

Under pop up 4.79 4.31 4.79 4.79 1.02 4.79 4.79

Under optima l 4.79 4.31 4.79 4.79 4.79 4.79 4.79

R a te o f o the r inte rve ntio ns

p e r 1000 scrip ts

(recording rate ad justed)

Current 6.93 4.62 6.93 6.93 6.93 4.62 12.37

under aspirin pop up 10.38 6.92 10.38 10.38 10.38 6.93 16.06

Under reduced pharmacy

work load 11.12 7.41 11.12 11.12 11.12 7.41 15.97

Under maximum 16.06 14.45 16.06 16.06 16.06 16.06 24.16

V a lue o f scrip ts

Max or Min va lue

(Min is 10% less) Max Max Max Max Max Min Max

attribution to pharmacist 75% 75% 95% 65% 75% 65% 75%

Outcomes (prevented)

P ha rma cist hours

Current 19.8 M 19.8 M 20.9 M 20.9 M 20.9 M 19.8 M 19.8 M

Under increased workforce 26.5 M 26.5 M 27.9 M 27.9 M 27.9 M 26.5 M 30.9 M

Curre nt situa tion

Days o f health loss('000) 53,179 35,453 70,953 48,547 56,015 27,653 94,850

per pharmacist hour 2.68 1.79 3.40 2.33 2.68 1.40 4.79

Days in hospita l ('000) 262 175 350 240 276 136 468


Financia l costs ('000) 349,275$ 232,850$ 466,010$ 318,849$ 367,903$ 181,623$ 622,965$

per pharmacist hour

Add itio na l e ffe ct o f a sp irin

inte rve ntio n



Days in hospita l ('000) 534 450 712 487 227 382 543


Financia l costs ('000) 319,512$ 247,110$ 426,300$ 291,679$ 215,284$ 204,146$ 331,999$

per pharmacist hour

Add itio na l e ffe ct o f incre a se d

hours



Days in hospita l ('000) 158 106 211 145 167 82 136


Financia l costs ('000) 210,818$ 140,545$ 281,278$ 192,453$ 222,061$ 109,625$ 181,423$

per pharmacist hour

Add itional FTE pharmacists 3,223 3,223 3,395 3,395 3,395 3,223 5,333

Add itio na l e ffe ct o f ma ximum

inte rve ntio n



Days in hospita l ('000) 749 735 999 684 789 652 850


Financia l costs ('000) 605,689$ 626,618$ 808,124$ 552,927$ 637,993$ 563,249$ 740,155$


Curre nt a s % o f ma ximum

Days o f health loss('000) 58% 37% 58% 58% 58% 32% 84%

Days in hospita l ('000) 35% 24% 35% 35% 35% 21% 55%Financia l costs ('000) 58% 37% 58% 58% 58% 32% 84%

S ensitivity ana lysis


Page 273 of 361

Table 6.2-4 demonstrates the minimum, base and maximum estimates for the current value of clinical

interventions in community pharmacies in Australia. As can be seen, even under very unlikely sets of

assumptions, the minimum estimate of value is significant.

Annual Estimates of Current Value of Interventions Value Indicator


Total Direct Costs Prevented

$182M $349M $623M


136,000 262,000 468,000

GP and/or Specialist

Consultations Avoided

0.77M 1.48M 2.64M

Days of Adverse Health Impact

Avoided 27.6M 53.1M 94.8M

Table 6.2-4: Minimum, Base Case and Maximum Estimates for Current Value of Community Pharmacists’ Clinical Interventions

6.3 Summary of method, data and results

The analysis we performed has five parts.

1. Estimate the rate of opportunity for interventions, viz. the rate of interventions under

perfect conditions (100% identification and action by pharmacists).

2. Estimate the actual rate of intervention in community pharmacies, per 1000 prescriptions.

3. Estimate the average value of an intervention and combine with 2) above to estimate the

total value of interventions recorded in the PROMISe study.

4. Estimate the rate and value of interventions nationally by extrapolating to national


5. Estimate the potential for increased or improved rate of intervention, combining each of

the above steps.

The key results from each of these steps, and the key methodological issues, are presented below.

We also provide a summary of the method we used within each step. Detailed methodological steps

are discussed in section 6.5.


Page 274 of 361

6.3.1 Step 1: Opportunity for Intervention

The main objective of this step was to determine the rate of opportunity for interventions in community

pharmacies if pharmacists were identifying and performing interventions at all possible opportunities.

To address this objective, we assessed information from a range of sources:

• The PROMISe data set – aggregated across information at the unit of pharmacy-days.

• Qualitative and quantitative information from trial observers in the pharmacies and qualitative

information from focus groups conducted with PROMISe participants.

• The estimate of rate of interventions on observed days compared to not observed days, and days

with and without the aspirin intervention prompt.

• The hours the pharmacy was open, and the hours worked by pharmacists, graduates and

dispensing technicians on each day.

• Estimates of rates of intervention by different levels of workload (measured as prescriptions per

pharmacist hour) and sorting data into 5 levels of activity.

• Estimates of rates of interventions by the number of different medications per patient (or more

than one).

A key assumption that we needed to make in this step of the analysis was the rate at which

interventions were recorded by pharmacists under a range of situations. We assumed that, while being

observed, 90% of all interventions were documented and that while not observed 50% of interventions

were documented. These estimates were based on data from the observers and feedback from

participating pharmacists. This analysis allows for variation of this assumed rate (can be varied in

Table 6.5-1).

Two key concepts were included in this step of the analysis:

• Clinical opportunity for intervention

o We used the rate of intervention per 1000 prescriptions, with and without an aspirin

intervention prompt, on observed days and in periods of low workload, as a guide to the

incidence of clinical opportunities per prescription that could be expected. That is, the

achievable maximum rate of interventions is guided by the rate achieved under the

circumstances where the pharmacist was not overly busy, and an observer was present.

We also considered the rate of interventions by number of medications per patients.

• Workload related opportunity.

o As workload increases, so does the opportunity to identify and act upon clinical

interventions. We considered the rate of interventions during busy and less busy periods

(defined as prescriptions per pharmacist hours) to estimate the effect of workload on the

incidence of interventions.


Page 275 of 361

6.3.1.1 Results of Opportunity for Intervention Step

The opportunity for clinical interventions is in the order of 16 to 20 interventions per 1000 prescriptions.

This is based on the rate of all clinical interventions on observed days with aspirin pop-ups. The lower

estimate is the rate for all interventions averaged over all levels of activity, adjusted for rate of

recording (see Table 6.6-16). The higher estimate is averaged over the two lowest quintiles of activity

for pharmacists, adjusted by recording rate (see Table 6.4-14, column 1 and 2, for observed aspirin

days, weighted average to adjust for the difference in prescriptions in each quintile, adjusted for 90%

recording rate). The first of these estimates was used as the base case for extrapolation (see Table

6.6-16), but other estimates of this opportunity for intervention can be specified in Table 6.5-2.

For patients who had more than one medication, the weighted average recorded rate of interventions

was higher than for the patients with one medication only (see Table 6.4-11). Intervention rates in

each quintile of workload activity for patients with more than one medication were five to ten fold

higher than intervention rates for patients with only one medication (see Figure 6.3-1).

8.15

3.942.54

1.81 1.59

46.5

29.01

17.4418.72

12.97

0

5

10

15

20

25

30

35

40

45

50

1 2 3 4 5

Quintile of Workload

Inte

rven

tio

n R

ate

(p

er

1000 p

rescri

pti

on

s)

with one medication

with more than one medication'

Figure 6.3-1: Intervention Rate for Patients with One or More Than One Medication Compared to Quintile of Workload

This result suggests that estimates of the opportunity for intervention for a given pharmacy will be

sensitive to the number of medications (and likely other characteristics) of patients.

The workflow adjusted opportunity is in the order of 19 interventions per 1000 prescriptions, using the

rate on the same days as for the opportunity for clinical intervention (days with pop-up, observed days,

adjusted for recording effect), but averaged over all levels of activity. This estimate can be specified

using Table 6.5-2.


Page 276 of 361

Further Research

A similar exercise with a larger number of observed days and observed hours would clarify the

optimum rate of interventions. A detailed review of each prescription dispensed and the nature of the

product and patient would also assist in determining the best estimate for optimum intervention rates.

6.3.2 Step 2: Rate of intervention in current practice

The objective for this step of the analysis was to estimate the rate at which a number of representative

pharmacies perform interventions, and to identify factors that result in variation in this rate, including

workload (prescriptions per hour while a pharmacist is working).

We used all of the data sources in the previous step to determine the results of this step of the

analysis. We assumed that the days that we examined in the PROMISe study were representative of

the days in these pharmacies. We also assumed, as before, that the rate of recording of interventions

on days where pharmacists were observed was 90%, and on days that were not observed was 50%.

By varying these assumptions, the extent to which the difference in the rate of intervention on these

two types of days (observed or unobserved) is due to reduced recording rather than reduced

performance of interventions is also varied (see Table 6.3-1). These assumptions can be varied in

Table 6.5-1.

Observed Rate 11.1 11.1 11.1 11.1 11.1 11.1

Unobserved Rate 3.7 3.7 3.7 3.7 3.7 3.7

Observed Recording Rate 90.0% 90.0% 90.0% 90.0% 90.0% 90.0%

Unobserved Recording Rate 30.0% 40.0% 50.0% 60.0% 70.0% 80.0%

Actual Rate in Observed Situation 12.3 12.3 12.3 12.3 12.3 12.3

Actual Rate in Unobserved Situation 12.3 9.3 7.4 6.2 5.3 4.6

Implied Reduced Rate of Performing Interventions 0.0 3.1 4.9 6.2 7.0 7.7

Varied Rates of Recording in Unobserved Pharmacies

Table 6.3-1: Effect of Varying Recording Rate Assumptions

6.3.2.1 Results of Current Rate of Intervention Step

The recorded rate of intervention, without an aspirin intervention prompt operating, and averaged over

all levels of activity, is 11.1 per 1000 prescriptions on observed days and 3.7 per 1000 prescriptions on

pharmacy days that are not observed.

The estimated actual rate of intervention per 1000 prescriptions, without an aspirin pop-up operating,

averaged over all levels of activity (adjusted for 50% recording in the unobserved pharmacies and

90% recording rate in observed pharmacies) is 12.3 on observed days and 7.4 on days that are not

observed.

The difference in the rates of 12.3 and 7.4 per 1000 prescriptions is by implication the result of

increased actual interventions on observed days rather than differences in rate of reporting. Hence,

the estimated actual rate of intervention in current practice is 7.2 per 1000 prescriptions. The 50%


Page 277 of 361

recording estimate is made on the basis of pharmacists’ self-reported estimate of the percentage of

clinical interventions documented (see section 5.8.11)

If the rate of reporting is assumed to be 30%, then the actual rate of intervention on unobserved days

is 12.3, the same as the rate on observed days. If the rate of reporting is assumed to be 90% on

unobserved days, then the estimated rate of actual interventions is 4.1 per 1000 prescriptions, and the

effect of observing is to increase rate of intervention 3-fold (see Table 6.3-1).

Further Research

The assumptions regarding the effect of changed rate of reporting vs. the effect of lower performance

of interventions on observed compared with non-observed days is a critical assumption. Improved

estimates of the rate of reporting interventions on observed and non-observed days would result in

better estimates of the actual intervention rate. A non-intrusive technique for capturing the activities of

pharmacists and matching these with the characteristics of each patient and prescription would be

necessary to gather this information.

6.3.3 Step 3: Value of an average intervention in current practice

The third step of the analysis was to determine the average value of community pharmacist

interventions. This was done using an indicator of value that considers benefits to patients in terms of;

• health loss

• days in hospital prevented and

• financial benefits to the heath care system, in terms of expenditure on MBS and hospital

admissions prevented.

In determining this value, we also considered the time taken to perform interventions as a measure of

the opportunity cost to pharmacies.

We assessed information concerning:

• the time taken to perform interventions (minutes per intervention)

• the consequences prevented in terms of indicators discussed above (see Table 6.2-1)

• an assessment of the probability that a series of consequences would have occurred in the

absence and presence of the intervention

• the attribution of the benefits of the intervention to the actions of the pharmacist.

Initially, we determined an estimate of the value of a series of possible consequences, in terms of the

indicators discussed above (see section 4.5.3.1.1). We then selected a series of interventions to be

assessed (see section 4.5.4) in terms of:

• the probability that a selected consequence will occur at a severe mild or moderate level, with and

without the intervention, and

• the probability that only the pharmacist could have performed the intervention before the

consequences had occurred.


Page 278 of 361

We then extrapolated the results of the clinical assessment panel assessments to all of the study’s

interventions.

A number of key assumptions were made in determining the average value of interventions:

• That the basis for determining the value of consequences, and assigning the pre and post

probability of these consequences, is appropriate.

• The attribution of the benefit to the pharmacist affects the overall value and can be varied in the

model.

• The extrapolation of assessed interventions to the whole of sample is appropriate.

• That the sample of interventions recorded in the PROMISe project is representative of national

interventions.

• That all interventions would result in patients acting on advice, including advice to go to their GP.

A measurement of the value of intervention was developed which comprised:

• days in any number of up to three types of health loss;

• days in hospital;

• general practitioner and specialist consultations; and

• the financial cost to the health care system – hospital costs and MBS.

This measurement is in contrast to an approach that simply assigns severe, moderate and mild as the

three possible consequences of an intervention. The procedure we developed consisted of assigning

the pre and post intervention probability of selected consequences, rather than assuming that the post

intervention probability is zero. This had the effect of attenuating the potential benefit of the

intervention as post intervention rates for consequences were frequently greater than zero. The

method also included assigning a probability of each level of severity of a series of consequences,

rather than a single probability of a single consequence, of a single severity (e.g. low risk of

myocardial infarction, a severe consequence). These assumptions can be varied in Table 6.5-1, Table

6.5-2, Table 6.5-4 and Table 6.5-6.


Page 279 of 361

6.3.3.1 Results of Determining Average Value Step

On average, an intervention took 6.5 minutes to complete (see Table 6.4-11). Aspirin interventions

were on average 4.6 minutes and other interventions averaged 6.6 minutes in duration.

In the lowest two quintiles of workload activity, the number of minutes per intervention were 8.0 (lowest

quintile) and 7.2 minutes (second lowest quintile) compared to 5.6 and 5.0 minutes in the top two

quintiles.

Interventions for patients with more than one medication took an average of 6.6 minutes compared to

5.3 minutes for interventions relating to patients with only one medication.

The number of minutes spent on interventions per pharmacist hour was 0.55 in the lowest quintile of

workload and 0.30 in the highest quintile.

Intervention time is only recorded if an intervention occurs. There may be investigations and actions

that the pharmacist undertakes, such as with the aspirin popup, that do not lead to an intervention. It

is therefore likely that the time spent in relation to interventions is greater than that indicated by our

results. However, the inverse relationship between level of activity and the time spent on the

intervention is likely to be maintained.

The economic cost of the intervention to the pharmacy in terms of pharmacist time is estimated as

$0.28 per hour of pharmacist time (assuming that the dollar value of a pharmacist hour is $45).

Without the intervention, the average situation would lead to a series of clinical consequences of the

following value (this example excludes aspirin interventions; see Table 6.4-24) :

o A loss of days in current health status of 7.6 with a severe loss, 32 with a moderate loss

and 32 with a mild loss, a total of 71 days of impaired health status.

o 0.3 days in hospital at a cost of $252

o 1.5 GP consultations, 0.3 specialist consultations at a total cost of $87 to MBS and a

further $82 in investigations.

o A total costs to the health system (MBS and hospitals) of $421

With the intervention, the consequences are:

o A loss of days in current health status of 2.5 with a severe loss, 11 with a moderate loss and 13 with a mild loss, a total of 27 days.


o 0.5 GP consultations, 0.1 specialist consultations (cost of $28 to MBS), further investigations at a cost of $24.50, making a total cost of $49 to MBS.

o A total cost to the health system (MBS and hospitals) of $131.


Page 280 of 361

Therefore, as a result of each intervention, there is a reduction of:

o 44 days in a lowered health status (5 days of level 3, 21 days of level 2 and 18 days of

level 3),


o 1.0 GP consultations and 0.23 specialist consultations at a cost of $59 to MBS,

o further investigations at a cost of $57 to MBS, and

o $283 in total costs (MBS and hospital combined).

Without the action of the pharmacist, in an average of 72% of cases, there would be no other health

professional who would have performed the intervention. The economic value of the consequences of

the pharmacist intervention has been adjusted to account for this.

Further Research

We were not able to compare the method we used to value the intervention, with the method used in

other studies. If we were able to do this, we could determine whether the result produced is more likely

to produce an accurate estimate, and whether other approaches are over-estimating the value of

pharmacists’ interventions as a result of the choice in method.

An additional study would be required to assess the costs to patients, as this would require that

patients were surveyed as well.

A more accurate estimate could be made by assessing a larger number of interventions.

A more accurate estimate of the value of an intervention could be facilitated by exploring the reasons

for differences between clinical assessors.

6.3.4 Step 4: Value of Pharmacist Interventions Extrapolated to National Situation

The objective of this step was to extrapolate the estimates of rate and value of interventions from the

PROMISe study to the national level. We used the results of the previous analyses and the number of

prescriptions dispensed each year from pharmacies across Australia. We assumed that the studied

pharmacies and their patients were representative of the national situation. These estimates are varied

in Table 6.5-6.

6.3.4.1 Results of National Extrapolation

We estimate that 1.7m clinical interventions would be undertaken at a rate of 1 intervention every 12

hours worked (Table 6.4-3.and Table 6.4-11). The financial value of these interventions to MBS and

hospitals is $436m, and 363,000 days in hospital are prevented. In addition, 65m days of adverse

health impact are prevented.


Page 281 of 361

Further Research:

Additional data on the characteristics of pharmacies and their patients across Australia would improve

our assumption regarding the representativeness of the study pharmacies and the extrapolations to

the national level.

6.3.5 Step 5: Value of improved rate of intervention.

The objective of this step was to estimate the benefits, and in one case the costs, of an increased rate

of intervention and to consider the feasibility of achieving an increased rate given the variation of rates

at current levels of activity.

To examine this objective, we used elements of the previously calculated steps, with additional

consideration of the rates of interventions under varying work loads and with and without the aspirin

intervention prompt.

We put in place assumptions regarding the current rate of interventions and the rate of opportunity for

intervention (see previous). These assumptions were interrelated with assumptions regarding the rate

of recording with and without observation. All of these assumptions can be varied in the following

tables: Table 6.5-4, Table 6.5-1, Table 6.5-3, Table 6.5-7, Table 6.5-8, Table 6.5-2 and Table 6.5-6.

Additional assumptions for this section are that:

• The average value of a current intervention is no greater than the average value of interventions

that are currently not occurring.

• Our estimate for extrapolation of sample results to national estimates is appropriate.

• Our estimate for the level of attribution of the intervention benefit to pharmacists is appropriate.

6.3.5.1 Results of Increasing Intervention Rates

For each of the effects that are modelled below, there are specific sets of assumptions. These

underlying assumptions have been varied according to our understanding of the data, and we show

below the results for one set of assumptions (those we consider the best estimate). However, for each

of these scenarios for increased intervention, a number of other assumptions regarding the

intervention rate after changed situations are possible. These assumptions can be varied in the tables

above (Table 6.5-4, Table 6.5-1, Table 6.5-3, Table 6.5-7, Table 6.5-8, Table 6.5-2 and Table

6.5-6.).

Our results indicate that:

If an aspirin intervention prompt (or something similar) is introduced, and all other factors such as

workload are constant, then there will be the following ADDITIONAL benefit at the national level per

annum:

o An additional 48.2M days of health loss prevented, 2.41 per capita.

o An additional 534,000 days in hospital prevented, 0.03 per capita.

o An additional 1.04M consultations with GPs or specialists prevented

o An additional $320M in financial savings to MBS and hospitals($16 per capita).


Page 282 of 361

If the workload of pharmacists is reduced to be similar to that of the average for the second two lowest

quintiles, and there is no-aspirin popup, then the rate of interventions will increase and the

ADDITIONAL benefits and costs per annum will be:

o An additional 32M days of health loss prevented, 1.60 per capita.


o An additional 894,000 GP and/or specialist consultations prevented.

o An additional $211M in financial savings to MBS and hospitals.

These benefits would require an additional 3,750 FTE pharmacists at a cost of $325M in salaries,

seeTable 6.6-14.

If an optimal level of intervention was to occur, at the rate specified in step one, the ADDITIONAL

benefit per annum is:

o An additional 92M days of health loss prevented, 4.59 per capita.


o An additional 2.26M GP or specialist consultations prevented

o An additional $606M in financial savings to MBS and hospitals, $30 per capita.

Further Research

In order to better estimate the extent of any potential increase in intervention rate, further research

would be required. It would be necessary to simulate the effect of other prompts that may be devised

and also to estimate the sustained effect of an aspirin or similar intervention prompt.

Further work would need to be done to be able to distinguish between the one-off benefits of, for

example, an aspirin intervention prompt, from the longer terms benefits of reduced work load. For

example, the benefits of an aspirin popup may reduce over time, as the number of people who are

identified as being able to benefit from the interventions reduces. In contrast, the benefits of reduced

workload are likely to be sustained.

6.4 Data set for the economic analysis

The economic analysis uses data from the PROMISe data base, on pharmacy rosters provided by

each pharmacy, the clinical panel that assessed the samples of interventions and the consequences

data base that assessed the dimensions of each possible consequence of the situation that lead to the

pharmacist intervention. The economic analysis used a data set constructed from the individual

record data set. In this section we present the details of the data set for the economic analysis.

6.4.1 The Key Descriptors

The economic analysis treated each pharmacy day as a separate unit record. It excluded all days for

which data was recorded but on which the pharmacy was not actively participating in the trial. It also

excluded days for which there was no information on prescriptions processed by the pharmacy.

For each of the 2,653 pharmacy days, we had information on the number of hours the pharmacy was

open, how many hours were worked by graduates (trainee pharmacists), pharmacists and dispensing


Page 283 of 361

technicians. We used this data to estimate the prescriptions per hour on that day, a key indicator of

activity which we used in our main analysis. We also had information for each day on number of

prescriptions, new prescriptions versus repeats, patients with more than one medication, whether

there was an intervention prompt installed on that day, and whether observed on that day or at some

day previously. We also had information on the number of clinical interventions that day, and how

many of these were related to the aspirin popup. Finally, we had information on the number of minutes

each of these interventions took.

The first summary table, Table 6.4-1, describes the days we analysed in terms of characteristics such

as the number of these days where observers were present, and the average number of pharmacist

hours for each hour the pharmacy was open. On average, dispensing technicians were working 16%

of the time, and for every hour a pharmacist worked, a dispensing technician would be there for 6.6

minutes.

This table can also be used as a guide to the question of how representative the pharmacy days we

analysed are of a full year of pharmacy activities in all pharmacies in Australia.

Pharmacy days 2,653

Pharmacy hours (Hours open) 24,694

Avg hours a day 9.3

Pharmacist hours (Hours worked) 35,941

Avg pharmacist hours per hours open 1.5

Graduate hours 9,659

Avg grad hours per pharmacy hour 0.39

Avg grad hours per pharmacist hours 0.27

Technician hours 3,866

Avg tech hours per pharmacy hour 0.16

Avg tech. hours per pharmacist hours 0.11

Observed days 171

As % of all days 6%

Ever observed days 1,159

As % of all days 44%

Days with aspirin interventions 1,598


Days with payments 1,984


Summary table: Days and hours

Table 6.4-1: Summary Table: Days and Hours

Interventions were proactive or reactive and, importantly, related to the aspirin popup or not related to

this popup. While only 8% of the clinical interventions were aspirin interventions, these occurred

almost entirely on days on which the aspirin popup was operating (see Table 6.4-2).


Page 284 of 361

Reactive interventions 469

As % of all clinical interventions 20%

Proactive interventions 1,904

As % of all clinical interventions 80%

Clinical Interventions 2,373

As a % of all scripts 0.56%

Aspirin interventions 200

As % of all clinical interventions 8.43%

% of all above on aspirin days 99.5%

Summary Tables: Interventions

Table 6.4-2: Summary Table: Interventions

Our sample pharmacy days had on average 158 prescriptions per day and 11.7 prescriptions per hour

a pharmacist worked. Prescriptions per pharmacist hour and pharmacy day varied substantially, and

were an important determinant of the rate of intervention. This rate was used as part of the more

detailed analysis. While there was also some variation in the proportion of prescriptions that were new

prescriptions, time has not allowed an in-depth analysis of the effect of this on the rate of intervention.

Patients had around 1.7 prescriptions each (see Table 6.4-3 ).

Summary Table: Scripts

New scripts 221,127

As % all scripts 53%

All scripts 420,152

Avg scripts per day 158

Avg scripts per pharmacy hour 17.0

Avg scripts per pharmacist hour 11.7

Avg scripts per patient 1.7

Table 6.4-3: Summary Table: Prescriptions

Around a third of all patients were using more than one medication, not all of which they may have

been dispensed a prescription for on the day. Around 7 patients were dispensed prescriptions on

average for each hour the pharmacists worked (see Table 6.4-4).

Summary Table: Patients

Patients 250,118

Avg patients per day 94

Avg patients per pharmacy hour 10.1

Avg patients per pharmacist hour 7.0

Patients with more than one medication 91,303

As % of all patients 37%

Patients with more than two medications 39,565

As % of all scripts 16%

Table 6.4-4: Summary Table: Patients


Page 285 of 361

6.4.2 Tables of Pharmacy Activity

We used the information on weekly rosters for the pharmacies to estimate the number of hours worked

by graduates, dispensing technicians and pharmacists each day. We then allocated each pharmacy

day into one of five quintiles, from quintile one, which contains the 20% of pharmacy days that had the

lowest activity on prescriptions per pharmacist hour basis, to quintile 5 which had the highest level of

activity. We also calculated three additional sets of quintiles, resulting in patients, then prescriptions,

per pharmacist hours and patients and prescriptions per staff hour. While there were some

differences, these were not in the order that would result in different estimates of the value of

pharmacists’ interventions. Furthermore, we can only extrapolate from prescriptions per pharmacist or

staff hour to the national level, as we only know the number of prescriptions at a national level, not the

number of patient visits.

Table 6.4-5 presents the key summary indicators of activity for each of the quintiles. Each quintile had

530 or 531 pharmacy days and the lowest quintile had 6 prescriptions per pharmacist hour and 18.8

prescriptions per pharmacist hour for the highest quintile. Up to 47 separate pharmacies where

represented in each quintile, indicating that most pharmacies had busy and less busy days. The

indicators we used required us to consider days when the aspirin popup was present (aspirin days)

and days when it was not present.


Page 286 of 361

Scripts per pharm

hour - QUINTILES 1 2 3 4 5 Total

Avge scripts per

pharmacist hour 6.0 8.7 11.1 14.1 18.8 11.7

Unique pharmacies

per quintile 34 45 46 47 42

Pharmacy days 530 531 531 531 530 2,653

Aspirin days 319 376 363 320 220 1,598

Non-aspirin days 211 155 168 211 310 1,055

Staff hours 8,113 10,202 9,778 10,576 10,798 49,465

Aspirin days 4,583 6,539 6,467 6,123 4,016 27,726

Non-Asprin days 3,530 3,663 3,311 4,453 6,783 21,739

Pharmacist hours 6,559 7,881 7,414 7,398 6,691 35,941

Aspirin days 3,928 5,335 4,949 4,261 2,555 21,026

Non-Asprin days 2,631 2,546 2,465 3,137 4,137 14,915

Scripts 39,382 68,835 82,422 104,035 125,478 420,152

Popup

Aspirin days 24,747 46,876 54,897 59,563 47,269 233,352

Non-Asprin days 14,635 21,959 27,525 44,472 78,209 186,800

Reimursment

Reim bursed 29,461 50,406 61,132 76,040 98,047 315,086

No reim bursem nt 9,921 18,429 21,290 27,995 27,431 105,066

Half of study

Firs t 21,190 35,712 41,713 55,056 57,409 211,080

Second 18,192 33,123 40,709 48,979 68,069 209,072

Patients 25,191 41,333 48,786 61,055 73,753 250,118

Number of Meds

one 17,062 26,408 30,668 38,133 46,544 158,815

m ore than one 8,129 14,925 18,118 22,922 27,209 91,303

Popup status

as rin days 16,281 28,713 32,717 34,800 27,145 139,656

not asprirn 8,910 12,620 16,069 26,255 46,608 110,462

Activity reported by quintiles (part 1 Scripts , patients , hours )

(pharm acy days sorted by scripts per pharm acis t hour

Table 6.4-5: Activity by Quintiles of Activity: Part 1

Table 6.4-6 summarises the number of days and number of prescriptions on each of these days, for

observed and non-observed days, and ever and never observed days. An observed day is a day

when a pharmacy was observed and ever observed is a day allocated to a pharmacy that had

previously had an observer.


Page 287 of 361

Quintiles 1 2 3 4 5 total

Observed days 25 40 34 48 24 171

Observed aspirin days 11 32 24 29 21 117

Observed non-aspirin

days 14 8 10 19 3 54

Non-observed days 505 491 497 483 506 2,482

Non-observed

aspirin days 308 344 339 291 199 1,481

Non observed non

aspirin days 197 147 158 192 307 1,001

Observed day scripts 2,608 5,486 6,603 10,899 6,114 31,710

Observed aspirin

day scripts 1,119 4,571 4,773 5,653 4,975 21,091

Observed non-

aspirin day scripts 1,489 915 1,830 5,246 1,139 10,619

Non- observed day

scripts 36,774 63,349 75,819 93,136 119,364 388,442

Non-Observed

aspirin day scripts 23,628 42,305 50,124 53,910 42,294 212,261

Non observed non-


Ever observed days 204 242 250 271 192 1,159

Ever observed

aspirin days 129 175 184 186 122 796

Ever observed non-

aspirin days 75 67 66 85 70 363

Never observed days 326 289 281 260 338 1,494

never observed

aspirin days 190 201 179 134 98 802

never observed non-

aspirin days 136 88 102 126 240 692

Ever observed day

scripts 16,280 32,450 39,039 55,322 45,583 188,674

Ever observed


Ever observed non-


Never observed day

scripts 23,102 36,385 43,383 48,713 79,895 231,478

never observed


never observed non-


Remuneration days 391 392 397 396 408 1,984

Days in first half

study 280 282 260 278 248 1,348

Proactive

intervntions 364 403 267 348 322 1,704

Patients with more

than one medication 17,062 26,408 30,668 38,133 46,544 158,815

Minutes on

interventions 3,606 3,485 2,072 2,436 2,027 13,626

Activity reported by quintiles (2 a, days and scripts, observation status)

(pharmacy days sorted by scripts per pharmacist hour

Table 6.4-6: Activity by Quintiles of Activity: Part 2a


Page 288 of 361

Table 6.4-7 shows the proportion of different days in the data that are in a range of categories across

the different quintiles of activity. There is a higher proportion of aspirin popup days in the second and

third quintile, and a lower proportion in the fifth quintile. These differences are taken into account as

the analyses looks at days with and without the aspirin pop-up separately.


Renumeration status

Renum erated days 74% 74% 75% 75% 77% 75%

Not renum erated

days 26% 26% 25% 25% 23% 25%

Asprirn day status

Popup 60% 71% 68% 60% 42% 60%

No popup 40% 29% 32% 40% 58% 40%

Number of

medications per

patient

Patients with One 57% 62% 63% 63% 63% 62%

Pateints with m ore

than one 43% 38% 37% 37% 37% 38%

Proactive/reactive

interventions

(ex asprirn)

Proactive interv. 74% 80% 77% 81% 81% 78%

Reactive interv. 26% 20% 23% 19% 19% 22%

By period of study

Days in firs t half 53% 53% 49% 52% 47% 51%

Days in second half 47% 47% 51% 48% 53% 49%

Activity reported by quintiles (part 2 b) % in quintile in each category


Table 6.4-7: Activity by Quintiles of Activity: Part 2b


Page 289 of 361

Table 6.4-8 presents the number of clinical interventions of any kind (either aspirin or other clinical

interventions) that occurred on aspirin and non –aspirin days, by observer status, and also by

characteristics of patients. The number of minutes taken for these interventions is also presented.


All interventions 517 537 394 498 427 2,373

Medications per

patient

one 139 104 78 69 74 464

tim e taken 785 662 403 284 311 2,445

m ore than one 378 433 316 429 353 1,909

tim e taken 3,338 3,219 2,030 2,490 1,814 12,891

Total tim e taken 4,123 3,881 2,433 2,774 2,125 15,336

Part of study

firs t half 347 389 271 362 285 1,654

second half 170 148 123 136 142 719

Reactive/proactive

Proactice 388 438 313 414 351 1,904

Reactive 129 99 81 84 76 469

Reimbursment

reim bursed 351 344 260 310 310 1,575

not reim bursed 166 193 134 188 117 798

Intervention by pop

up status

Aspirin days 325 405 287 357 231 1,605

observed 39 104 69 101 83 396

non-observed days 286 301 218 256 148 1,209

ever observed 214 270 174 293 208 1,159

never observed 111 135 113 64 23 446

Non-aspirin days 192 132 107 141 196 768

observed 36 15 17 39 11 118

non-observed days 156 117 90 102 185 650

ever observed 107 68 46 91 48 360

never observed 85 64 61 50 148 408

All days 517 537 394 498 427 2,373

observed 75 119 86 140 94 514


ever observed 321 338 220 384 256 1,519

never observed 196 199 174 114 171 854

Activity reported by quintiles (part 3a, all interventions)


Table 6.4-8: Activity by Quintiles: Part 3 a), All Interventions


Page 290 of 361

Table 6.4-9 presents the number of aspirin interventions on a range of days, and patient type. There

were almost no aspirin interventions recorded on days when the aspirin pop-up was not operating.


Aspirin interventions 24 35 46 66 29 200

patients with one

m edication 2 4 1 2 4 13

tim e taken 6 7 2 12 8 35

patients with m ore

than one

m edication 22 31 45 64 25 187

tim e taken 146 149 195 306 96 892

Total tim e taken 152 156 197 318 104 927

Part of study

firs t half 22 27 40 64 29 182

second half 2 8 6 2 - 18

Reactive/proactive

Proactice 24 35 46 66 29 200

Reactive - - - - - -

Reimbursment

reim bursed 18 21 30 39 15 123

not reim bursed 6 14 16 27 14 77

Aspirin days 24 35 46 65 29 199

observed 7 11 16 39 18 91


ever observed 16 20 32 63 29 160

never observed 8 15 14 2 - 39

Non-aspirin days - - - 1 - 1

observed - - - - - -

non-observed days - - - - - -

ever observed - - - 1 - 1

never observed - - - - - -

All days 24 35 46 66 29 200

observed 7 11 16 39 18 91


ever observed 16 20 32 64 29 161

never observed 8 15 14 2 - 39

Activity reported by quintiles (part 3 b) aspirin interventions)


Table 6.4-9: Activity by Quintiles: 3 b) Aspirin Interventions


Page 291 of 361

Table 6.4-10 presents the number of non-aspirin (other) interventions that were recorded on the

database by quintile.


Other interventions 493 502 348 432 398 2,173

patients with one

m edication 137 100 77 67 70 451

tim e taken 779 655 401 272 303 2,410

patients with m ore

than one

m edication 356 402 271 365 328 1,722

tim e taken 3,192 3,070 1,835 2,184 1,718 11,999

Total tim e taken 3,971 3,725 2,236 2,456 2,021 14,409

Part of study

firs t half 325 362 231 298 256 1,472

second half 168 140 117 134 142 701

Reactive/proactive

Proactice 364 403 267 348 322 1,704

Reactive 129 99 81 84 76 469

Reimbursment

reim bursed 333 323 230 271 295 1,452

not reim bursed 160 179 118 161 103 721

Aspirin days 301 370 241 292 202 1,406

observed 32 93 53 62 65 305


ever observed 198 250 142 230 179 999

never observed 103 120 99 62 23 407

Non-aspirin days 192 132 107 140 196 767

observed 36 15 17 39 11 118


ever observed 107 68 46 90 48 359

never observed 85 64 61 50 148 408

All days 493 502 348 432 398 2,173

observed 68 108 70 101 76 423


ever observed 305 318 188 320 227 1,358

never observed 188 184 160 112 171 815

Activity reported by quintiles (part 3 c) other interventions)


Table 6.4-10: Activity Reported by Quintiles: 3 c) Other (Non-Aspirin) Interventions

6.4.3 General indicators

The series of tables on activity (prescriptions, patients and hours) and clinical interventions (aspirin

and other) by day type (observed status and pop-up status), by pharmacist load (quintiles) and patient

type (one medication or more than one medication) were used to develop a series of indicators of the

rate of intervention.


Page 292 of 361

Table 6.4-11 presents a series of summary indicators, such as prescriptions per patient by quintile,

pharmacist hour and staff hour. Interventions occur at a much higher rate for people with more than

one medication, and take longer on average. Interventions take longer on average in the less busy

period (7.97 minutes) compared to 4.98 minutes for the most active quintile. The time spent on

interventions each hour also reduces as pharmacy activity increases, from 0.55 minutes per hour to

0.3 minutes per hour. The dollar value of intervention time, per pharmacist hour, is the dollar value of

the minutes spent by the pharmacist on interventions every hour. The dollar value used as the

estimate of a pharmacist’s hourly wage costs ($45) is presented and can be varied in Table 6.5-6.


Scripts

Per patient 1.56 1.67 1.69 1.70 1.70 1.68

Per pharm acis t

hour 6.00 8.73 11.12 14.06 18.75 11.69

Per s taff hour 4.85 6.75 8.43 9.84 11.62 8.49

Interventions

Per 1000 patients 20.52 12.99 8.08 8.16 5.79 9.49 with one

m edication 8.15 3.94 2.54 1.81 1.59 2.92 with m ore than

one m edication' 46.50 29.01 17.44 18.72 12.97 20.91

Minutes of

interventions

Min. per

pharmacist hour 0.55 0.44 0.28 0.33 0.30 0.38

Dollars value of

intervntion tim e,

per pharm acis t

hour 0.41$ 0.33$ 0.21$ 0.25$ 0.23$ 0.28$

Minutes per

intervention

Type of

intervention 7.97 7.23 6.18 5.57 4.98 6.46

aspirin 6.33 4.46 4.28 4.82 3.59 4.64

Other

interventions 8.05 7.42 6.43 5.69 5.08 6.63

Type of patient

patients with 1

m ed 5.65 6.37 5.17 4.12 4.20 5.27

patients with >

1 m ed 8.83 7.43 6.42 5.80 5.14 6.75

Patients

per pharm acis t hour 3.84 5.24 6.58 8.25 11.02 6.96

per s taff hour 3.11 4.05 4.99 5.77 6.83 5.06

Activity reported by quintiles (part 4 a) general indicators)


Table 6.4-11: Activity by Quintiles: 4 a) General Indicators


Page 293 of 361

Table 6.4-12 presents a series of indicators of recorded rate of aspirin interventions. The difference

in rate of intervention on observed compared to non-observed days is significant, even on days with

the aspirin pop-up operating. (4.3 per 1000 prescriptions compared to 0.5 on non-observed days).


Aspirin interventions 0.61 0.51 0.56 0.63 0.23 0.48

Patient medication

statusone, interventions

per 1000 patients 0.12 0.15 0.03 0.05 0.09 0.08

m ore than one -

interventions per

1000 patients 2.71 2.08 2.48 2.79 0.92 2.05

Per 1000 scripts

Part of study

firs t half 1.04 0.76 0.96 1.16 0.51 0.86

second half 0.11 0.24 0.15 0.04 - 0.09

Reimbursment

reim bursed 0.61 0.42 0.49 0.51 0.15 0.39

not reim bursed 0.60 0.76 0.75 0.96 0.51 0.73

aspirin days 1.0 0.7 0.8 1.1 0.6 0.9

observed 6.3 2.4 3.4 6.9 3.6 4.3

non-observed days 0.7 0.6 0.6 0.5 0.3 0.5

ever observed 1.9 0.9 1.1 1.8 1.1 1.3

never observed 0.5 0.6 0.5 0.1 - 0.3

non aspirin days

observed - - - - - -

non-observed days - - - - - -

ever observed - - - - - -

never observed - - - - - -

total days 0.6 0.5 0.6 0.6 0.2 0.5

observed 2.7 2.0 2.4 3.6 2.9 2.9


ever observed 1.0 0.6 0.8 1.2 0.6 0.9

never observed 0.3 0.4 0.3 0.0 - 0.2

Activity reported by quintiles (part 4 b) indicators, aspirin interventions per 1000

scripts)


Table 6.4-12: Activity by Quintiles: 4 b) Aspirin Interventions, Rate per 1000 Prescriptions


Page 294 of 361

Table 6.4-13 presents the summary indicators for the recorded rates of other interventions. There is a

significant variation across quintiles of workload and between aspirin days and non-aspirin days. The

difference in observed and not observed days is also significant and a key assumption we make in the

analysis is what proportion of this is due to better recording of intervention activity when observed, and

what proportion is due to less performance of interventions when not observed.


Other interventions 12.5 7.3 4.2 4.2 3.2 5.2

Patients

one, per 1000

patients 8.0 3.8 2.5 1.8 1.5 2.8

m ore than one - per

1000 patients 43.8 26.9 15.0 15.9 12.1 18.9

Part of study

firs t half 15.3 10.1 5.5 5.4 4.5 7.0

second half 9.2 4.2 2.9 2.7 2.1 3.4

Reactive/proactive

Proactice 9.2 5.9 3.2 3.3 2.6 4.1

Reactive 3.3 1.4 1.0 0.8 0.6 1.1

Reimbursment

reim bursed 11.3 6.4 3.8 3.6 3.0 4.6

not reim bursed 16.1 9.7 5.5 5.8 3.8 6.9

aspirin days 12.2 7.9 4.4 4.9 4.3 6.0

observed 28.6 20.3 11.1 11.0 13.1 14.5


ever observed 23.0 10.9 5.0 6.7 7.1 8.4

never observed 6.4 5.0 3.7 2.5 1.0 3.6

non aspirin days 13.1 6.0 3.9 3.1 2.5 4.1

observed 24.2 16.4 9.4 7.4 9.7 11.1


ever observed 14.0 7.1 4.2 4.3 2.4 5.2

never observed 12.2 5.1 3.7 2.1 2.6 3.5

total days 12.5 7.3 4.2 4.2 3.2 5.2

observed 26.1 19.7 10.6 9.3 12.4 13.3


ever observed 18.7 9.8 4.8 5.8 5.0 7.2

never observed 8.1 5.1 3.7 2.3 2.1 3.5

Activity reported by quintiles (part 4 c) indicators, other interventions per 1000 scripts)


Table 6.4-13: Activity by Quintiles 4 c) Other (non-Aspirin) Interventions per 1000 Prescriptions

Unsurprisingly, the rate of interventions for an intervention type varies by the status of the day and the

quintiles of pharmacist workload (see Table 6.4-13). There was a decrease in intervention actrivity

from 13.13 interventions per 1000 prescriptions to 3.40 interventions per 1000 prescriptions with

increasing workload (as defined by the quintiles of activity).


Page 295 of 361


All interventions 13.13 7.80 4.78 4.79 3.40 5.65

Patients

one, per 1000

patients 8.15 3.94 2.54 1.81 1.59 2.92

m ore than one - per

1000 patients 46.50 29.01 17.44 18.72 12.97 20.91

Part of study

firs t half 16.38 10.89 6.50 6.58 4.96 7.84

second half 9.34 4.47 3.02 2.78 2.09 3.44

Reactive/proactive

Proactice 9.85 6.36 3.80 3.98 2.80 4.53

Reactive 3.28 1.44 0.98 0.81 0.61 1.12

Reimbursment

reim bursed 11.91 6.82 4.25 4.08 3.16 5.00

not reim bursed 16.73 10.47 6.29 6.72 4.27 7.60

Aspirin days 13.13 8.64 5.23 6.01 4.89 6.88

observed 34.85 22.75 14.41 17.87 16.68 18.77


ever observed 24.81 11.77 6.18 8.48 8.23 9.69

never observed 6.88 5.64 4.23 2.56 1.05 3.92

non aspirin days 13.12 6.01 3.89 3.15 2.51 4.11

observed 24.18 16.39 9.40 7.43 9.66 11.13


ever observed 13.98 7.14 4.23 4.33 2.36 5.19

never observed 12.18 5.14 3.67 2.11 2.56 3.47

total days 13.13 7.80 4.78 4.79 3.40 5.65

observed 28.76 21.69 13.02 12.85 15.37 16.21


ever observed 19.72 10.42 5.64 6.94 5.62 8.05

never observed 8.48 5.47 4.01 2.34 2.14 3.69

Activity reported by quintiles (part 4 d) indicators, all interventions per 1000 scripts)


Table 6.4-14: Activity by Quintiles, All interventions per 1000 Prescriptions

6.4.4 Value of interventions

The value of interventions has a number of components.

1. The consequences that would be expected to occur had no intervention occurred. The range

of consequences is reported in Appendix 21

2. The consequences after the intervention (allowing for the possibility that there are still health

losses that may occur, even if the intervention is performed.)

3. The probability that a given consequence (eg a GI event) would manifest as severe (say

requiring hospitalisation), moderate (requiring a GP visit), or mild level.

4. The attribution of that benefit of an intervention to the actions of the pharmacist.


Page 296 of 361

5. The dimensions of a consequence, which were designed to go beyond a financial valuation,

are:

a. The days of loss in health status prevented:

i. We have used a measure that is relative (loss relative to health status that

would otherwise have occurred) rather than absolute (the absolute health

status that occurred). This is to account for the range in possible initial health

states of patients.

ii. In the absence of the intervention, patient could suffer a severe, moderate or

mild loss in previous health status, for a given number of days.

b. The days in hospital prevented:

i. measure was used instead of admissions prevented as we also considered

the possibility that an admission could occur after an intervention but the

effect of the intervention is to shorten that stay.

c. The consultations with GPs and specialists prevented:

i. These are the consequence of the health outcomes being prevented (for

example preventing a stroke through use of aspirin) and do not include the

outcomes of the actual intervention (for example seeing a GP to discuss the

use of aspirin).

d. The total financial costs to MBS and hospitals prevented:

i. These include hospital costs and costs to the MBS for consultations and

investigations.

ii. Costs to PBS were not considered. Unlike consultations, the impact on

pharmaceuticals would be more complex and would have required more

assumptions to be made in both the estimate of an average value of

intervention and the extrapolation to a population. Furthermore, there is

substantial variation in the costs to the PBS of medications and, unlike

medical consultations, it is difficult to determine an appropriate average cost.

iii. Costs to the patient were not considered, because of the variation in the co-

payment by patients, and the number of patients who are bulkbilled is likely to

be high. In addition, we found that the technique we used to value the

consequences of the cases where pharmacists intervened is not amenable to

the inclusion of costs to patients, apart from an average co-payment for a

consultation. If preferred, an average co-payment of $10 (for example) could

be applied to the figure for consultations prevented, of 1.8M, to provide an

estimate of $20M in costs to patients prevented. However, it may be that

some of these consultations would have otherwise occurred.

Table 6.4-21 presents the average effect of assessed consequences, separately for aspirin and non-

aspirin.

The method we used to develop these estimates is described in stages below.


Page 297 of 361

Stage 1. We used the clinical panel results data base (see section 4.5.3.4), the characteristics of

which are summarised in Table 6.4-15.

Number of interventions

assessed 291

Number of intervention

assessments 1779

Number of assessors 16

Average number of interventions

assessed per assessor 111

Total number of consequences

assessed 2445

Average number of

consequences per assessment 1.37

Attribution

Maximum 98%

Average 72%Minimum 10%

Summary of assessed interventions

Number of interventions

assessed 291

Number of intervention

assessments 1779

Number of assessors 16

Average number of interventions

assessed per assessor 111

Total number of consequences

assessed 2445

Average number of

consequences per assessment 1.37

Attribution

Maximum 98%

Average 72%Minimum 10%

Summary of assessed interventions

Table 6.4-15: summary of clinical panel data base

Stage 2. We then took two steps to aggregate the records relating to one intervention assessed by

several assessors and relating to several consequences, to a description of the value of a single

intervention: developing a weighted average probability for each consequence, and then assigning a

value to these consequences.

The panel data for a sample intervention was of the form in presented in Table 6.4-16. The

assessment is the probability, pre and post intervention of a severe, mild or moderate manifestation of

the selected consequences. The probability of no effect of a given consequence was derived form the

sum of the assigned probabilities for severe, moderate and mild. Attribution is the proportion of the

difference in pre and post probability attributed to the pharmacist.


Page 298 of 361

uid PollerType

Subgroup

Descriptio

n

Severe

_pre

Severe -

post

moder

ate_pr

e

modera

te_post

mild_p

re

mild_p

ost

Attrinbut

ion

19 GP Rash 1 1 15 5 35 20 90

19 GP

Allergic

reaction 1 1 5 3 10 5 90

20 GP Rash 8 7 0.5 2 0.5 35

20 GP

Allergic

reaction 7 8 0.5 5 2 35

21 GP Rash 1 2 25 25 20 20 90

22 GP Rash 50 30 20 30 100

23 GP Rash 10 80 5 10 10 100

24 Pharmacist Rash 50 100

24 Pharmacist

Allergic

reaction 50 100

25 Pharmacist Rash 5 70 20 20 20 100

27 Pharmacist Rash 0.3 0.1 20 5 60 2 99

28 Pharmacist Rash 100 95

28 Pharmacist

Allergic

reaction 50 95

30 Pharmacist Rash 5 60 35 80

31 Physician Rash 5 1 60 10 30 10 100

31 Physician

Allergic

reaction 5 1 60 10 30 10 100

32 GP Rash 10 5 10 5 100

33 Pharmacist Rash 5 1 25 1 50 1 95

34 GP Rash 5 15 75 10 95

35 Physician Rash 5 10 50 5 70

35 Physician

Allergic

reaction 1 0.2 5 0.5 20 0.5 70

Table 6.4-16: Clinical Panel Data for a Sample Intervention

Stage 3. We then derived an average assessment of the probability of a given consequence, at a

given severity level, for each assessed intervention. If no assessor included a given consequence for

a given intervention, it was given a weight of zero. If only one assessor out of, say, 5 identified a given

consequence for a given intervention, then this was averaged assuming that each of the other

assessors had given that consequence a probability of zero, that is, the results for that assessor were

weighted by 20% in this example. The probability for a given severity of a given consequence for a

selected intervention was calculated using an average of the assessors' responses, weighted for non-

response as described above.

The result of this exercise was an assessment for each intervention in the form presented below, but

with the addition of a severe, moderate and mild probability for each of the two consequences

identified, weighted as above.


Page 299 of 361

Table 6.4-17: Summary of an assessment of an intervention

Stage 4. The values in the Consequences table (see Appendix 21) were then linked with each of

these interventions and consequences to determine an average value pre and post for each

intervention. The example for the consequence “rash” is presented in Table 6.4-18.

Health loss

days in

health

loss

Duration

of

Admission

Cost of

Admission

Number

of GP

Consults

Cost of

GP

Consult

Number of

Specialist

Consults

Cost of

Specialist

Consults

Other

investigation

Rash

Severe 3 90 3.44 2166 1 37.775 2 192.15 0

Moderate 2 60 0 0 2 75.55 0 0

Mild 1 15 0 0 0 0 0 0

Table 6.4-18: Valuation of the consequence "rash"

Stage 5. Once aggregated across all the consequences and appropriately weighted by the number of

assessors who listed a given consequence and the probabilities they assessed to the outcome by

severity level (as described above), a valuation for each of the 291 assessed interventions was

calculated. The example intervention was assessed as in Table 6.4-19.

Intervention -

00654K-126

Duration of

loss 3

Duration of

loss 2

Duration of

loss 1

Days in

hosp.

Cost of

Admission

Number of

GP Consults

Cost of GP

Consult

Number of

Specialist

Consults

Cost of

Specialist

Consults

Other

investigatio

n

Pre 5.69 20.67 4.72 0.23 148.66 1.03 38.85 0.12 11.99 -

Post 0.30 2.96 1.30 0.01 8.99 0.14 5.18 0.01 0.61 -

Table 6.4-19: Example of the final valuation of an intervention

Stage 6. The results for assessed interventions were then extrapolated to the whole data set using

two different ways of classifying the interventions. The aspirin interventions were extrapolated

separately. The following table presents the figures used in the second extrapolation methods. The

results for each extrapolation were similar, but were used to generate maximum and minimum

valuations for average interventions, pre and post intervention.

Count assessors

Count of unique

consequences Allergic reaction

Rash

Intervention -

00654K-126 15 2 6 15


Page 300 of 361

Number in

assessed

Number in

remainder Total

1 COMPLIANCE 30 275 305

2 DRUG SELECTION 84 545 629

3 EDUCATION OR INFORMATION 22 418 440

4 MONITORING 7 47 54

5OVER OR UNDERDOSE

PRESCRIBED 76 465 541

6TOXICITY OR ADVERSE REACTION 51 264 315

7 UNTREATED INDICATIONS 21 382 403

total total 291 2396 2687

Summary of extrapolation from assessed to whole sample (Method 2)

Type

Table 6.4-20: Summary of extrapolation from assessed to whole of sample interventions

Table 6.4-21 presents the results for the assessed interventions (aspirin and non-aspirin interventions

separately), which can be compared with the extrapolated results in

Table 6.4-24. The minimum and maximum estimates from the clinical assessors are shown for the

“pre” situation (without the intervention) and the post situation (with the intervention). The effect of the

intervention (post subtracted from pre) is also shown.

Table 6.4-22 and Table 6.4-23 show how the results can be interpreted for aspirin and non-aspirin

interventions respectively. In both of these tables 1000 people with interventions are considered over

a one year period.

For 1000 people with Aspirin interventions, f there are 1000 people requiring non-aspirin interventions,

then without the intervention they would have 80% of days in an “underlying” health state, but 90% of

these days in a normal health state with the intervention.


Page 301 of 361

Days Health

loss 3

Days Health

loss 2

Days Health

loss 1

Total days

health loss

Days in

Hosp$ Adm ission

GP

Consults

$ GP

Consult

Specialis t

Consults

$ Special.

Consults

$Inves t-

igation Total cos t

Pre

Aspirin

Max 24.0 43.8 32.4 100.1 0.94 716.6 1.1 42.8 0.7 60.4 119.1 938.9

Min 21.4 38.0 31.8 91.2 0.83 634.7 1.1 39.9 0.6 54.3 113.1 841.9

Not aspirin

Max 7.7 33.1 31.8 72.6 0.31 240.1 1.6 59.1 0.3 30.4 76.0 405.7

Min 7.0 31.5 31.5 70.1 0.28 212.5 1.5 57.6 0.3 30.3 73.5 373.9

Post

Aspirin

Max 12.1 21.7 32.2 66.0 0.35 467.7 0.8 31.5 0.4 41.6 132.3 673.2

Min 11.9 21.4 30.5 63.8 0.31 437.5 0.7 28.0 0.4 38.6 118.9 623.0

Not aspirin

Max 1.6 9.2 13.4 24.2 0.05 52.9 0.5 18.3 0.1 7.8 22.6 101.6

Min 1.1 8.9 12.3 22.3 0.03 28.9 0.5 17.4 0.1 6.4 17.8 70.4

Effect

Aspirin

Max 12.01 22.34 1.93 36.3 0.63 279.11 0.39 14.76 0.25 21.81 0.19 315.9

Min 9.29 16.28 0.38- 25.2 0.48 167.01 0.22 8.32 0.15 12.69 19.30- 168.7

Not aspirin

Max 6.53 24.25 19.49 50.3 0.28 211.22 1.10 41.74 0.26 24.05 58.27 335.3

Min 5.87 22.64 19.29 47.8 0.24 183.61 1.07 40.27 0.25 23.89 55.68 303.4

D e ta ils o f a ve ra g e e ffe ct fo r a ve ra g e a sse sse d inte rve ntio n: a sp irin a nd o the r inte rve ntio ns

Table 6.4-21: Average effect of assessed interventions


Page 302 of 361

Pre Post Effect % reductionAttributed effect

Absolute Maximum Average Minimum

Total people with

interventions 1000 1000 1000 1000 98% 71% 25%

Total period (days) 365,000 365,000 365,000 365,000

Health status

Normal Health

status

Days 264,894 299,019 34,125 13% 33,443 24,229 8,531

% of total days 73% 82%

Loss level 1

Days 32,398 32,164 235 1% 230 166 59


Loss level 2

Days 43,755 21,725 22,030 50% 21,589 15,641 5,507


Loss level 3

Days 23,953 12,092 11,861 50% 11,624 8,421 2,965


Services

Days in hospital 941.96 348.43 593.53 63% 581.66 421.41 148.38

GP consultations 1,131.97 834.83 297.14 26% 291.20 210.97 74.29

Specialist

consultations 666.97 446.48 220.49 33% 216.08 156.55 55.12

Costs

Hospital $716,586 $467,718 $248,869 35% $243,891 $176,697 $62,217

Consultations $103,143 $26,095 $77,048 75% $75,507 $54,704 $19,262

Investigations $119,124 $22,614 $96,509 81% $94,579 $68,522 $24,127

Total costs $938,853 $516,427 $422,426 45% $413,978 $299,923 $105,607

Interpretation of above - aspirin interventions

Table 6.4-22: Interpretation of previous table: aspirin interventions


Page 303 of 361

Pre Post Effect Attributed effect

Absolute Maximum Average Minimum

Total people with

interventions 1000 1000 1000 98% 71% 25%

Total period (days) 365,000 365,000 365,000

Health status

Normal Health

status

Days 292,432 340,803 48,371 17% 47,404 34,344 12,093


Loss level 1

Days 31,752 13,402 18,350 58% 17,983 13,029 4,588


Loss level 2

Days 33,145 9,184 23,961 72% 23,482 17,013 5,990


Loss level 3

Days 7,671 1,611 6,060 79% 5,938 4,302 1,515


Services

Days in hospital 312 45 266 85% 261 189 67

GP cons ultations 1,565 485 1,080 69% 1,059 767 270

Specialis t

cons ultations 325.34 80.27 245.08 75% 240.18 174.00 61.27

Costs

Hospital 240,122$ 52,883$ 187,239$ 78% 183,494$ 132,940$ 46,810$

All interventions 89,544$ 26,095$ 63,449$ 71% 62,180$ 45,049$ 15,862$

Investigations 76,049$ 22,614$ 53,435$ 70% 52,366$ 37,939$ 13,359$

Total costs 405,715$ 101,592$ 304,123$ 75% 298,041$ 215,927$ 76,031$

Interpretation of whole of sample consequences - other interventions

%

reduction

Table 6.4-23: Interpretation of value of interventions: other interventions


Page 304 of 361

Days Health

loss 3

Days Health

loss 2

Days Health

loss 1

Total days

health loss

Days in

Hosp$ Adm ission

GP

Consults

$ GP

Consult

Specialis t

Consults

$ Special.

Consults

$Inves t-

igation Total cos t

Pre

Aspirin 21.4 38.0 32.4 91.8 0.83 660.4 1.1 39.9 0.6 54.3 113.1 867.5

Not aspirin 7.6 31.6 31.6 70.8 0.31 252.3 1.5 56.6 0.3 30.7 81.6 421.2

Post

Aspirin 11.94 21.42 32.16 65.5 0.35 486.61 0.83 31.54 0.45 41.57 132.35 692.1

Not aspirin 2.55 10.71 13.40 26.7 0.10 78.57 0.50 18.73 0.10 9.53 24.50 131.3

Effect

Aspirin 9.44 16.58 0.23 26.3 0.48 173.76 0.22 8.32 0.15 12.69 19.30- 175.5

Not aspirin 5.01 20.91 18.23 44.1 0.22 173.76 1.00 37.88 0.23 21.18 57.11 289.9

D e ta ils o f a ve ra g e e ffe ct o f a sse sse d inte rve ntio ns, e xtra p o la te d to the sa mp le

Table 6.4-24: Average effect of assessed interventions extrapolated to the sample


Page 305 of 361

Table 6.4-25 considers how sensitive the method we used to assess the value of interventions was.

On average, using our method, the assessed cases were assumed to have a value of 0.32 days in

hospital without the intervention and 0.1 days in hospital with the intervention. This would give an

outcome, prior to attribution to a pharmacist, of 0.12 days less in hospital. If the assessors had only

been able to assign severe, mild or moderate as an outcome of the intervention, that is, no pre and

post assignment and no probability of severe mild or moderate, these interventions would have been

assessed at 5.96 days in hospital if all were assessed as preventing a severe outcome, and 0.52 and

0 respectively for all moderate or all mild. Given that an assessor can specify the outcomes that can

be specified in a more simple approach (for example, by assigning 100% probability of a severe

consequence prior and 0% post), we would argue that our method will result in a more accurate

assessment of consequences, and probably a lower assessment than that which would be achieved

by other methods.

D ays in

hospita l D olla rs

E xisting

pre 0.32 415$

post 0.10 132$

V a ria tion 1

severe 5.96 5,202$

moderate 0.52 1,086$

mild - 96$

V a ria tion 2

severe 5.97 5,220$

moderate 0.49 1,077$

mild - 94$

Compa rison of scope o f

conse que nce s

Table 6.4-25: Comparison of scope of consequences by method used to elicit values

6.5 Economic analysis

The analysis we performed has five parts.

1) Estimate the rate of opportunity for interventions viz. the rate of interventions under perfect

conditions (100% identification and action by pharmacists).

2) Estimate the actual rate of intervention in community pharmacies, per 1000 prescriptions.

3) Estimate the average value of an intervention and combine with 2) above to estimate the total

value of interventions recorded in the PROMISe Melbourne study.

4) Estimate the rate and value of interventions nationally by extrapolating to national


5) Estimate the potential for increased or improved rate of intervention, combining each of the

above steps.

The methods used in the analysis and the key assumptions are presented in the following sections.


Page 306 of 361

6.5.1 Step 1: Opportunity for interventions

The opportunity for intervention was best estimated using a combination of two parameters – the rate

of intervention on observed aspirin days and the rate at which interventions were recorded on

observed days.

In relation to the assumption of the rate at which interventions were recorded, we considered two sets

of evidence. First, we looked at the rate of recorded interventions on days when the pharmacists were

observed, for aspirin and non-aspirin interventions. We then considered the likely rate at which these

were recorded on observed days. We considered both the pharmacists’ own estimates as reported in

the focus groups and the opinion of the observers. (see Table 6.5-1).

The rate at which interventions were recorded on non-observed days was similarly estimated and was

a key assumption in the development of estimates of the current rate of intervention in community

pharmacies. Differences in rates of recorded interventions on observed and unobserved days are

likely due to the combined effect of a lower rate of actual intervention and recording of these

interventions. The higher the estimate of recording on unobserved days, the greater the difference in

optimal rate of intervention (observed days) and actual rate of intervention (non-observed days).

The rate of 50% and 90% can be changed if this document is actively linked to the source files.

Rate of recording of actual interventions

Rate of interventions when observed

Aspirin (per 1000 aspirin day scripts ) 4.31

Not aspirin (per 1000 scripts ) 13.34

Rate of interventions when not observed first

Aspirin (per 1000 aspirin day scripts ) 0.51

Not aspirin (per 1000 scripts ) 4.51

Stated average rate of recording - source from focus

groups 75%

Current assum ption: Rate of recording when observed 90%

Current assum ption: Rate of recording when not

observed 50%

Key assumptions for extrapolations: part 2

Table 6.5-1: Key assumptions for extrapolation : Part 2

After making an assumption about the rate of recording, we decided which of the estimates of rates of

intervention on observed days were to be used. In relation to the estimate of aspirin interventions,

there was insufficient sample size to use the rates in the quintiles of activity on observed days, so we

used the rate of aspirin intervention on observed days when the aspirin popup was operating to

provide an estimate of the optimal rate of aspirin intervention. This figure was adjusted by the

recording rate of 90% to provide the estimate of optimal activity.


Page 307 of 361

The rate of other interventions was statistically significantly higher on days the aspirin pop-up

operated, compared to days when it did not operate. In addition, the rate of intervention was higher on

days when the pharmacist work load was smaller. We therefore generated three options based on the

indicators generated from the activity data, from which the assumed rate of optimal intervention can be

selected. (Table 6.5-2) Option 1 was used in the base case.

Rate of aspirin interventions

1. only one rate is poss ible as there is insufficient

sam ple s ize to cons ider effect of workload - this

is the rate of aspirin interventions with aspirin popup,

for observed days , averaged for all levels of activity 4.3

Rate of other interventions

(choose 1, 2 or 3) 1

1.

is the rate of other interventions , with or without popup,

for observed days , averaged for all levels of activity 14.5

2.

is the rate of other interventions for observed days , with

asprirn popup, avergaed for all levels of activty 19.7

3.

is the rate of other interventions , regardless of aspirin

popup, for observed days for average of lowes t and

second lowes t quintile of scripts per pharm acis t hour 21.7

Selected rate of intervention for aspirin 4.3

Selected rate of intervention for other interventions 14.5

Rate of recording applied to aspirin intervention

(either ob served or unob served) 90%

Rate of recording applied to other interventions


Estimate of level of activity with highest level of detection and

performance by pharmacist (per 1000 scripts)

Key assumptions for extrapolations: part 3 d)

Table 6.5-2: Key assumptions for extrapolations: 3 d)

6.5.2 Step 2: Current rate of intervention

The estimate of current rate of intervention required us to make two assumptions. The first is: the rate

of intervention from the general indicators we calculated. We considered rates on unobserved days,

averaged over all levels of activity, on days where there was no aspirin pop-up as these days were

most likely to be representative of existing practice. We also considered the option of never observed

days (days where there was no previous day that an observed was present, nor an observer on that

day.) We also included an option that considered the rate on observed days, in case there was a view

that the difference in observed and non-observed days was due entirely to differences in reporting

rate. (Table 6.5-3). We assumed that there would be almost no aspirin interventions in current

practice, which is reasonable given the results of our study.


Page 308 of 361

The second assumption we needed to make was the relationship between actual rate of intervention

and reported or recorded rate of intervention. In the base case we assumed that the rate of recording

on un-observed days was 50%, as presented in Table 6.5-1.


0 is the only option poss ible from our data 0


(choose 1, 2 or 3) : Note that the difference in the rate of

interventions on observed and non-observed days is

s tat. s ig, averaged across all levels of activity , so is the

difference between rate of other interventions on aspirin

and non-aspirin popup days 2

1.

is the rate of other interventions on days without aspirin

popup, for non-observed days , for all levels of activity 3.69

2.

is the rate of other interventions on days without

aspirin pop up , for never-observed days , for all levels of

activity 3.47

3.

is the rate of other interventions on days without

aspirin pop up , for observed days , for all levels of

activity 11.13

Selected rate of intervention for other interventions

used 3.47



Key assumptions for extrapolations: part 3 a)

Estimate of current level of activity

(pre effect of recording, per 1000 scripts )

Table 6.5-3: Key assumptions for extrapolation: 3 a)

6.5.3 Step 3: Average value of interventions

The average value of aspirin and non-aspirin interventions was estimated separately. The average

value was calculated by combining:

• the results of the clinical panel data (assigning consequences to each intervention and

probability to the severe, moderate and mild manifestations of these consequences),

• the consequences table (assigning value to consequences);

• the rate of attribution to the pharmacist (from the clinical panel data); and

• the estimates of rates of intervention from the activity data base – the PROMISe data base.

A range of possible pre and post intervention values of consequences was derived. This range was

based on two methods of extrapolation of the results of the interventions assessed by the clinical

panel to all interventions in the data base. The analysis allows for choice between a maximum and

minium effect, and for any rate of attribution of intervention to be assigned. In the base case we used


Page 309 of 361

the maximum effect of the intervention and an attribution of 75%. Changing attribution is the simplest

way to perform a sensitivity analysis.

The important point regarding the conservatism or otherwise of the method we used to value the

consequences of interventions is that it is likely to result in lower values for an average intervention

compared to the techniques used in similar studies. The fundamental reason is that the technique we

have used allows for greater flexibility in valuing an outcome, and was discussed in more detail in

section 6.4.4.

Value of average interventions (full attribution, non-

aspirin)

Maxim um

Cos ts prevented per intervention 211.22$

Minim um

Cos ts prevented per intervention 183.61$

Current Assum ption: Enter Max or Min Max

Attribution of effect to pharmacists

Maxim um attribution weighted average) 98%

Average attribution weighted average) 72%

Current assum ption: Attribution to pharm acis ts 75%


Table 6.5-4: Assumptions for extrapolation: Part 1

For the above set of assumptions, the average value of an aspirin or other intervention is presented

below. This is the value that was assigned to the rates of interventions derived from the PROMISe

data base in order to estimate the value of interventions per 1000 prescriptions and nationally.

Interventions Aspirin Other

Value (selected value) Max

Days in health state loss prevented

Loss 3 9.4 5.0

Loss 2 16.6 20.9

loss 1 0.2 18.2

Service use prevented

Days in hospital 0.48 0.22

GP and specialis t consultations 0.37 1.23

CostsCos ts to health sys tem prevented 175$ 290$

Value of consequences- full attribution

Table 6.5-5: Value of consequences for extrapolation: to be adjusted by attribution


Page 310 of 361

6.5.4 Step 4: National value of interventions

Without detailed information on the characteristics of Australian pharmacies and their patients, we

used the number of prescriptions dispensed in Australia as the main extrapolation variable.

Salary per hour

Pharm acis t $45

Prie deflator for each year s ince 2002/03

applied to the hospiatl cos ts - ets iam ted 2%

Dispensed scripts nationally

2002/03

Num ber HIC scripts nationally 173,000,000

Num ber of Total scripts dispensed from com m unity

pharm acies (es tiam te from AIHW) 220,000,000

HIC as % total - derived 79%

2004/05

Num ber of HIC (PBS and RPBS) 2004/05 183,000,000

Proportion of all s cripts that go through HIC

(copayem nt has increased over period so expect 79%

Total scripts nationally 231,645,570


Table 6.5-6: Key assumptions for extrapolation: 4

6.5.5 Step 5 Value of improved interventions

The first opportunity to increase the rate of intervention, and make some gains towards the optimal

rate of intervention, was the example of an aspirin pop-up. To simulate the effect of the aspirin pop-up

in practice, we could use the rate for either observed days with pop-up or unobserved days. In the

case of the aspirin pop-up, we decided to use the rate on observed days, for the following reasons:

• In a program, there could be intensive prompting through reminders, newsletters and

educational visits

• It could be tied to a payment for meeting a target rate of aspirin interventions.

We accept that in practice we would not have observers in pharmacies, but unlike the effect on

general interventions, we expect it would be feasible to use a targeted payment incentive.


Page 311 of 361


(choose 1 or 2) 1

1.

is the rate of aspirin interventions on days with aspirin

popup, for observed days , averaged for all levels of

activity 4.3

2.

is the rate of aspirin interventions on days with aspirin

popup, for unobserved days , averaged for all levels of

activity 0.5


(choose 1, 2 or 3)

Note that the difference in the rate of interventions on

observed and non-observed days is s tat. s ig, averaged

across all levels of activity , so is the difference between

rate of other interventions on aspirin and non-aspirin

popup days 1

1.

is the rate of other interventions on unobserved days

with aspirin popup, averaged for all levels of activity 5.2

2.

is the rate of other interventions on never-observed

days with aspirin popup, averaged for all levels of

activity 3.6

3.

is the rate of other interventions on observed days with

aspirin popup, averaged for all levels of activity 14.5

Selected rate of intervention for aspirin 4.31

Selected rate of intervention for other interventions 5.19

Rate of recording applied to aspirin intervention




Key assumptions for extrapolations: part 3 b)

Estimate of level of activity w ith aspirin pop up (pre effect of recording,

Table 6.5-7: Key assumptions for extrapolations: 3 b)

The second opportunity is to reduce the workload of pharmacists. The details of how we simulated the

effect of reduced workload are detailed in Section 6.6.3. In simple terms we assumed that the number

of prescriptions nationally remains the same, but the number of hours pharmacists work to dispense

these prescriptions is calculated using the rate of pharmacist hours per prescription for the lowest two

quintiles. (see Table 6.4-11). In the lowest two quintiles, pharmacists were dispensing 6 and 8.7

prescriptions per hour, and in the highest two quintiles, they were dispensing 14 and 18.7

prescriptions per hour. We assumed that the rate of aspirin interventions would be zero, as there

would be no popup. The analysis can be performed for either of the two options listed in Table 6.5-8.

The hours per 1000 prescriptions for the selected option is included in the table.


Page 312 of 361


0

0 is the only option poss ible from our data


(choose 1 or 2) 1

1.

is the rate of other interventions with no aspirin popup,

on unobserved days , for second lowes t quintile of

pharm acis t script per hour activity 5.6

2.

is the rate of other interventions with no aspirin popup,

on unobserved days , for average of second and lowes t

level of quintile of pharm acis t script per hour activity 8.0

Selected rate of other interventions 5.56

Hours per scripts for chosen activity level 114.48



Key assumptions for extrapolations: part 3 c)Estimate of level of activity with no aspirin pop up, but additional

pharmacist hours (pre effect of recording, per 1000 scripts)

Table 6.5-8: Key assumptions for extrapolations: 3 c)

6.6 Results of Economic analyses

The analysis described in the previous section has four sets of results:

1) The estimated value of current interventions

2) The additional value of introducing aspirin pop-ups

3) The additional value of reduced pharmacist workload

4) The additional value of complete identification and intervention

In this section we explore these results using a range of alternative assumptions.

The assumptions that are used in this analysis are referred to from Table 6.6-1 to Table 6.6-7. In

addition, it is assumed that the 435,520 scripts from this study and 52 participating pharmacies are

representative of the 230M scripts annually. See section 6.3.5 for details of these assumptions.

The results of the analyses that are detailed in the previous section, using the assumptions outlined

are summarised from Table 6.6-8 to Table 6.6-16. These tables address:

o The four alternative levels of intervention activity

i. current,

ii. with aspirin popup,

iii. increased pharmacist hours and workforce, and

iv. the rate of optimal or maximum intervention rate.


Page 313 of 361

o For each of four different indicators

i. intervention numbers,

ii. pharmacist hours,

iii. hours spent on interventions, and

iv. their dollar value

o Expressed as each of three different calculated rates:

i. per 1000 scripts,

ii. per annum (nationally), and

iii. per pharmacist hour.

o For both total and additional (improved intervention rate) scenarios compared to

current activity.

Table 6.6-1 presents intervention rates nationally and per pharmacist hour. Overall there were 6.93

interventions per 1000 scripts. This extrapolates to 1.6M interventions annually and 0.08 interventions

per pharmacist hour. The introduction of an aspirin popup could result in an increase in interventions (

both those related and those not related to the popup), from 8.2 to 15.2 interventions per 1000 scripts.

At this rate, there would be around 0.18 interventions per hour of pharmacist activity, (one intervention

every 5.6 hours). The optimal intervention rate is approximately four times that of the current rate.

Intervention rate

per 1000

scripts

Nationally

('000)

Per pharmacist

hour

Current

Interventions 6.93 1.61 M 0.08

Pop up

Additional 8.24 1.91 M 0.10

Total 15.17 3.51 M 0.18

Workforce

Additional int 4.19 0.97 M 0.04

Total int 11.12 2.58 M 0.10

Optimal intervention

Additional 13.92 3.22 M 0.16 Total 20.85 4.83 M 0.24

Interventions

per 1000 scripts, nationally and per pharmacist hour ,

summary by current and improved intervention rate

Table 6.6-1: Interventions Nationally and per Pharmacist Hour by Current and Increased Rate


Page 314 of 361

One of the changes to pharmacist activity that we simulated was to reduce the script load per hour for

pharmacists. If pharmacists work at the same workload (scripts per hour) as the average load in the

lowest two quintiles of activity, then 1000 scripts would be dispensed in 114 hours instead of 86 hours

currently. At a national level , given that the number of pharmacists is fixed, this would require an

additional 3,223 FTE pharmacists nationally (see Table 6.6-2). Whether this is feasible is

questionable.

Pharmacist hours

per 1000

scripts Nationally

Current

Hours 86 19,815,622

Total FTE pharmacists n/a 9,527

Workforce

additional hours (K) 29 6,704

total hours (K) 114 26,520

Additional FTE n/a 3

Total FTE pharmacists n/a 13

Pharmacist hours

per 1000 scripts and nationally , summary by current and

improved intervention rate

Table 6.6-2: Pharmacist Hours per 1000 Prescriptions and Nationally

The number of hours spent on these interventions is small (see Table 6.6-3). On a per 1000 scripts

basis, the current time spent on interventions for every 1000 scripts is 0.77 hours, which extrapolates

to 177,518 hours nationally and 0.009 hours (0.54min) per hour of pharmacist time. Even if the

additional interventions under each of the scenarios tested occurred, the maximum time per

pharmacist hour would be 0.023 hours (1.4min) per pharmacist hour. The average length of an aspirin

related intervention is shorter than other interventions and because aspirin interventions are not

performed currently, the percentage increase in number of interventions per hour, is greater than the

increase in the minutes per pharmacist hour.

This time does not include time spent investigating the situation, which in the case of some

combinations of patients and scripts, could be a number of minutes, without an actual intervention

being performed. If there are 2 minutes of pharmacist time for investigation on potential interventions

that do not occur, for every minute of actual intervention, then this would mean that would mean 1.8

minutes per hour of pharmacist time is currently spent on intervention related activity. At the maximum

level of activity, this would be 4.2 minutes of intervention related activity per hour of pharmacist time.


Page 315 of 361

Intervention hours

per 1000

scripts Nationally (K)

Per pharmacist

hour

Current

Total 0.77 177,518 0.009

Pop up

additional 0.75 173,897 0.009

total 1.52 351,415 0.018

Workforce

additional 0.46 107,147 0.004

total 1.23 284,665 0.011


additional 1.16 269,659 0.014total 1.93 447,177 0.023

Intervention hours,

per 1000 scripts, nationally and pharmacist hour

Summary by current and improved intervention rate

Table 6.6-3: Number of Hours Spent by Pharmacists on Interventions per 1000 Prescriptions and Nationally

The dollar value of this intervention time was calculated by multiplying the number of hours spent on

interventions by an estimated $45 per hour as the average salary, including on-costs, for pharmacists

(see Table 6.6-4). At the current level of intervention, this represents $0.40 per pharmacist hour at the

current level of activity and $1.02 per hour at maximum level of activity.

Dollar va lue of

pharmacist hours on

interventions

per 1000

scripts Nationally

Per pharmacist

hour

Current

Total $34.49 $7.99 M $0.40

Pop up

Additional $33.78 $7.83 M $0.39

Total $68.27 $15.81 M $0.80

Workforce

Additional $20.81 $4.82 M $0.18

Total $55.30 $12.81 M $0.48


Additional $52.38 $12.13 M $0.61

Total $86.87 $20.12 M $1.02

Dollar va lue of pharmacist intervention hours

per 1000 scripts, nationally and by pharmacist hour


Table 6.6-4: Dollar value of pharmacist intervention hours

The number of days of loss in health status prevented are currently 230 per 1000 scripts and 2.7 per

each hour a pharmacist works (see Table 6.6-5). There is a potential to increase this by 4.6 days of

health loss prevented, to a total of 7.3 days of health loss prevented for every hour of pharmacist’s

work.


Page 316 of 361

Days

per 1000

scripts Nationally

Per pharmacist

hour

Current

Total 230 53.2 M 2.7

Pop up

additional 208 48.3 M 2.4

total 438 101.4 M 5.1

Workforce

additional 139 32.1 M 1.2

total 368 85.3 M 3.2


additional 396 91.8 M 4.6total 626 145.0 M 7.3

Days of health loss prevented

per 1000 scripts, nationally and per pharmacist hour


Table 6.6-5: Days of Health Loss Prevented per 1000 Prescriptions, Nationally and Per Pharmacist Hour

Some of these days of health loss are days spent in hospital. For every 1000 scripts, there are 1.13

days in hospital prevented and an optimal intervention rate would result in an additional 3.2 days in

hospital prevented per 1000 scripts (see Table 6.6-6). Nationally, around 262,000 days in hospital are

currently prevented but an additional 749,000 days could be prevented if the rate of interventions were

improved.

Days

per 1000

scripts Nationally

Per pharmacist

hour

Current

Total 1.13 262.4 K 0.013

Pop up

additional 2.30 533.7 K 0.027

total 3.44 796.2 K 0.040

Workforce

additional 0.68 158.4 K 0.006

total 1.82 420.8 K 0.016


additional 3.23 748.7 K 0.038total 4.37 1011.2 K 0.051

Days in hospital prevented



Table 6.6-6: Days in Hospital Prevented per 100 Prescriptions, Nationaly and Per Pharmacist Hour

We estimated a total of $1,508 of health system costs per 1000 scripts are prevented, which is a rate

of $17.63 per hour of pharmacist’s work (Table 6.6-7). This could increase by $30.57 per pharmacist

hour if the rate of intervention were improved to the maximum level. Nationally, the costs prevented


Page 317 of 361

are currently $349M, but could increase by $606M to a total of $955M if the rate of interventions were

improved.

Dollars

per 1000

scripts Nationally

Per pharmacist

hour

Current

Total $1,508 $349.3 M 17.6

Pop up

additional $1,379 $319.5 M 16.1

total $2,887 $668.8 M 33.8

Workforce

additional $910 $210.8 M 7.9

total $2,418 $560.1 M 21.1


additional $2,615 $605.7 M 30.6total $4,123 $955.0 M 48.2

Financia l Costs to health system prevented



Table 6.6-7: Financial Costs to Health System Prevented per 1000 Prescriptions, Nationally and Per Pharmacist Hour

Table 6.6-9 presents the summary results of these estimates, in relation to the current situation.

Sections 6.6.1 to 6.6.4 include a more detailed examination and discussion of these results.

The main point is that the current value of pharmacists’ interventions in both health and financial terms

is already high, but there is scope for increasing this rate, as evidenced by the difference between the

current and our estimate of a possible optimal rate of intervention in community pharmacies. Both the

rate and the financial value of pharmacists’ interventions could be increased up to three-fold from the

existing amount.

It is also feasible to achieve some gain towards this optimal rate through the use of techniques such

as the aspirin pop-up and reduced pharmacist workload.


Page 318 of 361

Interven-tions

Hours on inter-vention

s

days of loss 3 pre-

vented

days of loss 2 pre-

vented

days of loss 1 pre-

vented

Total days

health loss pre-vented

Days in hospital

pre-vented

Consultations pre-

vented

Total costs pre-

vented

Estimate of value of current rate of intervention

Per 1000 scripts 6.9 0.7 26 109 95 230 1.13 6.4 1,508$



per capita annual 0.08 0.01 0.30 1.26 1.10 2.66 0.01 0.07 17.46$ Estimate of ADDITIONAL value of aspirin pop up

Per 1000 scripts 8.2 0.79 47 114 47.89 208 2.30 4.51 1,379$



per capita annual 0.10 0.01 0.54 1.32 0.55 2.41 0.03 0.05 15.98$ Estimate of ADDITIONAL value of increased pharmacist hours

Per 1000 scripts 4.19 0.40 16 66 57 139 0.68 3.86 910$

Per pharmacsit hour 0.04 0.00 0.14 0.57 0.50 1.21 0.01 0.03 $ 7.95

National annual ('000s) 970 255 3,640 15,204 13,254 32,098 158 894 210,818$

per capita annual 0.05 0.01 0.18 0.76 0.66 1.60 0.01 0.04 10.54$ Estimate of ADDITIONAL value of maximum interventions in current hours

Per 1000 scripts 13.92 1.33 68 203 126 396 3.23 9.75 2,615$



per capita annual 0.16 0.00 0.79 2.35 1.45 4.59 0.04 0.11 30.28$

Summary of value of existing and changed rate of interventions

Table 6.6-8: Summary of existing and changed rate of interventions


Page 319 of 361

6.6.1 Estimate of current value of pharmacist activity

The current value of community interventions as presented in Table 6.6-9 and Table 6.6-10

(extrapolated to the national level) are based on the assumptions presented in the section 6.5.2. For

every hour a pharmacist works, their interventions prevent $17.63 in costs to MBS and hospitals, and

for every 100 hours worked, their interventions prevent 1.3 days in hospital.

At a national level, a total of $349M in costs to MBS and Hospitals are prevented and around 262,000

days in hospital are prevented by the estimated 1.6M interventions that are performed each year.

75%

Aspirin

interventions

Other

interventions

Total

interventions

Per

pharmacist

hour

Rate - no adjustment for

recording 0 3.47 3.47 0.04

Rate - increased for

recording effect - 6.93 6.93 0.08

Value (attributed to

pharmacists)

Days in health state

loss prevented

State 3 0 26.03 26.03 0.30

State 2 0 108.74 108.74 1.27

State 1 0 94.80 94.80 1.11


Days in hospital 0 1.13 1.13 0.01

GP and specialist

consultations 0 6.39 6.39 0.07

Costs preventedCosts to health system

prevented 0 $ 1,508 $ 1,508 $ 17.63

Estimates 1 a) 1The value of current level of interventions in community

pharmacies,

Table 6.6-9: Estimates of current value of community pharmacy interventions; per 1000 prescriptions


Page 320 of 361

National

Aspirin

interventions

Other

interventions

Tota l

interventions

Estimated

interventions for total

national scripts,

includes estimated non-

recorded - 1.61 M 1.61 M


pharmacists)


loss prevented

State 3 0 6.03 M 6.03 M

State 2 0 25.19 M 25.19 M

State 1 0 21.96 M 21.96 M


Days in hospital 0 0.26 M 0.26 M

GP and specialist

consultations 0 1.48 M 1.48 M


prevented 0 $349.27 M $349.27 M

Estimates 1 b)

The value of current level of interventions in community pharmacies,

Table 6.6-10: Estimates of Current Value of Community Pharmacy Interventions; National

6.6.2 Estimate of improved rate of activity – aspirin popup

Without the aspirin pop-up, the aspirin intervention rarely occurred. In the data set used for the

economic analysis, this occurred only once. The aspirin pop-up is an example, then, of both an

opportunity and mechanism to increase the rate of intervention in community pharmacies. What would

the rate of intervention be if an aspirin pop-up were introduced nationally? As presented in Table

6.5-7, we could assume that the rate of interventions that would occur would be as that for either

observed days or unobserved days. In the case of the aspirin pop-up, we decided to use the rate on

observed days, for reasons explained in section 6.5.5.

In addition, we noted that there was an increase in the rate of other interventions on the days that the

aspirin pop-up was operating. Accordingly, the analysis assumes that there would be an increase in

the rate of other interventions as a result of introducing an aspirin pop-up. At the base case we

assumed that this rate was the rate of other interventions on unobserved days (as days would be

unobserved in practice) with the aspirin pop-up operating. This is in contrast to the assumption for the

aspirin interventions, but there would be no accompanying targets for the other interventions, or

specific education material. This is a recorded rate of 5.2 per 1000 prescriptions, which would be

adjusted by the assumed reporting rate to calculate the actual rate of interventions. The analysis

allows for this assumption to be varied either up (by using the rate on observed rather than

unobserved days) or down (by using the rate on never observed days). (See Table 6.5-7)


Page 321 of 361

The additional benefits per 1000 prescriptions are significant, partly because the aspirin intervention

has significant average value, up to three times that of the average for other interventions (see Table

6.6-12; per 1000 prescriptions and Table 6.6-13; national extrapolation). There is a possibility that the

gains from this could reduce over time as an initial benefit for existing cases is experienced, but over

time the benefit is from incident cases only. The total additional costs prevented are around $2,900 per

1000 prescriptions and $669M nationally per annum. This saving is the result of both the increased

aspirin interventions (around 57%) and the increased rate of other interventions (around 43%).

75%

Aspirin

interventions

Other

interventions

Total

interventions

Per

pharmacist

hour


recording 4.31 5.19 9.50 0.11


prevailing recording

effect 4.79 10.38 15.17 0.18


pharmacists)


loss prevented

State 3 33.9 39.0 73 0.85

State 2 59.6 162.7 222 2.60

State 1 0.8 141.8 143 1.67


Days in hospital 1.74 1.70 3 0.04

GP and specialist

consultations 1.34 10 11 0.13

Costs prevented Costs to health system

prevented $ 631 $ 2,256 $ 2,887 33.75

Effect of pop up

Additional interventions 4.79 3.44 8.24 0.10

Additional days in

hospital prevented 1.74 0.56 2.30 0.03 Additional costs

prevented $ 631 $ 748 $ 1,379 16.12

Estimate 2 a) Improved rate of intervention, per 1000 scripts

increase due to an aspirin pop-up

Table 6.6-11: Estimate of improved rate of intervention (aspirin pop-up), per 1000 prescriptions


Page 322 of 361

National

Aspirin

interventions

Other

interventions

Tota l

interventions

Number of

interventions-

estimated actual 1.11 M 2.40 M 3.51 M


pharmacists)


loss prevented

State 3 7.86 M 9.02 M 16.89 M

State 2 13.81 M 37.69 M 51.50 M

State 1 0.20 M 32.86 M 33.05 M


Days in hospital 0.40 M 0.39 M 0.80 M

GP and specialist

consultations 0.31 M 2.22 M 2.53 M

Costs prevented Costs to health system

prevented $146.15 M $522.63 M $668.79 M

National effect of pop

up

Additional interventions 1.11 M 0.80 M 1.91 M

Additional days in

hospital prevented 0.40 M 0.13 M 0.53 MAdditional costs

prevented $146.15 M $173.36 M $319.51 M

Estimate 2 b) improved rate of intervention, national

increase due to an aspirin pop-up

Table 6.6-12: Estimate of improved rate of intervention (aspirin pop-up), with national extrapolation


Page 323 of 361

6.6.3 Estimate of improvements resulting from reduced pharmacist workload

One determinant of the rate of intervention was the pharmacist’s workload. Table 6.4-14 presents the

reported rate of intervention by quintile of pharmacist activity. Overall, pharmacists performed

interventions at a rate of 13 per 1000 prescriptions in the periods of lowest workload, and at a rate of

3.4 in the periods with the highest workload.

The objective of the analysis was to determine the rate of intervention that would occur per 1000

prescriptions and nationally, if pharmacists intervened at a rate consistent with the lower workloads,

and also to estimate the additional pharmacist hours required to achieve these reductions in workload.

To perform the analysis, we assumed that if the pharmacist workload was reduced to be consistent

with that of the lower two quintiles, then the rate of intervention at the lower levels of workload

observed in our study would then apply to all hours of activity and all prescriptions. We then used the

number of hours per 1000 prescriptions that the pharmacist worked at the lower quintiles, and applied

this rate to the volume of activity in the higher quintiles to estimate the additional pharmacist hours

required to achieve a lower work load.

We also used our assumption of the cost per hour of a pharmacist and applied this to our estimate of

additional hours to consider the costs of additional pharmacists.

We have not adjusted the estimate based on the average of the lowest two quintiles for the effect of

having a higher proportion of patients with more than one medication in this group, (see Table 6.4-11,

Table 6.4-12 and Table 6.4-13). The higher proportion of more complex patients (who have a higher

rate of intervention) in the low workload quintiles made a small contribution to the higher rate of

intervention in this quintile. Table 6.6-13 presents the analysis on the basis of 1000 prescriptions. An

additional 28 hours per 1000 prescriptions compared to existing practice would be required in order to

achieve these higher rates of intervention. At around $1,300, these additional hours would have a

greater financial value than the additional expenditure prevented ($1,050), but there would also be

substantial gains in health status and days in hospitals. This suggests that it is likely that the

additional financial cost of the additional hours would be less than the additional benefit.

Table 6.6-14 extrapolates these results to a national level, this increase would mean around 3,500

additional FTE pharmacists. It is unlikely that this number of pharmacists is available in Australia, but

other ways to reduce the impact of a high number of prescriptions per pharmacist hour could be

considered, including greater usage of dispensing technicians. The gains to society of reducing the

pharmacist workload could be significant; however, the feasibility of achieving this through increased

numbers of pharmacists alone would be low.


Page 324 of 361

75%

Aspirin

interventions

Other

interventions

Total

interventions

Per pharm

hour

Rate - not adjusted for

recording effect - 5.56 5.56 0.05



effect - 11.12 11.12 0.10

Value


loss prevented

State 3 - 41.7 42 0.36

State 2 - 174.4 174 1.52

State 1 - 152.0 152 1.33


Days in hospital - 1.82 2 0.02

GP and specialist

consultations - 10.25 10 0.09


prevented $ - $ 2,418 $ 2,418 21.12

Sample effect of

additional pharmacist

hours

Additional interventions - 4.19 4.19 0.04

Additional days in

hospital prevented - 0.68 0.68 0.01 Additional costs

prevented $ - $ 910 $ 910 7.95

Estimate 3 a) of improved rate of intervention, per 1000 scripts

increase due to more pharmacist hours

Table 6.6-13: Estimate of improved rate of intervention (more pharmacist hours), per 1000 prescriptions


Page 325 of 361

National

Aspirin

interventions

Other

interventions

Tota l

interventions

Number of

interventions-

estimated actual - 2.58 M 2.58 M

Value attributed to

pharmacists


loss prevented

State 3 - 9.67 M 9.67 M

State 2 - 40.39 M 40.39 M

State 1 - 35.21 M 35.21 M


Days in hospital - 0.42 M 0.42 M

GP and specialist

consultations - 2.38 M 2.38 M


prevented $ - $560.09 M $560.09 M

National effect of

additional pharmacist

hours

Additional interventions - 0.97 M 0.97 M

Additional days in

hospital prevented - 0.16 M 0.16 MAdditional costs

prevented $0.00 M $210.82 M $210.82 M

Estimate 3 b) of improved rate of intervention, nationally

increase due to more pharmacist hours

Table 6.6-14: Estimate of improved rate of intervention (more pharmacist hours), nationally

6.6.4 Estimate of improvement resulting from maximum possible rate of interventions

If all pharmacies were identifying and performing interventions at an optimal level, then we estimate a

total of $955M in financial savings to MBS and hospitals and 1.01B days in hospital could be

prevented (see Table 6.6-16). This represents an addition 749,000 days in hospital prevented and

$606M in savings compared to the existing level of activity. This estimate assumes the optimal level

comprises:

• aspirin interventions as the only additional type of intervention, and

• an increased rate of the existing range intervention.

There are likely to be other types of interventions not currently being performed, so this estimate of

optimal rate of interventions could be an underestimate of the actual opportunity to improve rate of

interventions. The analysis also assumes that the average value of additional interventions is the

same as the average value of existing interventions. This would appear to be a reasonable

assumption for two reasons:


Page 326 of 361

• the rate at which pharmacists recorded their interventions appeared to be similar across

different types of pharmacist-assessed severity – so the reported interventions appear

representative of all interventions

• the additional interventions such as those that result from the aspirin pop-up actually have a

greater value than the average value of interventions performed currently.

Table 6.6-15 presents the results at the level of per 1000 prescriptions and Table 6.6-16 at a national

level. The total time spent on interventions would be 1.9 hours per 1000 prescriptions, which

represents $87 in terms of the financial value of the pharmacists’ time and 1 hour of interventions per

every 44 hours of pharmacists’ time.

75%

Aspirin

interventions

Other

interventions

Total

interventions

Per

pharmacist

hour


recording 4.31 14.45 18.77 0.219



effect 4.79 16.06 20.85 0.244

Value


loss prevented

State 3 33.9 60.3 94 1.101

State 2 59.6 251.8 311 3.641

State 1 0.8 219.5 220 2.576


Days in hospital 1.74 2.62 4 0.051

GP and specialist

consultations 1.34 15 16 0.189

CostsCosts to health system

prevented $ 631 $ 3,492 $ 4,123 48.192

Effect of maximum

intervention rates per

1000 scripts

Additional interventions 4.79 9.12 13.92 0.163

Additional days in

hospital prevented 1.74 1.49 3.23 0.038 Additional costs

prevented $ 631 $ 1,984 $ 2,615 $ 30.57

Estimate 4 a) maximum rate of intervention, per 1000 scripts

increase due to optimal detection and performance by pharmacists , no additional hours

Table 6.6-15: Estimate of improved rate of intervention (using maximum interventions), per 1000 prescriptions


Page 327 of 361

National

Aspirin

interventions

Other

interventions

Total

interventions

Number of

interventions- 1.11 M 3.72 M 4.83 M

Value attributed to

pharmacist


loss prevented

State 3 7.86 M 13.96 M 21.83 M

State 2 13.81 M 58.33 M 72.14 M

State 1 0.20 M 50.85 M 51.05 M


Days in hospital 0.40 M 0.61 M 1.01 M

GP and specialist

consultations 0.31 M 3.43 M 3.74 M

Costs

Costs to health system

prevented $146.15 M $808.81 M $954.96 M

National effect

maximum

intervention rates

Additional interventions 1.11 M 2.11 M 3.22 M

Additional days in

hospital prevented 0.40 M 0.35 M 0.75 MAdditional costs

prevented $146.15 M $459.54 M $605.69 M

Estimate 4 b) maximum rate of intervention, national

increase due to optimal detection and performance by pharmacists , no

additional hours

Table 6.6-16: Estimate of improved rate of intervention (using maximum interventions), national


Page 328 of 361

7 Results and Discussion Part 3: Barriers and Facilitators to Performing and Recording Clinical Interventions

Each pharmacist involved in the Melbourne PROMISe trial was encouraged to record each of their

clinical interventions. The records of clinical interventions performed in community pharmacy provide a

unique repository of information.

Information about barriers and facilitators to recording and performing clinical interventions was

gathered from participants via an independent body - DeBoos Associates. Information was gathered

by DeBoos Associates via three focus group sessions and ten individual interviews. Other issues

discussed in these sessions included remuneration, general uptake of the system and software

concerns; these issues will be discussed in sections 7.4 and 8.1. Information was also gathered via a

questionnaire distributed to each pharmacist involved in the trial (Post Study Questionnaire)

Finally, a telephone survey gathered opinions from both pharmacists involved in the trial and a

randomly selected representation of community pharmacists from across Australia on the barriers and

facilitators to clinical interventions.

7.1 Barriers to performing clinical interventions

As regularly recording clinical interventions was generally a new procedure for the pharmacists

involved in the PROMISe project, a number of barriers were found to adopting this new practice. The

pharmacists involved in the feedback were able to formulate solutions to some of these barriers.

7.1.1 Definition and Identification of Clinical Interventions

The PROMISe project provided the participants with a fairly broad definition for an intervention:

Any professional activity (outside of the basic dispensing and counselling procedures) directed

towards improving health outcomes or the quality use of medicines, or the provision of health-related

information.

However, from the feedback received it seems that this definition was not fully utilised by participants.

There were in fact a range of definitions found, from drug interactions and changes in dose to any

recommendation made to patients about their medications. The Melbourne trial of the PROMISe

project aimed to gather as much information as possible on the clinical activities of community

pharmacists (which is why a broad definition was adopted), with a focus on prescription medications.

However, the feedback provided indicates that more information is required by the participants to

adequately recognise clinical interventions. Where pharmacists had varying understanding of the

definition of clinical interventions it is not surprising that these pharmacists would also have had

difficultly identifying when they were performing a clinical intervention. The provision of on-site training

is a possible way to overcome these problems with definition and perception of clinical interventions.


Page 329 of 361

7.1.2 Business Culture and the Individuals Who Operate in These Systems

The culture or business style of the pharmacy was seen as a determinant for the rate of clinical

services performed.

Pharmacists were asked to indicate what proportion of their time was involved in performing clinical

services (see Appendix 16) as opposed to administrative tasks. It can be seen in Figure 7.1-1 that

most pharmacists involved in PROMISe spent less than 20% of their work day providing clinical

services. When reviewing this information it is also important to keep in mind that the pharmacies

which enrolled in the PROMISe project were informed that the project involved recording clinical

interventions - so by means of self selection these pharmacies would probably have had some interest

in providing clinical services.

44

60

10 3 20

10

20

30

40

50

60

70

<5% 5-20% 20-50% >50% no response

Figure 7.1-1: The Estimated Proportion Of Time At Work Involved In “Clinical Services”

The business culture of a pharmacy is generally determined by the owner/s of that pharmacy. Some

of the pharmacies involved in the PROMISe project were observed to be focused on business

aspects, such as meeting dispensing time expectations. Others gave attention to professional services

including the provision of advice and counselling to patients. Whilst each pharmacy has the potential

to explore both styles, in general it was found that one style would dominate. The Melbourne

PROMISe trial aimed to encourage both styles of pharmacy to participate by including payment for

clinical interventions and also by providing an educational drug alert.

Whilst business culture has an influence on the employed pharmacists, there was also a degree of

variability which was attributed to individual approaches to performing and recording clinical services.

Some pharmacists were observed to choose to remain in the dispensary and perform the role of

medication suppliers; these pharmacists generally did not employ dispensary technicians. Another


Page 330 of 361

individual factor was the pharmacist’s clinical knowledge and confidence to use their clinical

knowledge. A poor relationship with their local prescribers was also seen as a contributing factor to the

reluctance to perform clinical interventions. As with the adoption of any new system, there was also

the influence of inertia to change.

7.1.3 Time and workload

A lack of time was cited as the most significant barrier to both recording and performing clinical

interventions in both the feedback sessions facilitated by DeBoos Associates and in the post-study

questionnaire. In Figure 7.1-2 the lack of time can be seen to be considered as a barrier by the

greatest proportion of pharmacists (2.5 times higher than the next barrier considered).

The perception of lack of time can also be linked to the pharmacy’s workload, its staffing levels and the

mix of staff employed. Where a pharmacist is working with very high workloads and poor staffing

levels it was observed that the interventions of a serious nature were performed but other activities

were reduced.

39.6%

14.9% 13.9%11.9%

5.0%3.0% 2.0% 1.0%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

45.0%

lack of time workflow

restrictions

forgetfulness software

concerns

no barrier

found

other lack of

motivation

lack of

clinical

knowledge

Figure 7.1-2: The Actual Barriers To Recording Interventions As Seen By Participants In The PROMISe Trial

The general observation was that there was not enough time for clinical interventions. When the

pharmacists were asked to indicate the average amount of time they took to both perform and record

their interventions, over half of the respondents indicated that it took them between 2 and 5 minutes.


Page 331 of 361

Therefore, the interventions for the most part did not actually take a long time. It is more likely that

prioritisation and importance of these activities is considered by many pharmacists as quite low.

Therefore, any time taken to complete these tasks is seen as an interruption to normal daily practices.

This perception is also an indicator that broad changes to community pharmacy practice are required

to increase the rate of clinical service provision.

2

60

46

9

1

0

10

20

30

40

50

60

70

<1 min 2-5 min 6-10 min 11-30 min >30 min

Figure 7.1-3: The Average Amount Of Time To Perform And Record Clinical Services

7.1.4 Clinical knowledge and continuing education

During the Melbourne PROMISe trial an educational drug alert was activated in 31 pharmacies. The

pharmacists from these pharmacies found the alert to be beneficial and provided considerable positive

feedback during both the focus sessions and the individual interviews. The pharmacists saw it as a

good way to incorporate continuing education into the workplace.

It was found through the feedback provided that those pharmacists who indicated their clinical

knowledge was not fully up to date did not feel confident dealing with some clinical interventions. In

some situations these pharmacists would refer the situation onto another pharmacist in their practice.

Hence, having a good mixture of pharmacists is important to optimise the clinical services provided.

Practical continuing education which highlights proactive interventions was highlighted by the

participants as a possible way of increasing intervention rate.

Although many pharmacists perceived lack of clinical knowledge as a barrier before being involved in

the trial (14.9% or respondents), it can be seen in Figure 7.1-4 that once the pharmacists had

participated in the trial their concerns about clinical knowledge reduced (from 14.9% to 4.2% of

respondents).


Page 332 of 361

The high initial perception may have been increased by concern about being involved in a 'research

project', whereas the aim of the project was to encourage the pharmacists to record their interventions

and not to assess their interventions. It was also stressed in the sessions prior to commencing the trial

that many useful interventions do not require an extensive clinical knowledge e.g could be a simple

matter of checking whether a patient, in error, is taking two drugs from the same therapeutic group (or

a generic and innovator brand concomitantly).

3.6%

25.8%

14.9%

10.2%

32.0%

13.5%

3.2%

21.6%

4.2%

25.3%

34.2%

11.6%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

lack of

motivation

workflow

restrictions

lack of clinical

knowledge

forgetfulness lack of time software

concerns

Perceivedbarrier torecording

Actualbarriers torecording

Figure 7.1-4: Barriers to Recording Clinical Services; Before And After Participating in the PROMISe Trial


Page 333 of 361

The pharmacists were able to identify that additional educational opportunities would increase their

ability to recognise drug-related problems (see Figure 7.1-5).

25

61

10

17

3 21

strongly agree

agree

neutral

disagree

strongly disagree

unsure

no response

Figure 7.1-5: I Require An Update Of Clinical Knowledge To Optimise Identification Of Drug Related Problems

7.1.5 Other barriers to performing clinical interventions

A number of other barriers were proposed to have an influence on pharmacists performing clinical

interventions. These included

• Language barriers

• Exposure to litigation

• Poor relationships with other health professionals, including GPs

• Inertia to change in practice

• Inadequate recorded patient information on which to base decisions

These barriers were also explored in the telephone survey of Australian community pharmacists (see

section 7.3)

7.2 Facilitators to performing clinical interventions

The pharmacists who participated in the Melbourne PROMISe trial found clinical interventions to be

both professionally rewarding and to have an important role in pharmacy practice.


Page 334 of 361

7.2.1 Professional satisfaction

Personal and professional satisfaction were key motivators for continuing to perform clinical services.

All respondents to the feedback facilitated by DeBoos Associates indicated that they would like to be

able to carry out more interventions and felt they were hindered by time constraints or poor clinical

knowledge. When the pharmacists were asked if they thought recording clinical services helped to

demonstrate the ability of pharmacists to improve medication therapy, the following responses were

seen (see Figure 7.2-1)

83

32

1201

strongly agree

agree

neutral

disagree

strongly disagree

unsure

Figure 7.2-1: Does Recording Clinical Services Demonstrate The Ability Of Pharmacists To Improve Medication Therapy?

7.2.2 Recognition for providing clinical services

It is generally accepted that the general public have a poor understanding of the role of pharmacists

both within hospital and community practice. The trial participants proposed that conducting clinical

interventions could be a way to increase the awareness of pharmacists' activities. Also by

pharmacists carrying out clinical interventions, patients would be able to gain a better understanding of

their individual medical conditions and medications. The consequences of this improved knowledge

could be extensive and would also increase the understanding of the role of pharmacists. There is the

potential that through this recognition that the pharmacists are able to increase customer loyalty and

potentially the profitability of their business.

Pharmacists participating in the PROMISe trial also identified that being able to submit their

interventions provided greater motivation for recording. The concept that someone was receiving and

analysing the information was seen as a positive benefit to the trial.


Page 335 of 361

When the pharmacists undertook interventions related to the educational drug alert, they generally

received good feedback from their patients. This, in turn, was a driver for performing more

interventions.

7.2.3 Information Continuity

Recording clinical interventions allowed information to be passed on between pharmacists working

within the same pharmacy. As an organised recording system was developed, the need to record

extensive notes for other pharmacists was reduced. Some of the pharmacists involved in the trial had

been using information entered into the previous system to report to the management of the nursing

home they serviced. The reporting of interventions was seen as an aspect of the system which could

be improved, so that various groups could receive reports relating to the pharmacists’ interventions.

By maintaining a relatively complete record of interventions the pharmacists were able to accurately

follow-up individual patients. This also contributed to increased professional satisfaction.

7.2.4 Continuing education

The provision of a unique educational drug alert was well accepted by participants who were exposed

to it during the trial. The pharmacists who submitted feedback on this section of the trial indicated that

they would be interested in other educational drug alerts. These alerts would need to be activated and

de-activated in a timely manner and be relevant to the current practice of the pharmacy. They were

identified as a subtle but powerful educational tool, allowing on-the-job training.

As well as “on the job” continuing education, the pharmacists involved in the feedback saw that

specific continuing education could increase their rate of clinical interventions. As the continuing

education could be linked to their routine practice the pharmacists felt this could also improve the

uptake of the continuing education.

7.2.5 Work environment

Each pharmacist involved in the PROMISe trial identified that when they were busy it was only the

most serious interventions or those which prevented the prescription from being dispensed which

occurred. Therefore, ensuring that pharmacists are working with optimal conditions could potentially

increase the intervention rate. The conditions would vary between sites and pharmacists. Ensuring

adequate staffing rates and mixture of staff would be an obvious improvement. Also important would

be ensuring that the work environment allowed for a sufficient counselling area and that the

pharmacist could spend uninterrupted time with their patients when required.

However, although it would be possible to create the 'ideal' conditions there are still individual

characteristics which could outweigh these changes and hence there may be no change in

intervention frequency.


Page 336 of 361

7.3 National Survey of Clinical Intervention Documentation

A telephone survey was conducted by I-view.25

A questionnaire was developed with DeBoos

Associates in consultation with the Project Team. This explored barriers and facilitators to clinical

interventions in community pharmacy.

Of the voluntary respondents 83 participants had taken part in the PROMISe trial and 417 were

sampled from across Australia, providing a total sample of 500 community pharmacists.

142

106

72

35 40

114 7

83

0

20

40

60

80

100

120

140

160

180

200

NSW Vic Qld SA WA Tas NT ACT

PROMISe Respondents

Australian Respondents

Figure 7.3-1: Representation Of Respondents To Telephone Survey

Each participant in the telephone survey was either the pharmacist in charge or a participant in the

trial. They were given the PROMISe definition of clinical intervention and responded accordingly. The

time taken for the interview was between 10 and 15 minutes. When responding to a statement the

participant was given a scale of 0 -10, where 0 was strongly disagree and 10 was strongly agree.

Table 7.3-1 displays the responses to the statements about potential barriers to performing and

recording interventions. The results are displayed as the mean and standard deviation of the

responses from both the PROMISe participants and the Australian sample.

25 I-view is a data collection and dissemination agency with 20 years experience in market and social

research data processing and other allied services, I-view, Hawthorn, Victoria


Page 337 of 361

Barriers PROMISe Non-

PROMISe

Some of the pharmacists I work with or know don’t seem interested in performing clinical interventions.

4.28 (2.7) 4.46 (2.9)

Having an efficient and profitable business is more important than worrying about clinical interventions.

2.72 (1.9) 3.13 (2.3)

Some of my patients speak a different language and I don’t have a way to communicate with them well enough to perform clinical interventions.

4.89 (2.7) 4.33 (3.0)

Some pharmacists really don’t like change so they are not interested in doing anything more than run the business and dispense what is needed.

5.40 (2.3) 5.27 (2.6)

When it is busy I don’t have time to do anything more than dispense and intervene in the most important or potentially serious clinical situations.

5.60 (2.9) 5.77 (2.8)

I don’t like performing too many clinical interventions as it exposes me to being sued.

2.04 (1.9) 2.37 (2.1)

My other activities in the pharmacy prevent me from spending time with customers talking about their medications

4.01 (2.7) 4.30 (2.6)

I feel my clinical knowledge limits the number of clinical interventions I perform.

4.18 (2.1) 4.10 (2.4)

It is impossible to perform clinical interventions when you don’t have a complete patient history to consult.

5.80 (2.5) 5.79 (2.7)

Doctors are always very approachable when I contact them.? 5.94 (1.9) 6.15 (2.2)

There is not time to record all clinical interventions. 5.77 (2.5) 5.72 (2.8)

Customers have a good understanding of what pharmacists are trained to do.

5.66 (2.3) 5.68 (2.3)

Most customers believe pharmacists stick labels on containers and not much more.

5.63 (2.2) 5.11 (2.4)

When dispensing repeat preprescriptionions I am less vigilant to possible interventions.

5.39 (2.1) 5.40 (2.3)

Table 7.3-1: The Responses To The Statements Used In The I-View Telephone Survey

One of the barriers identified in the focus group sessions and individual interviews was the impact of

work environment and business culture. These two factors could act as both barriers and facilitators to

clinical interventions. From the sample gathered during the telephone survey, the majority did not rate

their business over clinical role. There was representation from those pharmacies where the business

structure is a significant driver to their pharmacy (see Figure 7.3-2).


Page 338 of 361

0 2 4 6 8 10

Having an efficient and profitable business is more important than worrying about

clinical interventions.

0

20

40

60

80

100F

req

ue

nc

y

Mean = 3.064Std. Dev. = 2.25967N = 500

Figure 7.3-2: Responses Of All Pharmacist About Business Vs. Clinical Performance

From the results seen in the telephone survey it appears that language can be a barrier to clinical

interventions, but only in certain circumstances. During the focus session one of the pharmacists

highlighted that language could be a significant barrier; however, when the intervention was of a

serious nature, then means were found to ensure adequate communication (see Figure 7.3-3)


Page 339 of 361

0 5 10

Language as a barrier to conducting clinical interventions

0

10

20

30

40

50

60

70F

req

uen

cy

Mean = 4.42Std. Dev. = 2.96127N = 500

Figure 7.3-3: Responses To Language Difficulties As A Barrier To Clinical Interventions

The potential for litigation may be a barrier for some pharmacists when considering offering advice and

recommendations to their patients. The respondents concurred that this, however, is not generally a

barrier to clinical interventions (see Figure 7.3-4).


Page 340 of 361

0 5 10

I dont like performing too many clinical interventions as it exposes me to being

sued.

0

20

40

60

80

100

120F

req

ue

nc

y

Mean = 2.318Std. Dev. = 2.10088N = 500

Figure 7.3-4: Responses To Potential For Litigation As A Barrier To Clinical Interventions

The requirement of good communication skills to facilitate clinical knowledge includes communication

between health professionals. When the respondents were asked about their relationships and

personal experiences with GPs, most indicated that they found them to be approachable (see Figure

7.3-5).


Page 341 of 361

0 5 10

Doctors are always very approachable when I contact them.

0

20

40

60

80

100

120F

req

uen

cy

Mean = 6.114Std. Dev. = 2.16203N = 500

Figure 7.3-5: Responses To Professional Relationships As A Barrier To Clinical Interventions

The facilitators are displayed in Table 7.3-2. As with the previous table, the mean response is

displayed and the standard deviation for both PROMISe participants and the Australian sample of

community pharmacists.

The telephone survey canvassed a large number of community pharmacists (n=500) about their

opinions on clinical interventions. In general, the respondents were very positive about their responses

when considering the facilitators to undertaking clinical interventions, compared with the barriers.


Page 342 of 361

Facilitators PROMISe Non-

PROMISe It is important for all pharmacists to perform clinical interventions.

9.35 (0.9) 8.99 (1.2)

The more clinical interventions I can perform, means better health outcomes for my customers.

9.0 (1.1) 8.7 (1.5)

Looking at the prescription workload in this business, I could justify employing another person to dispense prescriptions.

4.04 (2.8) 3.71 (2.9)

When I am dispensing I try to interact with patients as much as possible or counsel them in the proper use of their medication.

9.00 (1.1) 8.67 (1.3)

If I ‘handed out’ more prescriptions to patients I could identify more clinical interventions to perform.

6.54 (2.7) 6.07 (2.1)

When I perform a clinical intervention I get a great deal of personal satisfaction.

8.64 (1.4) 8.26 (1.8)

Performing more interventions builds relationships and is good for business

8.76 (1.2) 8.71 (1.5)

If patients are better informed they have a greater chance of their disease being better controlled on their medication.

8.95 (1.3) 8.71 (1.5)

It is important to me that my customers are well informed about their condition and medication.

9.04 (1.2) 8.75 (1.4)

Recording clinical interventions is important to inform other pharmacists what has taken place.

8.99 (1.2) 8.24 (1.8)

Recording clinical interventions is important as it reminds me what I have done.

8.60 (1.4) 8.29 (1.7)

If I was to receive additional payment for performing clinical interventions then I would endeavour to do more.

6.87 (2.6) 7.00 (2.9)

Payment for conducting clinical interventions is not required as it is a service that pharmacy should provide free of charge.

3.18 (2.2) 4.88 (3.0)

If patients know that I record my clinical recommendations, it is a demonstration of the value of pharmacists to the community.

8.25 (1.5) 7.76 (2.1)

If pharmacists gained CPE points from performing interventions then they would record more interventions.

6.66 (2.5) 6.67 (2.6)

Performing clinical interventions is a way of differentiating my business from others.

8.24 (1.6) 7.90 (1.9)

Training on clinical interventions is not required as I know enough to do interventions.

4.29 (2.0) 4.39 (2.4)

It would be good if the dispensing software was able to remind me to perform particular clinical interventions (for example , suggesting patients with type 2 diabetes should be taking aspirin) *

7.01 (2.8) 7.30 (2.5)

* PROMISe participants who had been involved in the education arm of the study n= 42

(51% or respondents) therefore had had exposure to a reminder system

Table 7.3-2: Facilitators To Interventions As Assessed By The I-View Telephone Survey


Page 343 of 361

As can be seen in the tables above, the respondents held fairly similar opinions about clinical

interventions when compared between project participants and the randomly selected pharmacists.

When asked about the importance of clinical interventions the majority of respondents indicated that it

was an important role to fulfil (see Figure 7.3-6).

0 2 4 6 8 10

It is important for all pharmacists to perform clinical interventions.

0

50

100

150

200

250

300

Fre

qu

en

cy

Mean = 9.05Std. Dev. = 1.20765N = 500

Figure 7.3-6: Response From All Pharmacists About The Importance Of Clinical Interventions

The respondents also agreed that it was a professional and personally rewarding role within their

practice (see Figure 7.3-7)


Page 344 of 361

0 5 10

When I perform a clinical intervention I get a great deal of personal satisfaction.

0

50

100

150

Fre

qu

en

cy

Mean = 8.326Std. Dev. = 1.72327N = 500

Figure 7.3-7: Responses Of All Pharmacists To The Personal Satisfaction Of Conducting Clinical Interventions

The receptiveness for technology changes to provide continuing education was high (see Figure

7.3-8).

This initial good acceptance of this means of providing continuing education is encouraging. Of the

PROMISe participants who had been exposed to this type of system (n= 42, 51% or respondents from

the PROMISe trial) there was also agreement that this type of system was workable.


Page 345 of 361

0 5 10

It would be good if the dispensing software was able to remind me to perform particular

clinical interventions

0

20

40

60

80

100

120

140F

req

ue

nc

y

Mean = 7.75Std. Dev. = 1.95426N = 500

Figure 7.3-8: Responses From All Pharmacists About The Use Of Dispensing Software In Continuing Education

The telephone survey responses supported the facilitators to clinical interventions assessed in the

focus groups and the post-study questionnaire. In Figure 7.3-9 facilitators have been displayed. In

general, it is the influence of each of these facilitators collectively which has an impact on clinical

interventions. For example, where continuing education is provided without the appropriate work

environment it would be difficult to put the knowledge into practice.

Remuneration and its impact on undertaking and recording clinical interventions is discussed in

section 5.8.9.


Page 346 of 361

Figure 7.3-9: Facilitators to Undertaking And Recording Clinical Interventions In Community Pharmacy

A similar diagram has been developed with the barriers to recording and performing interventions (see

Figure 7.3-9). The influence of the different barriers, as with the facilitators, often are linked, for

example professional confidence is linked to clinical knowledge.

BarriersBarriers

Clinical Clinical

KnowledgeKnowledge

Professional Professional

and Personal and Personal

ConfidenceConfidence

Work Work

environment environment

and business and business

cultureculture

Workload Workload

Available Available

Time Time

Figure 7.3-10: Barriers To Undertaking And Recoding Clinical Interventions In Community Pharmacy


Page 347 of 361

7.4 Opinions Regarding Remuneration for Interventions

The pharmacists who participated in the PROMISe project and a sample of pharmacists selected at

random across Australia were canvassed regarding their opinions of different remuneration models for

recording and performing interventions. This involved determining their opinions of the model and an

estimate of the rate of remuneration within that particular model. The potential economic implications

for a national implementation each of the models explored have not been considered in this report.

7.4.1 Preferred Payment Models

Participants in the Melbourne PROMISe trial were exposed to payment and non-payment periods

during the trial; 109 out of 118 participants who returned the post-study questionnaire considered

payment for recording clinical services to be important. (see Figure 7.4-1)

109

5 4

0

20

40

60

80

100

120

Yes No Unsure

Figure 7.4-1: Should Pharmacists Be Remunerated For Recording Their Clinical Services?


Page 348 of 361

Others who considered payment as less important highlighted continuing education, feedback and

better patient care as drivers for recording interventions. Different models of remuneration were

discussed. These included

• Set payment per intervention

• Payment for interventions related to educational alert when a certain target had been reached

• A matrix which included seriousness of interventions and the time taken for investigation

(Figure 7.4-2)

Severe

Moderate

Mild

Nil

LongModerate Short TimeSeverity

Figure 7.4-2: Matrix Of Remuneration

7.4.2 Remuneration Models Assessed in the Post Study Questionnaire

Five models were proposed and included into the post-study questionnaire (entire questionnaire can

be viewed in Appendix 16). Each pharmacist then expressed their level of agreement with that model

and suggested an appropriate amount of remuneration. The rate of remuneration will be discussed in

section 7.4.4.

Models of remuneration

1) A payment to each pharmacy for clinical services documented

2) A payment to each pharmacist for clinical services documented

3) A payment to each pharmacy for those interventions which are of high or moderate

significance

4) A payment to each pharmacist for those interventions which are of high or moderate

significance

5) An increase in the dispensing fee for prescriptions for drugs with a high frequency of clinical

services (eg warfarin)

The responses to these different models can be seen in Table 7.4-1


Page 349 of 361

In general both groups supported all of the models and their level of support was fairly high. For each

model proposed those that agreed with the model were approximately 60% of respondents whereas

disagreement for any of the proposed models did not exceed 30% of respondents.

Model

39 34.2% 39 34.2% 14 12.3% 11 9.6%

40 35.1% 31 27.2% 18 15.8% 14 12.3%

51 44.7% 30 26.3% 11 9.6% 14 12.3%

46 40.4% 32 28.1% 13 11.4% 13 11.4%

44 38.6% 34 29.8% 12 10.5% 11 9.6%

Strongly

Disagree

68.4% 21.9%

11 9.6%

Strongly

AgreeAgree Neutral Disagree

22.8%

68.4% 20.2%

62.3% 28.1%

71.1% 21.9%

1 :Pharmacy Payment for

Each Intervention

2: Pharmacist Payment for

Each Intervention

3: Pharmacy Payment for

Selected Interventions

4: Pharmacist Payment for

Selected Interventions

5: Overall Increase In

Dispensing Fee

11 9.6%

8 7.0%

10 8.8%

13 11.4%

68.4%

Table 7.4-1: PROMISe Pharmacists’ Opinions of Different Remuneration Models

As expected, owner pharmacists generally supported models of remuneration based around payment

to the pharmacy (models 1 and 3) rather than models based on payments to the pharmacist (Models 2

and 4) (see Figure 7.4-3).

Figure 7.4-3: Preferred Remuneration Models: Owner vs Employee Pharmacists

80.0%

49.0%51.6%

75.2%

14.4%

37.0% 35.8%

18.0%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

Owners opinions

on payment to

pharmacy

Employee opinions

on payment to

pharmacy

Owner opinions on

payment to

pharmacist

Employee opinions

on payment to

pharmacist

Agree

Neutral

Disagree


Page 350 of 361

7.4.3 Remuneration Models Explored During the I-view Telephone Survey

After the feedback in relation to remuneration was gathered from the focus sessions, individual

interviews and the post study questionnaire, three models of remuneration were developed. The

acceptance of these models was assessed by the 500 community pharmacists contacted during the I-

view telephone interviews.

Model A: A set payment each month to all pharmacies for conducting interventions regardless

of the number of interventions

Model B: Payment to the pharmacy for each intervention recorded

Model C: A set payment when a pharmacy attained a certain rate of intervention (this could be

linked to a specific educational alert)

The selection for each of these models can be seen in Figure 7.4-4, with model B being the most

widely accepted model for remuneration.

0.00%

10.00%

20.00%

30.00%

40.00%

50.00%

60.00%

Model A Model B Model C None of the

above, other

model

No payment

required

PROMISe

Other

Figure 7.4-4: Remuneration Models Explored During the Telephone Survey


Page 351 of 361

7.4.4 Preferred Rates of Remuneration

When responding to the various remuneration models (within the post-study questionnaire) the

pharmacist was given the opportunity to nominate an amount of remuneration for each model. For

analysis of the remuneration value those respondents which did not nominate a value were not

considered in the evaluation.

7.4.4.1 Preferred Rate if a Payment to Each Pharmacy for Each Intervention is Made

The average preferred amount for this model was $12 (range $1 to $50) per clinical service

documented (see Figure 7.4-5). Over 50% of the pharmacists who responded indicated that a

payment to the pharmacy of less than $15 per intervention would be a suitable payment.

0.00 10.00 20.00 30.00 40.00 50.00

A payment to each pharmacy for clinical services documented

0

5

10

15

20

25

30

Fre

qu

en

cy

Mean = 12.0139Std. Dev. = 8.17156N = 72

Figure 7.4-5: Preferred Rate of Remuneration for Per Intervention Pharmacy Payment

7.4.4.2 Preferred Rate if a Payment to Each Pharmacist For Each Intervention is Made

The average preferred rate of payment for this model was $11.67 (range $1 to $70) for each

clinical service documented (see Figure 7.4-6). This result is very similar to that for the first model,

and implies that pharmacists who responded to the survey felt that remuneration of this order was

suitable for a per intervention payment.


Page 352 of 361

0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00

A payment to each pharmacist for clinical services documented

0

5

10

15

20

25

30

Fre

qu

en

cy

Mean = 11.6714Std. Dev. = 9.14888N = 70

Figure 7.4-6: Preferred Rate of Remuneration for Per Intervention Pharmacist Payment

7.4.4.3 Preferred Rate if a Payment to Each Pharmacy for Selected Interventions (Severe and Moderate only) is Made

The average rate of remuneration for the model where only severe or moderately significant

interventions were remunerated was $18.92 (range $2 to $100) (see Figure 7.4-7). This was

higher than the preferred rate for all interventions, indicating that pharmacists felt that more severe

interventions should be remunerated at a higher rate.


Page 353 of 361

0.00 20.00 40.00 60.00 80.00 100.00

A payment to each pharmacy for high/mod interventions

0

5

10

15

20

25

30

Fre

qu

en

cy

Mean = 18.92Std. Dev. = 15.72205N = 75

Figure 7.4-7: Preferred Rate of Remuneration for Selected Intervention Pharmacy Payment

7.4.4.4 Preferred Rate if a Payment to Each Pharmacist for Selected Interventions (Severe or Moderate only) is Made

The average rate of remuneration for payment to pharmacists for the severe or moderate

interventions only was $17.86 (range $3 to $80) (see Figure 7.4-8). This was in keeping with the

result for a payment to the pharmacy for the same type of intervention.


Page 354 of 361

0.00 20.00 40.00 60.00 80.00

A payment to each pharmacist for high/mod interventions

0

10

20

30

40

Fre

qu

en

cy

Mean = 17.863Std. Dev. = 13.51164N = 73

Figure 7.4-8: Preferred Rate of Remuneration for Selected Intervention Pharmacist Payment

7.4.4.5 Preferred Rate if an Increase in the Dispensing Fee for Prescriptions for High-risk Drugs Occurred

The average increase in dispensing fee suggested in this model was $9.32 (range $1 to $50) (see

Figure 7.4-9).

The current dispensing fee (Aug 2005) is $4.75, and an increase of $9.32 for selected items with a

high frequency of interventions may have significant implications on the overall PBS budget. The

extent of impact would depend on which drug groups were selected and their frequency of

dispensing.


Page 355 of 361

0.00 10.00 20.00 30.00 40.00 50.00

An increase in the dispensing fee for drugs with high frequency of intervention eg

warfarin

0

5

10

15

20

Fre

qu

en

cy

Mean = 9.3167Std. Dev. = 8.2657N = 60

Figure 7.4-9: Preferred Rate of Remuneration for Increased Dispensing Fee


Page 356 of 361

8 Potential Improvements to the Intervention Recording System

There are a number of improvements that could be made to the software and to the classification

system used in the PROMISe intervention study. Pharmacists who participated provided a number of

useful ideas for improvements, and the Project Team also identified opportunities for improvements.

The pharmacists involved in the Melbourne PROMISe trial used the recording software for

approximately eight weeks. The participants interacted with the software at varying levels. Individual

recording rates of interventions varied from 0.1 to 6% of prescriptions. Feedback concerning the

software was facilitated by DeBoos Associates from both pharmacists who had used the software

extensively and those who had used the software minimally. Feedback on the software was also

gathered from the participants via the post-study Questionnaire. The following suggestions were

made by the participants.

8.1 Software and IT Changes Suggested

8.1.1 System Requirements

Throughout the PROMISe intervention study the variation in individual computer system setups

caused a number of problems. For future uptake of the intervention documentation software, we

recommend specifying a set of minimum system requirements would improve this situation. Many

pharmacists were not aware of their own pharmacy’s operating system.

A suggested set of minimum system requirements would be:

� Operating System: at least Windows 2000 on all terminals

� Server terminal to be clearly identified

� Broadband internet connection

� Details of firewall and anti-virus software are provided

� Pharmacies to inform the Project Team of any changes in hardware made during the trial

8.1.2 Improvement of Software Installation and Testing

The pharmacists involved in the feedback sessions had experienced ‘teething’ problems with the

software which was loaded onto their system. Subsequennt problems were also experiences with the

update to WiniFRED and the Comm Server. These problems were compounded by the variations in

firewall, virus software, operating systems, broadband and dial up internet seen from pharmacy to

pharmacy. The impact of these variations was greater than anticipated. In some instances these

problems caused the system to shut down. Allowing additional time for initiation of the software could

improve this aspect of the trial rollout.


Page 357 of 361

8.1.3 Incorporate More Educational Alerts

Each respondent who had experienced the educational alert (activated in 31 pharmacies) were

enthusiastic to incorporate more alerts in the future. They also provided a few suggestions as to how

to minimise potential fatigue of these alerts. These included

� Educational alerts which are automatically deactivated when the pharmacist had seen the

alert for a specified number of times

� Educational alerts which only displayed for each patient once

� For the alert to turn off after a certain period of time

� Education alerts which could be selected each month with the dispensing software update

8.1.4 Adjustments to the Interface

The pharmacists involved in the trial found that the software was easy to access; through either Alt I or

through the Activities menu. The adjustments suggested included

� Where the notes sections were available, for these to have an unlimited number characters

� A separate window with which to show interventions so the pharmacist could switch between

the patients dispensing history and the intervention

� Interventions to appear in a different colour within the dispensing history (during the

Melbourne PROMISe trial they appeared as the same colour as the note function within the

dispensing history)

� A system which flagged an intervention where it needed further follow-up

� A way to view all the interventions for that particular day

� Improve the appearance of the intervention within the list of dispensed medications

� Allow the comments sections around significance to be accessed for interventions of low or

mild significance

� Utilise more drop down menus

� Utilise more keyboard functionality and reduce the requirement for using the mouse

� Improve the reporting functionality so it could be more widely used


Page 358 of 361

8.2 Changes to the PROMISe System Overall

8.2.1 Refinement of Classification

Some pharmacists involved found it difficult to identify quickly the appropriate category which to

classify their intervention under. Implementation of a way to record interventions when the category

was not evident was suggested. It may also be appropriate to include a facility whereby the recording

pharmacist submits their intervention and gets suggestions for the appropriate category, particularly

when the participants are initially using the software.

8.2.2 Mentor program

One of the pharmacists involved in the feedback on the project suggested that a mentor system could

improve the initial usage of the system and possibly the amount which was recorded. This system

would overcome some of the initiation problems and hence improve the uptake of the program

generally. A mentor would have been involved in the trial previously and be able to provide the user

with practical feedback on the program.

8.2.3 Educational alerts in the dispensing software

Those who had access to the educational drug alert found it was useful to improve patient

relationships. The use of more alerts or a range of alerts was seen as a way to improve the existing

system. It was the general consensus that the individual pharmacist needed control of the activation

of these alerts.

8.2.4 Online training

For the clinical skills cases used in the online training it was suggested that review be undertaken of

case 1. The pharmacists felt that were they were given an overall score and if that score was low this

was a disincentive to continue with the training. Using a system which first trained the pharmacists in

how to complete the clinical problem solving tool could be implemented. This training module could

include prompts to identify the most appropriate answer.

With regard to the online training in the DOCUMENT system, 20 case scenarios was seen as a flexible

way to complete training in the system. Training such as this would allow for uptake of the program by

a range of pharmacists, and distance and isolation would no longer be an issue.

8.2.5 Continuing Education

Targeted continuing education which incorporated potential instances of intervention was seen as a

broader approach to improving the uptake of the system. Some of the pharmacists involved in the

Melbourne PROMISe trial experienced difficulty identifying interventions. Hence increasing their rate of

intervention would be extremely difficult. Targeted continuing education which improved clinical

knowledge would in turn improve the quantity and quality of interventions.


Page 359 of 361

8.2.6 Workshop sessions

Providing workshop sessions was seen as a potential way to improve awareness of what could be

recorded as part of the PROMISe project. These type of sessions could involve face to face meetings

or be held remotely via telephone or the internet. Although the pharmacists were provided with

information in both the Pharmacy Resource manual and the ‘trial participant’ section of the PROMISe

website about examples of interventions they also requested more information on this aspect. Using a

workshop format could encourage the pharmacists to formulate suggestions and increase the general

knowledge of what could be recorded.


Page 360 of 361

9 Conclusions and Recommendations

We believe that our study has demonstrated that the current value of Australian community

pharmacists’ interventions in both health and financial terms is high. However, there is considerable

scope for increasing this impact; and it is likely that both the existing rate and the financial value of

pharmacists’ interventions could be increased three-fold.

There are a number of recommendations for further investigations that are a logical followon from this

study.

Educational Prompts

Automated educational alerts within computerised dispensing systems possess significant potential for

increasing pharmacists’ intervention rates, and thereby improving the quality use of medicines.

1. We recommend that other key interventions that are amenable to this educational alert

technique be investigated.

Economic Analysis

The economic analysis technique we used to evaluate the interventions in this study is unique. We

believe it to be accurate and conservative, and capable of use in a range of situations where changes

to therapeutic management are made.

2. We recommend that the economic assessment technique used in this study be used for the

assessment of other intervention data sets that may be made available.

3. We recommend a comparison of our current analysis with alternative economic analysis

techniques for the PROMISe data set.

Detailed Economic Analysis

Our current economic analysis examines all clinical interventions and their value. More detailed

analysis could focus on particular types of interventions such as proactive interventions, interventions

relating to drug selection problems, or interventions resulting in a recommendation to seek medical

attention.

4. We recommend that a detailed economic analysis of the PROMISe dataset be undertaken

with a view to establishing values for different types of interventions.

Factors Affecting Intervention Rate

Our study has established that there are underlying factors within pharmacies and pharmacists that

influence theor rate of performance of clinical interventions. Remuneration, observation and the

presence of an educational intervention prompt had a significant impact in parts of our study.


Page 361 of 361

Establishing what other factors are involved may enable targeting of educational and other techniques

that could be used to increase rates of interventions.

5. We recommend that pharmacist and pharmacy specific factors associated with increased

clinical intervention rates be established.

Representability of the PROMISe Data

Our study utilised information from 52 representative pharmacies in the greater Melbourne area. There

were no pharmacies involved from other States in Australia, and no pharmacies from Rural areas

involved.

6. We recommend that clinical intervention recording be trialled in rural and remote areas of

Australia to increase the representativeness of the information.

Intervention Classification System

The DOCUMENT classification system was pilot tested in Tasmania, modified and used for the

PROMISe intervention study. Several improvements to the classification have been suggested and

identified.

7. We recommend that the DOCUMENT classification system be reviewed and simplified for

more widespread use.

Intervention Recording Software

The intervention rrecording software we developed was sufficient for our purposes, but there are a

number of improvements that could be made to the installation process and the software itself.

8. We recommend that the WiniFRED Intervention Recording Software be modified according

to feedback from users in the PROMISe study.

Documents

PROMISe Intervention Study