How to Invesitigate Antimicrobial Drug Use in Hospitals

8/12/2019 How to Invesitigate Antimicrobial Drug Use in Hospitals

1/76

How to InvestigateAntimicrobial Drug Use in Hospitals:

Selected Indicators

Working Draft

May 2001

Rational Pharmaceutical Management Plus ProgramCenter for Pharmaceutical Management

Management Sciences for Health4301 North Fairfax Drive

Arlington, VA 22203 USAPhone: 703-524-6575

Fax: 703-524-7898e-mail: [email protected]


2/76

H ow to Investigate Anti micr obial Dr ug Use in H ospitalsi i

This publication was made possible through support provided by the U.S. Agency forInternational Development, under the terms of cooperative agreement numbers HRN-A-00-92-00059-13 and HRN-A-00-00-00016-00. The opinions expressed herein are those of the authors

and do not necessarily reflect the views of the U.S. Agency for International Development.

Recommended Citation

Rational Pharmaceutical Management Plus Program. May 2001. How to Investigate Antimicrobial Drug Use in Hospitals: Selected Indicators. Draft published for the U.S. Agencyfor International Development by the Rational Pharmaceutical Management Plus Program.Arlington, VA: Management Sciences for Health.


3/76

H ow to I nvestigate Anti microbial Dr ug Use in H ospitals ii i

CONTENTS

ACRONYMS...................................................................................................................................vI. FOREWORD........................................................................................................................... 1II. INTRODUCTION................................................................................................................... 1III. PURPOSE 3IV. Objectives of a Hospital Antimicrobial Use Study ................................................................. 3V. BACKGROUND TO DRUG USE INDICATORS ................................................................ 4VI. Indicator Format ...................................................................................................................... 5VII. Drug Use Indicators................................................................................................................. 6

A. Hospital Indicators ............................................................................................................. 6B. Prescribing Indicators......................................................................................................... 6C. Patient Care Indicators ....................................................................................................... 6D. Supplemental Indicator ...................................................................................................... 7E. Description of Hospital Indicators...................................................................................... 8

Indicator 1. Existence of standard treatment guidelines (STGs) and a list of officiallysanctioned antimicrobial drugs in a formulary list........................................ 8

Indicator 2. Availability of a set of key antimicrobial drugs in the hospital stores on theday of the study ............................................................................................. 9

Indicator 3. Average number of days that a set of key antimicrobial drugs is out of stockin a 12-month period ................................................................................... 10

Indicator 4. Expenditure on antimicrobial drugs as percentage of total hospital drugcosts ............................................................................................................. 11

F. Description of Prescribing Indicators ............................................................................... 12Indicator 5. Percentage of hospitalizations with one or more antimicrobial drugs

prescribed .................................................................................................... 12Indicator 6. Average number of antimicrobial drugs prescribed per hospitalization with

antimicrobial drugs prescribed .................................................................... 13Indicator 7. Percentage of antimicrobial drugs prescribed consistent with the hospital

formulary list ............................................................................................... 14Indicator 8. Average cost of antimicrobial drugs prescribed per hospitalization with

antimicrobial drugs prescribed .................................................................... 15Indicator 9. Average duration of prescribed antimicrobial drug treatment .................... 16Indicator 10. Percentage of surgical inpatients who received antimicrobial drug

prophylaxis .................................................................................................. 18Indicator 11. Percentage of patients with pneumonia who are prescribed antimicrobial

drugs in accordance with standard treatment guidelines (STGs) ................ 19Indicator 12. Percentage of antimicrobial drugs prescribed by generic name .................. 21

G. Description of Patient Care Indicators ............................................................................. 23Indicator 13. Percentage of doses of prescribed antimicrobial drugs actually administered

23Indicator 14. Average duration of hospital stay of patients who receive antimicrobial

drugs (H/W)................................................................................................. 24H. Supplemental Indicator .................................................................................................... 25

Indicator 15. Number of antimicrobial drug sensitivity tests reported per hospitaladmission including antimicrobial treatment .............................................. 25


4/76

H ow to Investigate Anti micr obial Dr ug Use in H ospitalsi v

VIII.How to Conduct an antiMICROBIAL DRUG use study...................................................... 26A. Purpose and Design of the Study..................................................................................... 26

Types of AMD Studies ........................................................................................................................ 27B. Design Criteria ................................................................................................................. 27C. Planning and Field Methods............................................................................................. 29

IX. How much does an assessment cost? .................................................................................... 33REFERENCES ............................................................................................................................. 33ANNEX A..................................................................................................................................... 35DETAILED INSTRUCTIONS AND SAMPLE DATA FORMS................................................ 35ANNEX B. .................................................................................................................................... 55BLANK DATA COLLECTION FORMS .................................................................................... 55


5/76

H ow to I nvestigate An timi crobial D ru g Use in H ospitals

ACRONYMS

ADR adverse drug reactionAMD

BAN

antimicrobial drug

British Approved NameCH clinical historyDTC drug and therapeutics committeeFL formulary listICIUM International Conference on Improving Use of

MedicinesINN international nonproprietary nameIV intravenousLQASPI

lot quality assurance system principal investigator

RPM Rational Pharmaceutical Management [Project]STG standard treatment guidelineUSAIDUSAN

U.S. Agency for International DevelopmentU.S. Approved Name

VEN vital, essential, nonessentialWHO World Health Organization


6/76

H ow to Investigate Anti micr obial Dr ug Use in H ospitalsvi


7/76

H ow to I nvestigate An timi crobial Dr ug Use in H ospitals 1

I. FOREWORD

The Rational Pharmaceutical Management (RPM) Project of the U.S. Agency for InternationalDevelopment (USAID) has produced two drug assessment manuals based on indicators ( Rapid

Pharmaceutical Management Assessment: An Indicator-Based Approach and Drug Management

for Childhood Illness Manual). Both of these manuals were aimed at assessing outpatient druguse as opposed to hospital inpatient use and were directed at a broad range of pharmaceutical products, not specifically antimicrobial drugs (also referred to in this document as antimicrobialsor AMDs). In 1993, the World Health Organization (WHO) published How to investigate druguse in health facilities: selected drug use indicators . Only 2 of the 19 indicators in that

publication specifically referred to antibiotic use or management.

In response to antimicrobial drug resistance legislation by the U.S. Congress, RPM proposed toUSAID the development of a manual for assessment of antimicrobial drug use in hospitals. Sucha manual would be a tool for hospital managers to assess antimicrobial drug management anduse and thus contribute to reducing antibiotic misuse, a congressionally mandated objective forUSAID worldwide.

The present manual is intended for use by hospital drug and therapeutics committees (DTCs), physicians, pharmacists, and managers, as well as drug use researchers who wish to evaluate andimprove antimicrobial drug use in hospitals. It will allow basic comparisons of antimicrobialdrug use in a hospital over time and between hospitals.

The manual is divided into two sections. The first describes the indicators for antimicrobial druguse and management according to a standard format, and the second suggests procedures toapply them in a hospital study.

II. INTRODUCTION

The WHO conference on the rational use of drugs in 1985 marked the beginning of efforts toimprove the use of drugs, especially in developing countries (1). In 1993, the WHO ActionProgramme on Essential Drugs (WHO/DAP) published the manual How to investigate drug usein health facilities in response to the increased awareness of the problems impeding the rationaluse of drugs (2). This manual presented 12 indicators to assess drug use in outpatient healthfacilities and has been instrumental in standardizing drug use studies.

The manual has been used to assess drug use in hospitals, even though the drug use indicators foroutpatient settings do not address a number of the factors and situations that affect drug use inhospitals such as the duration of stay or the different diseases treated. For example, an indicatorsuch as the time to dispense a prescription to an ambulatory patient is meaningless in aninpatient setting. Similarly, the type and severity of illness that causes patients to be hospitalizedoften necessitates the use of intravenous drugs. Therefore, the indicator percentage ofinjectables prescribed would be expectedly higher in hospitals than in outpatient facilities, andthus less meaningful for inpatient drug use.


8/76

H ow to I nvestigate An timi crobial D ru g Use in H ospitals2

The First International Conference on Improving Use of Medicines (ICIUM), held in Thailand in1997, identified the need for a set of indicators and appropriate methodology to assess the use ofdrugs in hospitals, particularly antimicrobial drugs (3). The detection of problems withantimicrobial drug use in hospitals is the first step in evaluating the underlying causes and takingremedial action.

The management and use of antimicrobial drugs has clinical, economic, and environmentalimplications. In many countries, antimicrobial drugs are the most frequently prescribedtherapeutic agents, accounting for 30 to 50 percent of drug prescriptions. From a clinicalstandpoint, there are three principal concerns surrounding the use and management ofantimicrobial drugs:

1. They are necessary for treatment of most bacterial infections. If they are not available inhospital pharmacies, lives may be jeopardized.

2. They may be prescribed inappropriately by physicians and drug sellers and especially bythe general public (self-prescribing) where antimicrobial drugs are sold over the counter.Inappropriate prescribing includes use of antimicrobial drugs without proof of infectionor to treat viral infections or noninfectious diarrhea. The wrong drug may be prescribedor taken for a particular infection or, if the correct drug is used, it may be prescribed ortaken at the wrong dosage or by an inappropriate route of administration. Perhaps thegreatest misuse of antimicrobial drugs is shortened duration of therapy.

3. Adverse drug reactions (ADRs) constitute the third critical area of antimicrobial drug use.Such reactions include nephrotoxicity and allergic reactions as well as antibiotic-associated diarrhea. It is estimated that 25 percent of ADRs are caus ed by antimicrobialdrugs (4).

In conclusion, hospitals must ensure availability of antimicrobial drugs while at the same timecontrolling and improving prescribing practices of physicians and minimizing untoward sideeffects.

In economic terms, expenditures on antimicrobial drugs in the year 2000 are projected at $40 billion with about one-third of that in developing countries. Worldwide, antimicrobial drugsaccounted for 12 percent of all pharmaceutical sales in 1990 while in developing countries theyaccounted for 19 percent; this proportion was projected to increase to 34 percent in 2000 (5).Thus, antimicrobial drugs are a large and growing component of pharmaceutical expenditures indeveloping countries and must be managed effectively in the face of limited financial resources.

Finally, cases of antimicrobial drug resistance have drawn public attention to the public healthand ecologic implications of drug-resistant microorganisms in the environment. The veryubiquity of antimicrobial drugs and their frequent over- and under-use are associated with theappearance of antimicrobial drug-resistant strains of major pathogens such as Staphylococcusaureus and Pseudomonas (6, 7) . Highest levels of resistance to antimicrobial drugs occur incountries with highest levels of antimicrobial drug use (8, 9, 10). It has been estimated thathospital-acquired infections due to drug-resistant organisms cost the United States $4 billion in


9/76


1990 (11). The high cost of treating drug-resistant infections may exceed the financial capacityof many patients and hospitals in developing countries. Thus, managers must monitor andminimize antibacterial resistance in their hospitals.

III. PURPOSE

The purpose of this manual is to define a limited number of indicators that will objectivelydescribe the management and use of AMDs in hospitals and to provide tools and step-by-stepinstructions to design and carry out an assessment of antibiotic use and management in hospitals.The indicators in this manual will complement the existing WHO indicators of outpatientantimicrobial drug use suggested in How to investigate drug use in health facilities (including

percentage of encounters in which an antibiotic was prescribed and percentage of drug costsspent on antibiotics) and will address the need for AMD indicators for inpatient conditions. Themanual will follow the pattern of previous RPM assessment guides and the WHO publication by

presenting a limited number of indicators useful for screening, monitoring, and assessing impact.Because these indicators do not need adaptation and can be used in any indicator-based AMDuse study, they provide a simple tool for quickly and reliably evaluating critical aspects of AMDuse in hospitals. A supplemental indicator of the use of sensitivity testing is also presented. Thismay have limited application because of limitations in laboratory services.

This manual is intended as a rapid assessment tool that can be used by hospital administrators,DTCs, researchers, and program managers in developing countries to identify problems withantimicrobial drug use in their hospitals. Once problems have been detected, investigators willneed to interpret the meaning of the results in the context of the hospital (size, type of patient,level of complexity) and probe more deeply to uncover possible underlying causes. Thus, forexample, the hospital indicator expenditure on antimicrobial drugs as percentage of totalhospital drug costs may show that antimicrobial drugs account for 80 percent of a hospitals

budget. This may seem excessive, but circumstances of the hospital may warrant such a figure.The cost of antimicrobial drugs in a pulmonary hospital with many tuberculosis cases would behigh, while in a maternity hospital it would probably be low. If the cost of antimicrobial drugs isinappropriately high, this may be due to several factors, including physicians using expensive,

brand-name antimicrobial drugs instead of generic products on the formulary list; physicianstreating the majority of patients with multiple antimicrobial drugs when this is not indicated;antimicrobial drugs being procured at high cost due to poor procurement practices; or acombination of all these factors. Further analysis will therefore be needed to determine the rootcause of the problem.

IV. OBJECTIVES OF A HOSPITAL ANTIMICROBIAL USE STUDY

Hospital administrators, researchers, and DTCs will want to study antimicrobial drug use inorder to

1. Describe antimicrobial drug prescribing practices2. Compare performance among hospitals or prescribers


10/76


3. Monitor performance and orient supervision4. Assess changes resulting from interventions

Once problems have been detected, investigators will decide whether further study is warrantedto explore causes of the problems detected in the first round. These additional studies will

explore areas such as?? Antimicrobial drug selection procedures and criteria?? Antimicrobial use in specific wards or specialties or by individual prescribers?? Purchasing and financing of drugs and antimicrobial drugs?? Impediments to patient compliance with antimicrobial drug prescriptions?? Comparison of antimicrobial drug use among hospitals

Investigators should clearly state why the study is needed and what is expected as the outcome.For example, hospital management may want to use the following wording: Undertake a rapid,hospital-wide review of antimicrobial drug use and management to detect problem areas and

assign responsibility for correction to the respective departments.

V. BACKGROUND TO DRUG USE INDICATORS

Drug use indicators are standardized measurements of various aspects of hospital operationsrelated to drug management and use that can be compared to normative ranges in order toestablish adequacy of performance or other diagnostic conditions. They may be quantitative orqualitative. In order to be useful, indicators should be

Important Each indicator must reflect an important dimension of performance. Even

though data may be readily and consistently available, they may not say anythingimportant about the system performance.

Measurable Each indicator must be measurable, within existing constraints of time andvariable quality and availability of source data. It might be desirable to measure AMD

prescribing practices retrospectively, but if prescriptions are not written completely inclinical records, the indicator is not measurable.

Reliable Each indicator must give consistent results over time and with differentobservers. If one observer reports a certain result from a set of data, it is expected that asecond observer will report the same result.

Valid Each indicator must allow a consistent and clear interpretation and have a similarmeaning across different environments.

The indicators described in this manual are not all-inclusivethey do not measure every aspectof AMD use in hospitals. Also, they are best understood as proposed, standardized measuressince normative ranges have not been established that would define cut-off points betweenacceptable and unacceptable performance.


11/76


VI. INDICATOR FORMAT

The indicators described in the following section follow the format summarized below.

Indicator Name: The name of the indicator along with a code for the different hospital

locations that may be studied. The location codes areH Hospitalthe facility as a wholeW Ward, service, or specialtyinternal medicine, pediatrics,

OB/GYN, surgery, etc.P Pharmacythe hospital pharmacy and ward drug cabinets

For example, H, W, P after the name of the indicator signals that it may bestudied at the hospital, ward, and pharmacy levels.

Rationale: The reason that this indicator is important.

Definition: The meaning of the indicator and the terms used to describe this indicator.

Data Collection : The most likely source(s) of information are summarized in a tableindicating where the data are to be collected, whom to ask for assistance,and what documents and records to review.

Brief discussions of methods and issues related to data collection.

Citations of the data collection forms to be used, if any. Data for 10 of theindicators are collected using five different forms. There is a discussion ofhow to develop the required forms in Section VIII, and examples are givenin Annex A.

Calculation: Calculations, if any, which are needed to derive the indicator.

Instrument: The specific data collection instrument and location of the data necessaryto calculate the indicator.

Example: Example of the use or results of the indicator.

Notes: Suggestions for additional information or discussion required to put theindicator in proper context or to provide more detail.


12/76


VII. DRUG USE INDICATORS

Fifteen drug use indicators related to antimicrobial drug use in hospitals are described below.Four are hospital-related, eight are prescribing indicators, and two relate to patient care. Asupplemental indicator is related to drug sensitivity testing.

A. H ospital I ndicators

Indicator 1. Existence of standard treatment guidelines (STGs) and a list of officiallysanctioned antimicrobial drugs in a formulary list

Indicator 2. Availability of a key set of indicator antimicrobial drugs in the hospital stores onthe day of the study

Indicator 3. Average number of days that a set of key indicator antimicrobial drugs is out ofstock in a 12- month period

Indicator 4. Expenditure on antimicrobial drugs as percentage of total hospital drug costs

B. Prescribin g I ndicators

Indicator 5. Percentage of hospitalizations with one or more antimicrobial drugs prescribed

Indicator 6. Average number of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed

Indicator 7. Percentage of antimicrobial drugs prescribed consistent with the hospitalformulary list *

Indicator 8. Average cost of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed

Indicator 9. Average duration of prescribed antimicrobial drug treatment

Indicator 10. Percentage of surgical inpatients who received antimicrobial drug prophylaxis

Indicator 11. Percentage of patients with pneumonia who are prescribed antimicrobial drugs inaccordance with standard treatment guidelines (STGs)

Indicator 12. Percentage of antimicrobial drugs prescribed by generic name

C. Patient Care Indicators

Indicator 13. Percentage of doses of prescribed antimicrobial drugs actually administered

* This may or may not be a part of the national essential drugs list or formulary list.


13/76


Indicator 14. Average duration of hospital stay of patients who receive antimicrobial drugs

D. Supplemental I ndicator

Indicator 15. Number of antimicrobial drug sensitivity tests reported per hospital admission

including antimicrobial treatment


14/76


E. Description of H ospital I ndicators

Indicator 1. Existence of standard treatment guidelines (STGs) and a list of officiallysanctioned antimicrobial drugs in a formulary list(H, W)

Rationale The existence of an STG and a list of essential (antimicrobial) drugs withaccurate, unbiased, and reasonably current information for prescribers concerningthe antimicrobial drugs approved for use in the hospital is a measure of thehospitals commitment to standards of patient care and rational drug use. Theformulary list ensures that authorized antimicrobial drugs will be procured on a

priority basis.

Definition For purposes of this indicator, the STG must be intended as a clinical referencefor prescribers and contain treatment protocols for the most frequent conditionsseen in the hospital. The latest revision must be no more than two years old. Theformulary list must be derived from the STG and be no more than one year old.

Where to look Whom to ask What to get Datacollection Hospital directors

officeDTCPharmacy

Hospital directorService chiefsDTC chairPharmacist

Copy of STGCopy of formulary list

The STG and formulary list must officially exist for this indicator to bemeaningful. Obtain the most recent copy of each and evaluate whether they have

been revised within two years and sanctioned by the hospital administration

and/or the DTC.

Calculation Record the existence of the STG and formulary list and when they were lastrevised.

Instrument The information for this indicator can be found on Instrument 1, questions 3 and6.

Example Hospital Y has a formulary list that was revised within the last 12 months and isintended for use by physicians, nurses, and the pharmacy. STGs exist only for the

pediatric service and were revised three years ago.


15/76


Indicator 2. Availability of a set of key antimicrobial drugs in the hospital stores on theday of the study(H, P)

Rationale Rational prescribing is based on the availability of needed drugs. If key

antimicrobial drugs are not present in hospital stores, patients may not receive thedrug of choice for their infections or may receive no treatment at all, with risk ofdeath.

Definition Indicator 2 measures the availability of key antimicrobial drugs in the hospital andthe hospital pharmacy drug supply management. The hospital must have aformulary list of key antimicrobial drugs authorized for use (see Indicator 1). Ifkey antimicrobial drugs are not defined by the hospital, these will need to bedetermined before using this indicator.

Where to look Whom to ask What to get Data

collection Hospital pharmacy Chief pharmacistDTC chair

Hospital formulary listGeneric and brandnames of antimicrobialdrugs on the formularylistPharmacy stock records

Calculation Percentage, calculated by dividing the number of key antimicrobial drugs actuallyin stock on that day by the number of key drugs that should be available,multiplied by 100.

Number of key antimicrobial drugs actually in stock x 100 Number of key antimicrobial drugs that should be available

Instrument The necessary information is found on Instrument 7 and is calculated by addingthe total numbers of entries in column 2 (current stock) that are more than 0 anddividing by the total number of products in column 1.

Example At hospital Z, only 75 percent of a set of key indicator antimicrobial drugs wasavailable on the day of the study. The pharmacist indicated that because the drug

budget was so low the hospital had decided to suspend purchase of the mostexpensive drugs, including five antimicrobial drugs. The DTC conducted an ABC

analysis and a VEN (vital, essential, and nonessential) analysis of all drugs andfound that the hospital was purchasing quantities of IV solutions and analgesics(nonessential) because of their low cost and was not purchasing several vitaldrugs (including the expensive antimicrobial drugs). The situation was explainedto the hospital director with the suggestion that hospital purchasing policyconsider the therapeutic importance of the drugs as well as their cost.


16/76


Indicator 3. Average number of days that a set of key antimicrobial drugs is out of stockin a 12-month period(H, P)

Rationale The average number of days that key antimicrobials are out of stock for the 12

months prior to the study is a measure of the availability of antimicrobial drugagents.

Definition Indicator 3 measures the probability that any of the key antimicrobial drugs wereout of stock during the past year. The average number of days that antimicrobialsare out of stock assesses a hospitals capacity to procure and distribute drugs andmaintain a constant supply.

Where to look Whom to ask What to get Data collection Hospital medical stores

Pharmacy

Manager

Pharmacist

Inventory records for

preceding 12 months

Calculation The average is calculated by dividing the sum of the number of days that each keyantimicrobial drug is out of stock over a 12-month period by the total number ofkey antimicrobials drugs.

Number of days that each key antimicrobial drug is out of stock over 12 months Number of key antimicrobials in the review

Instrument The information is found on Instrument 7 and is calculated by adding the totalnumbers of days out stock in column 15 and dividing by the number of productsin column 1.

Example In hospital I, key antimicrobial drugs were out of stock for an average of 66 daysover the past 12 months. The acting manager of the hospital medical storesindicated that the purchasing department was ordering drugs only when inventorywas completely depleted. When a permanent manager was hired, she appliedgood procurement practices and percentage of time out of stock decreased to 28days.


17/76


Indicator 4. Expenditure on antimicrobial drugs as percentage of total hospital drugcosts(H)

Rationale Drug costs generally represent a major expense for hospitals and, as such, should

be closely monitored. This indicator documents the cost of antimicrobial drugsrelative to other hospital drug costs. High percentages may suggest prescribing ofmultiple antimicrobial drugs, unjustified use, or the use of expensive, brandedantimicrobial drugs.

Definition Indicator 4 measures the relative expenditure on antimicrobial drugs as a portionof drug costs.

Where to look Whom to ask What to get Data collection Hospital pharmacy Pharmacist Records of all units of

drugs received or

purchased

Hospital medical stores ManagerPharmacist

Price list, purchaseorders

Calculation Percentage is calculated by dividing the total cost of all antimicrobial drugsdispensed by the total cost of all drugs dispensed and multiplying by 100.

Total cost of all antimicrobial drugs dispensed x 100

Total cost of all drugs dispensed

Instrument The information is found on Instruments 4 and 6 and is calculated by adding thetotal costs of antimicrobial drugs in column 2 on Instrument 6, dividing by thetotal cost of all drugs from column 4 on Instrument 4, and multiplying by 100.

Example In hospital M, antimicrobial drugs account for 45 percent of all drug costs. Aninvestigation by the chief pharmacist showed that 25 percent of expenditures wasfor antimicrobial drugs used in surgical prophylaxis. The DTC assisted

prescribers in preparing STGs for surgical prophylaxis. After implementation ofthe STGs, the percentage of hospital drug costs spent on antimicrobial drugs fell

by 10 percent.


18/76


F . Description of Pr escribin g I ndicators

Indicator 5. Percentage of hospitalizations with one or more antimicrobial drugsprescribed(H, W)

Rationale Antimicrobials used in hospitals to treat infections or for surgical prophylaxis areoften used inappropriately. This may result in prolonged morbidity, increasedduration of therapy, and development of antimicrobial drug resistance.

Definition Indicator 5 measures the extent of antimicrobial drug use in hospitals. When usedover time, it allows observation of changes in trends. When combined withIndicator 4 (antimicrobial drug cost per hospitalization), it will give informationon cost-effectiveness (e.g., if antimicrobial drug cost per hospitalization goesdown but antimicrobial drug prescribing remains constant, then cost of therapyhas been reduced). The interpretation of the indicator will depend on the type ofhospital and patients seen (e.g., psychiatric versus maternity versus infectiousdisease hospitals and patients). The indicator may provide valuable informationon prescribing behavior by ward, specialty, or diagnosis.

Where to look Whom to ask What to get Data collection Medical records

departmentManager or clerk Inpatient records

(treatment charts,nurses notes, doctorsnotes)

Calculation Percentage is calculated by dividing the number of patient hospitalizations during

which one or more antimicrobial drugs are prescribed by the total number ofhospitalizations studied and multiplying by 100.

Number of patient hospitalizationswith one or more antimicrobial drugs prescribed x 100

Total number of hospitalizations studied

Instrument The information is found on Instrument 2 and is calculated by adding the total ofYs in column 3, dividing by the total number of patients in column 1, andmultiplying by 100.

Example In hospital B, one or more antimicrobial drugs were prescribed in 47 percent of allhospitalizations. This was neither extremely high nor extremely low, so no furtherinvestigation was done.


19/76


Indicator 6. Average number of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed(H, W)

Rationale Hospital patients may receive more than one antimicrobial drug during a

hospitalization. This may be justified on clinical grounds but also may be due tounnecessary combination antimicrobial drug therapy; duplication of drugs; orfrequent, unjustified changes of therapeutic regimen. The purpose of this indicatoris to determine the extent of antimicrobial drug use in hospitals for those patients

prescribed antimicrobial drugs.

Definition Indicator 6 measures the average number of antimicrobial drugs prescribed perhospitalization.


department

Manager or clerk Inpatient records


Calculation The average is calculated by dividing the number of antimicrobial drugs prescribed for all hospitalizations in which an antimicrobial drug is used by thetotal number of hospitalizations with antimicrobial drugs studied. Differentformulations of the same antimicrobial drug should be counted as one.

Number of antimicrobials prescribed for all hospitalizationsTotal number of hospitalizations with antimicrobial drugs

Instrument The information is found on Instrument 2 and is calculated by dividing the total ofcolumn 11 by the total of column 3.

Example In hospital A, patients with antimicrobials prescribed are prescribed an average of2.3 antimicrobial drugs per hospitalization. This is an acceptable rate.


20/76


Indicator 7. Percentage of antimicrobial drugs prescribed consistent with the hospitalformulary list(H, W)

Rationale Formulary lists represent the drugs of choice for a hospital, as defined by thecompetent medical authority, and represent one way to optimize the use ofmedicines. Nonadherence to such hospital policy may be due to prescribers not

being aware of or in agreement with the list, listed antimicrobial drugs not beingavailable at the hospital, or prescriptions listed with brand names while drugs arestocked and dispensed under generic names.

Definition Indicator 7 measures the degree of prescriber adherence to the hospital formularylist. The formulary list is defined as the drugs approved by the DTC for purchaseand prescribing in the hospital. If such a list does not exist, it will be necessary torefer to an essential drugs list provided by the Ministry of Health.

Where to look Whom to ask What to get Data collection Hospital pharmacy Chief pharmacistDTC secretary

Hospital formulary list

Medical recordsdepartment

Manager or clerk Inpatient records(treatment charts,nurses notes, doctorsnotes)

Calculation Percentage is calculated by dividing the number of antimicrobial drugs prescribed

that are on the hospital formulary list by the total number of antimicrobial drugs prescribed and multiplying by 100.

Number of antimicrobial drugs prescribed that are on the formulary list x 100 Number of antimicrobial drugs prescribed

Instrument The information is found on Instrument 2 and is calculated by adding the numberof Ys in column 8, dividing by the total of column 6, and multiplying by 100.

Example In hospital C, only 54 percent of antimicrobial drugs prescribed were on thehospital formulary list. To the hospital DTC this appeared to be a low percentage,

and an assessment was done to examine the cause of nonadherence. The DTCfound that prescribers were not in agreement with the list and preferred to use brand names in their prescriptions. The DTC undertook a program to developtreatment protocols in each service and ward, insisting that prescribers achieve aconsensus on therapies and preferred drugs. The formulary list was revised and

prescribing adherence monitored on a monthly basis with results prominentlydisplayed. This program resulted in increased use of AMDs that were approvedand on the formulary list.


21/76


Indicator 8. Average cost of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed(H, W)

Rationale Antimicrobials typically account for 2040 percent of hospital drug expenditures.

Inappropriate treatment such as prescribing more antimicrobial drugs thanrecommended, prescribing higher doses or longer treatments than required, and prescribing brand name instead of generic antimicrobial drugs may increase costs.Determining the cost of antimicrobial drugs used during a hospitalization maylead to interventions that decrease hospital expenditures on antimicrobial drugs.

Definition Indicator 8 measures the cost of antimicrobial drug therapeutic practices in thehospital. If the only antimicrobial drugs administered in the hospital are thosesupplied by the hospital pharmacy, then the cost is defined as the most recent

purchase price for the drug. However, if the patients family purchasesantimicrobial drugs on the street or if inflation creates large price movements,then the cost should be defined as the published prices to the public on the day ofdata collection. If the indicator is to be measured over time, prices must bestandardized by an inflation factor.

Where to look Whom to ask What to get Data collection Inpatient records

(treatment charts,nurses notes, doctorsnotes)Hospital pharmacy

Chief pharmacistDTC secretary

Hospital formulary listGeneric and brandnames of drugs on thelist

Medical records

department

Manager or clerk

Calculation To find the average cost, divide the total cost of all antimicrobial drugs prescribed by the number of hospitalizations in which at least one antimicrobial drug was prescribed.

Cost of all antimicrobial drugs prescribed Number of hospitalizations in which at

least one antimicrobial drug was prescribed

Instrument The information is found on Instrument 2 and is calculated by dividing the total ofcolumn 16 by the total of column 3.

Example A study of hospital L found that the average cost of antimicrobial drugs prescribed per hospitalization was US$13.00. Two years before, this indicator wascalculated at US$4.50. Investigation revealed that the hospital was using morecostly sources of supply than in the past. Appropriate changes in supply sourceswere subsequently made to lower the average cost.


22/76


Indicator 9. Average duration of prescribed antimicrobial drug treatment(H, W)

Rationale The optimal duration of therapy for many bacterial infections has not beendetermined, but the current recommendation is usually 710 days of treatment.

Too short a course of treatment may prolong patient morbidity and promoteemergence of resistant organisms. Too long a course of therapy increases patientexposure to antimicrobial drugs, increasing the risk of adverse drug reactions, ofthe incidence of AMR, and of unnecessary expenditure on antimicrobial drugs.Frequent, unnecessary changes in antimicrobial therapy contributes to AMR, highcosts, and increased patient morbidity.

Definition Indicator 9 measures the intensity of patient exposure to antimicrobial drugsduring a hospitalization. The number of days on antimicrobial drug treatmentincludes the number of days of all antimicrobials prescribed for a patient duringthe hospitalization and does not distinguish routes of administration or a change indosage. This indicator measures the number of days of acute antimicrobialtreatment for each generic antibiotic and does not include antimicrobials for

prophylaxis or the treatment of tuberculosis.




Calculation The average duration is calculated by dividing the total number of days of

antimicrobial drug treatment by the total number of antimicrobials prescribed.Different dosage forms of the same generic drug (i.e., ampicillin injection andampicillin capsules) are counted as one. Surgical prophylaxis and treatment oftuberculosis are not included in this indicator .

Total number of days on antimicrobial drug treatmentTotal number of antimicrobials prescribed

Instrument The information is found on Instrument 2 and is calculated by dividing the total of

column 10 by the total of column 11.

Example In hospital L, the average duration of prescribed antimicrobial drugs treatmentwas 3.1 days. According to recommendations, this is a short duration of therapyfor most antimicrobial drugs, and a more detailed analysis of medical records wasundertaken. It was found that 71 percent of antimicrobial drugs prescribed werecephalosporins, for which the short treatment period was appropriate. However,


23/76


the high cost of these had led to exhaustion of the pharmacy budget by June, andantimicrobial drugs were out of stock by August.


24/76


Indicator 10. Percentage of surgical inpatients who received antimicrobial drugprophylaxis(H)

Rationale Antimicrobial prophylaxis is recommended before certain surgical procedures.

However, studies have shown that surgical prophylaxis is often administeredwhen there is no recognized indication for its use. Unnecessary prophylaxisincreases patient exposure to antimicrobial drugs, the likelihood of ADRs, andexpenditure on antimicrobial drugs, and it promotes the emergence of resistantorganisms.

Definition Indicator 10 measures the level of antimicrobial drug prophylaxis among surgical patients.

Where to look Whom to ask What to get Data collection Operating theater Nurse in charge Records of surgical

procedures performedon inpatients

Medical recordsdepartment


Calculation The percentage is the number of surgical inpatients receiving prophylaxis divided by the total number of surgical inpatients, multiplied by 100.

Number of surgical inpatients receiving prophylaxis x 100Total number of surgical inpatients

Instrument The information is found on Instrument 3 and is calculated by adding the total Ysof column 3, dividing by the total number of surgical patients from column 1, andmultiplying by 100.

Example In hospital L, 90 percent of all surgical inpatients received antimicrobial drug prophylaxis. In conducting a drug use review, the DTC found that antimicrobialdrug prophylaxis was being given even when it was not indicated, for example

before suturing clean lacerations. An algorithm to guide decisions on prophylaxiswas designed and implemented. Three months later, the percentage of surgical patients receiving antimicrobial drug prophylaxis had decreased to 60 percent.


25/76


Indicator 11. Percentage of patients with pneumonia who are prescribed antimicrobialdrugs in accordance with standard treatment guidelines (STGs)(H, W)

Rationale Indicator 11 measures the quality of patient care for a common infectious disease

based on treatment guidelines. Definition The compliance of prescribers with hospital treatment standards is defined by (1)

only the use of antimicrobial drugs of choice as defined in the STG and (2)observance of dosing indications for the same antimicrobial drugs. The indicatorcan only be evaluated if STGs for infectious diseases exist in the hospital (seeIndicator 1).


DTC office and/orhospital directors office

Records manager orclerk

DTC chairService chiefs and/orhospital director

Inpatient records(treatment charts,

nurses notes, doctorsnotes)STG

Calculation 1. The percentage is calculated by dividing the number of patients with pneumonia treated only with antimicrobial drugs contained in the hospitalsSTG by the total number of patients with pneumonia and multiplying by 100.

Number of pneumonia patients treatedonly with antimicrobial drugs per STG x 100

Total number of patients with pneumonia

2. The percentage is calculated by dividing the number of patients with pneumonia who were prescribed the correct dosage of the antimicrobial drugrecommended in the STG by the total number of patients with pneumoniareceiving the recommended antimicrobial drug according to the hospitals STGand multiplying by 100.

No. of pneumonia patients prescribed correct dose of correct AMD per STG x 100 No. of pneumonia patients who received the recommended AMD per STG

Instrument The information is found on Instrument 2 and is calculated by looking at eachtreatment for pneumonia shown by a Y in column 2 and comparing it with theSTG. Calculations can then be accomplished with the formulas listed above.

Example In hospital H, 54 percent of cases of pneumonia were treated only with therecommended antimicrobial drug according to the hospitals STG. Of these, 95

percent were prescribed the correct dose, according to the STG. This suggests thatsome prescribers are either not aware of the STG or are in willful noncompliance.


26/76


However, of those prescribers who prescribe the recommended antimicrobialdrugs, nearly all use the appropriate dose.


27/76


Indicator 12. Percentage of antimicrobial drugs prescribed by generic name(H, W)

Rationale If health care providers prescribe by generic names instead of brand names,confusion is avoided about multiple names for the same product. This simplifies

procurement and dispensing, facilitating generic substitution and improvinghospital efficiency.

Definition Indicator 12 measures the percentage of antimicrobial drugs that are prescribedusing their internationally recognized generic names, as identified in the WHO listof international nonproprietary names (INN) (12). The availability of genericallynamed drugs in the market, and the information available to prescribers, willinfluence the pattern observed. Although the INN is used as the official genericname for this indicator, in some countries, generic drugs may be available underINN and other variations such as United States Approved Name (USAN) orBritish Approved Name (BAN). For example, the antihistamine chlorphenamine(INN) may be marketed as chlorpheniramine (USAN). Or, only the USAN orBAN generic product may be marketed. In such situations, any of the commongeneric names may be counted for purposes of this indicator.



(treatment charts, nursesnotes, doctors notes)

Study organizers must develop a list of (or an explicit way of defining) the

specific product names to be included as generic drugs. Usually the generic namesof drugs are identified on the national drug formulary list. Data collectors must beable to observe the actual names used to describe the drugs prescribed, as opposedto having access only to the names of the products dispensed. Thus, thedispensing ledger (if there is one) may not be an accurate source for this indicator.

Calculation The indicator is recorded as a percentage, computed by dividing the number ofantimicrobial drugs prescribed by generic name by the total number ofantimicrobial drugs prescribed, and multiplying this quotient by 100.

Total number of antimicrobial drugs prescribed by generic name x 100

Total number of antimicrobial drugs prescribed

Instrument The information is found on Instrument 2. The percentage is calculated bydividing the total of column 7 by the total of column 6 and multiplying by 100.

Example For example, results for hospital X are calculated as follows:


28/76


Column 6 = Total AMDs = 51 Column 7 = INN = 41

% of AMDs = 41 x 100 = 79% prescribed by 51

generic name


29/76


G. Description of Patient Care I ndicators

Indicator 13. Percentage of doses of prescribed antimicrobial drugs actually administered(H, W)

Rationale To be effective, the doses of antimicrobial drugs that are prescribed must beadministered. If they are not, the reasons may include unavailability of drugs,dispensing errors, or nursing errors.

Definition Indicator 13 measures the extent to which prescribed antimicrobial drugs actually reach the patient in the dosages prescribed. The basic assumption is thatadministration of the drug, no matter what its source (hospital pharmacy, patientsfamily), is recorded on the patients treatment record or in nurses notes. If thisassumption is not true, the indicator cannot be calculated.


collection Medical recordsdepartment


Calculation The percentage is calculated by dividing the number of doses of antimicrobialdrugs administered by the total number of doses of antimicrobial drugs prescribedand multiplying by 100.

Number of doses of antimicrobial drugs administered x 100Total number of doses of antimicrobial drugs prescribed

Instrument The information is found on Instrument 2 and is calculated by dividing the total ofcolumn 14 by the total of column 13 and multiplying by 100.

Example In hospital D, 80 percent of doses of prescribed antimicrobial drugs were actuallyadministered in the hospital. Investigation by the DTC revealed that doses werenot given because nurses were often busy with other duties at the time themedication would have been administered. Some of the nurses duties werereassigned to nurses aides, and a study six months later showed that 90 percent of

prescribed doses of antimicrobial drugs were being administered.


30/76


Indicator 14. Average duration of hospital stay of patients who receive antimicrobial drugs(H/W)

Rationale The length of treatment with antimicrobial drugs should not exceed therecommended duration. If patients who received antimicrobial drugs do not

improve within that time, it is possible that treatment is not appropriate or thediagnosis is incorrect. Prolonged hospital stay is costly, increases the risk to the patient of nosocomial infections, and promotes the emergence of organismsresistant to antimicrobial drugs. If the hospital stay is too short, antimicrobial drugtreatment may be ineffective because of subtherapeutic treatments, relapses,

promotion of antimicrobial drug resistance, and finally, increased costs.

Definition Indicator 14 measures the duration of hospital stay as an index of treatmenteffectiveness.


collection Medical recordsdepartment


Calculation The average is calculated by dividing the total number of days for allhospitalizations for patients treated with antimicrobial drugs by the number of

patients treated with antimicrobial drugs.

Total number of days of hospitalization for patients receiving antimicrobial drugs

Number of patients receiving antimicrobial drugs

Instrument The information is found on Instrument 2 and is calculated by finding the total ofcolumn 4 and dividing by the total of Ys in column 3 (remembering that if noAMDs were given to a patient, this column should be blank).

Example In hospital F, the average duration of stay for patients treated with antimicrobialdrugs was seven days. According to recommendations, this duration wasacceptable. However, further examination revealed that 10 percent of patientstreated with antimicrobial drugs stayed more than 30 days. These patients had

numerous, unwarranted changes in antimicrobial drug therapy.


31/76


H. Supplemental I ndicator

The following indicator measures the degree of compliance with STGs (where these exist) andthe use of antimicrobial drug sensitivity information (antibiograms or culture and sensitivitytesting) for establishing appropriate treatments. Because not all hospitals have STGs or conduct

sensitivity tests, this indicator is optional and may be used to supplement the information derivedfrom the core indicators at hospitals where the necessary data are available.

Indicator 15. Number of antimicrobial drug sensitivity tests reported per hospitaladmission including antimicrobial treatment

Rationale The use of effective antimicrobial drug therapy depends on knowing thesensitivity of infectious microorganisms to possible therapeutic agents. Thefrequency of sensitivity tests performed is a measure of the hospitals ability to

provide rational antimicrobial drug therapy.

Definition Indicator 15 measures the availability of antimicrobial drug sensitivityinformation to determine optimal treatment of infections.

Prerequisites A microbiology laboratory capable of cultivating specimens and testing forantimicrobial drug sensitivity is required. Reports from the laboratory indicatingthe results of sensitivity tests must be available in the patients clinical history inorder to calculate this indicator.

Where to look Whom to ask What to get Data collection Clinical history or

laboratoryManager or clerkLaboratory chief or

technician

Reports of antimicrobialdrug sensitivity tests

performed

Calculation Percentage is calculated by dividing the total number of sensitivity tests performed by the number of patients with antimicrobials prescribed.

Number of patients with a sensitivity test performed x 100 Number of patients with curative antimicrobials prescribed

It is recommended that hospitals utilize a minimum of 100 records involvingcurative antimicrobial use to obtain this information. Patients receiving

antimicrobials for prophylaxis must be excluded.

Instrument The information is found on Instrument 2 and is calculated by dividing the total Ysin column 5 by the total Ys of column 3 (that are curative and not prophylactictreatments) and multiplying by 100.

Example In hospital J, one antimicrobial drug sensitivity test was performed for every 5.3hospital admissions treated with curative antimicrobial drugs, or about 19 percent.


32/76


VIII. HOW TO CONDUCT AN ANTIMICROBIAL DRUG USE STUDY

A. Pu rpose and Design of the Study

AMD use studies serve one of the following four purposes:1. A facility assessment to screen AMD management and use in a hospital will assess

facilitywide aspects of supply and distribution and overall cost of AMDs. Results may bedisaggregated by ward or service. Multiple facilities may be compared.

2. Comparisons between groups or prescribers are generally cross-sectional surveys to compare prescribing and patient care between selected groups (wards, patients with specificdiagnoses, etc.).

3. To monitor performance and orient supervision with regard to a norm or goal, selectedindicators will be measured periodically in prospective, cross-sectional surveys to observechanges at facility, ward, or patient levels.

4. Assessments of changes resulting from interventions in control and intervention groups aresimilar to 3 above but usually follow a before-after design.

An AMD use study may be proposed by hospital management, the DTC, or individualinvestigators. Multifacility studies may be undertaken by ministries of health, nongovernmentalorganizations, or international organizations.

The agency sponsoring the study must determine the purpose of the study and assign a person orcommittee to develop an appropriate study design according to the type of record-keepingsystem, availability of resources, depth of information required, and so on, as suggested inTable 1.


33/76


Table 1. Basic Parameters of Different Types of Antimicrobial Drug Use Studies*

TYPES OF AMD STUDIESParameters

Facility Assessment Service/PrescriberComparison

Monitoring andSupervision

Assess InterventionImpact

Study objective Screen for AMDuse/management

problems

Detect AMD over-/underuse

Continuousobservation of key

problems

Determine whetheran interventionachieved itsobjective

Number of prescribingencounters

100 100 Apply lot qualityassurance systemaccording toinstruction in thereference below*

100

Type of prescribingdata

Retrospective, if possible, otherwise prospective

Retrospective or prospective

Prospective Retrospective

Time frame for prescribing data(retrospective)

One year One year Days, weeks, ormonths

At least 46 monthsafter intervention

Type of patient caredata

Retrospective, if possible, otherwise prospective

Prospective Prospective Retrospective

Time frame for patient care data

One year One year Days or weeks 12 months afterintervention

Type of hospitaldata

Retrospective purchasing and

stocking data

N/A Prospective purchasing and

stocking data

Prospective purchasing and

stocking dataTime frame forhospital data

One year N/A Weeks or months At least 46 monthsafter intervention

* Adapted from Table 4 in How to investigate drug use in health facili ties . WHO/DAP/93.1, World HealthOrganization, Geneva, 1993. Permission to adapt and reprint requested.

B. Design Cri teri a

1. Hospital type and groups to study. Antimicrobial drug use will vary depending on thesize/type (general, teaching, number of beds) and location (urban, rural) of the hospital.

Specialty hospitals (e.g., pediatric, cancer) also have expected patterns of AMD use. Within ahospital, wards or services may be studied if disaggregation is necessary and, within these, patients with specific diagnoses if warranted.


34/76


In an initial AMD study, it is likely that the whole hospital will be studied in order todetermine whether problems exist, how significant they are, and where they appear to belocated. In this case, the sampling unit will be the hospital and some 100 usable patientrecords will be studied, as suggested in the above table of basic parameters for an AMD usestudy.

2. Retrospective versus prospective clinic records. Generally, retrospective data collection froma sample of patient records will be the quickest way to conduct a study. The principalconcern in deciding for or against retrospective data collection is the availability andaccessibility of medical records. Before actually undertaking the study, a preliminary reviewof patient records should be conducted to ensure that it is possible to extract the requiredinformation efficiently. The following questions should be answered during this review:

?? Is there a chronological list of patients from which the sample can be drawn?

?? Are the clinical histories reasonably complete with regard to diagnoses, drug prescriptions, and drug administration?

?? Are patient records available for study for the period of the investigation?

?? If prescribing information is kept in the pharmacy, is it possible to link this with patient records?

With this information, investigators can make an informed decision about the accessibilityand usability of patient records and whether to adopt a retrospective or prospective design. Iftoo many records must be discarded as unusable (30 percent or more), a prospective study

may have to be conducted.Prospective data collection is also appropriate for measuring short-term changes in

performance following an intervention. However, since prescribers and dispensers will beaware of the data collection, their behavior is likely to be influenced by the study,introducing bias.

3. Data sources. Two source for collecting data are patient charts and hospital pharmacy andfinancial records.

?? Patient charts . A thorough understanding of the sections of the patient record (medical

orders/prescribing page, physician notes, nurses notes, treatment records, etc.) will benecessary as data may need to be collected from various sections of the patient chart.

?? Hospital pharmacy and financial records. These can be explored both retrospectively and prospectively, and stocks can be examined at various points in the distribution chain.


35/76


Once the facilitywide AMD study has revealed problems, these should be studied intensively,in which case the sampling unit may be the ward or service or a group of patients with thesame diagnosis.

4. Period of study. A long study period will minimize biases due to interruption of the drug

supply or seasonal outbreaks of febrile illnesses that may cause more AMDs to be prescribedthan usual. In general, AMD assessments should cover 12 consecutive months in order toensure that any seasonal variations are taken into consideration. All data collectors shoulduse the same 12-month period. If the study is prospective, the study period will probably beshorter, in the range of one to three months.

C. Plannin g and F ield M ethods

1. Preparation. The first step is to inform hospital management of the study and obtainconcurrence. Management will want to know why the study should be carried out, what thelikely outcomes are, how long it will take, how much staff effort will be required, how muchit will cost and where financing will come from, and finally, whether the study will interferewith usual operations. Investigators should ask for a meeting with the hospital director(s) andchiefs of services to explain these points and resolve questions.

Once hospital management has authorized the study, investigators will have to decide thefollowing details of the investigation.

?? Who . The typical team would consist of a principal investigator (PI) and one or twodata collectors who can be doctors, pharmacists, nurses, or students. The criticalconsiderations for data collectors are (1) experience with the clinical record system,

(2) knowledge of diseases and diagnoses, and (3) ability to interpret drug names anddosages. At least one data collector should be familiar with drug purchasing procedures and be responsible for this aspect of data collection.

?? How . The data collection instruments are the basic tools for the study. They must berevised to reflect the technical and administrative conditions of the hospital and thenreproduced in sufficient quantity. Also, a list should be prepared of AMDs availablein the market with generic and trade names and prices. The PI and/or data collectorscan usually process the data manually using a calculator or, if a computer is available,with an electronic spreadsheet.

??Where . Define data collection siteswards, statistics department, pharmacy,administration, and finance. It is also a good idea for the study to have an office orwork space with desks and chairs.

?? When . Develop a calendar of activitiestraining, data collection, critique, tabulation,and analysis. Allow about two days for training and a pilot trial, one week for datacollection, three days for critiquing and tabulation of data, and two to three days foranalysis and report writing.


36/76


?? What . Inform hospital staff of the study and what will be required of them. Usually,the heaviest loads are on statistics and finance or accounting departments and on the

pharmacy. Staff will generally make records available to the data collectors, and theactual time needed to complete this task will depend on how well these departmentsare organized.

2. Define a key set of antimicrobials. This may already be defined by the formulary list, VENanalysis, or by other mechanisms. It should include antimicrobials that are consideredessential and ones that have high use within the hospital.

3. Draw the sample according to the following instructions :

If the hospital cases are listed chronologically (the usual situation):1. Select a 12-month period and count or sum how many patients were hospitalized during

that period (e.g., 4,235).2. Assume that you will randomly and systematically sample 100 cases. Take the total

number of cases in the 12-month period and divide by 100. In the example, the result is42.35. This will be the sampling interval after rounding down to 42. That is to say,every forty-second clinical history (CH) will be examined for the study.

3. In order to start the sample, you need a random entry number. For example, take thefirst digit of the serial number on a bill of currency. Say it is 8. Find the eighth CH

beginning in the 12-month period.4. Now count CHs and select the forty-second as the first sampled CH. Count another 42

CHs and select that one. Continue counting and selecting until you have 100 CHs to besampled.

5. Examine each selected case and record the data on Instrument 2 for sampled cases.

If the cases are not listed chronologically and CHs are shelved in numerical order:1. Locate the CHs having the same date as yesterdays, one year earlier.2. Subtract the number of the first CH of that day a year ago from the last CH of yesterday

(e.g., yesterday was November 4, 2000, and the last CH opened yesterday was number12397). One year ago, the first CH on November 4 was number 10946. Subtracting thisfrom the former, the result is 1,451 CHs opened in one year.

3. Assume that 100 CHs from the last year are to be randomly and systematically sampled.Thus, divide the 1451 CH from last year by 100; the result is 14.51. Round this to thelower unit (i.e., 14). This gives the sampling interval which means that everyfourteenth CH will be selected for examination.

4. To begin the sample selection, a random index number is needed. Take the first digit othe serial number on a bill of currency (e.g., 9). Now, add this to the first CH from a yearago (10946) to produce the entry number (i.e., 9 + 10946 = 10955). Find the CH withthis number. This will be the first CH in the sample.

5. To select the next case, add the interval number to the last case (i.e., 9 + 10955 = 10964).Examine this case and collect the appropriate data.

6. Repeat step 5 until all of the 100 cases have been sampled through yesterdays date.

7. Examine each selected case and record the data on Instrument 2 for sampled cases.

If the cases are shelved in alphabetical order:1. For each letter, pull all the cases opened with a date in the last 12 months (e.g., A = 47

cases). Compile and count all CHs pulled (e.g., 2,481).2. Count all the CHs and divide by 100 to determine the sampling interval (e.g., 24.81).3. Select a random index number as in step 4, above.4. Beginning with the first CH pulled from letter A, add the index number and select the

CH of the patient. For example, if the index is 4, take the fourth case.5. Count through the remaining CHs up to the sampling interval to select the next CH.6. Repeat step 5 until all of the CHs pulled from last year have been sampled.7. Examine each case and record the data on Instrument 2 for sampled cases.


37/76


4. Train data collectors . Two days of training and practice are recommended, according to thefollowing training schedule. Staff from statistics, accounting, and nursing departments shouldexplain how their respective information systems record data, where the data are found, andhow to interpret them.

Illustrative Training Plan

5. Disseminate instructions for data collection instruments . See Annex A for detailedinstructions and sample data collection forms (Instruments 1-7) and Annex B for blankforms.

?? Basic information?? Form to record antimicrobial drug treatments?? Form to record surgical prophylaxis during the last year?? Form to record drug purchases during the past year?? Antimicrobial drugs purchased in the last year?? Cumulative purchase of antimicrobial drugs in the last year?? Availability of a set of key indicator antimicrobial drugs and time out of stock

Annex A includes detailed instructions for using each instrument. Each form has a note at thefoot of the page that identifies the indicator(s) for which the data are collected, as well asinstructions to data collectors about how to record the data and process the information.

Day Training Activities Time Opening introductions approval and support of hospital managementPurpose of the study sponsorsObjectives of the training master data collection instruments and datasourcesWhere to collect data clinical records, pharmacy, accounting/financeWhat to do if cases are incomplete or unusable, data are missing, etc.Calendar of activities data collection, critique, analysis, reporting

1 hr

2 hr

1 hr

1(a.m.)

(p.m.) Overview of data collection formsHow to interpret and record diagnoses, drug names, dosages, andtreatment regimensPractice filling in the forms and conduct exercises with recordscontaining missing or uninterpretable data

1 hr

2 hr

Conduct a pilot trial sample 10 records from a different period (e.g., 2years before) and collect per methodology and instruction

4 hr2(a.m.)(p.m.) Debrief on pilot trial critique and troubleshoot problems

Organize the data collection by distributing tasks among the teammembersHand out data collection instruments

2 hr1 hr


38/76


6. Collect data .

a. Locate the data to be collected in the clinical records or accounting documents.

b. Fill in forms. Once data have been transcribed from sources, additional information (e.g.,

prices, generic names) will have to be entered in order to complete the forms.c. Supervise collection. The PI should observe data collectors periodically to ensure

consistent interpretation and recording. This is especially important during the first day ofdata collection.

7. Critique data . Distribute the forms among the team so that each person has the forms ofanother team member. Read through the recorded data to ensure completeness andcoherency. A medical doctor should review the forms for diagnostic and treatment data.

8. Tabulate and process information. The team will process the data following the instructionson each form and in Annex A. The initial procedures are primarily counts and sums withsome rank ordering.

9. Calculate and analyze indicators . Each indicator is calculated according to the instructionsgiven in the text of Section VII of this guide. Most indicators are percentages or rates. Initialanalysis of the indicators will be primarily subjective, since standards do not exist. Each DTCor hospital management will have to interpret the results in light of its own treatment andadministrative norms and objectives.

10. Prepare a report of results. A brief written report is necessary to present the result to hospitaladministrators and the DTC. This will generally consist of a simple table or list. Ifcomparative results are presented (e.g., between facilities or in a facility at different times),these should be displayed graphically using bar charts. The report may be presented for

publication in house journals.

11. Report to the hospital or service. Call a meeting of hospital (or service) staff to present thestudy results. Time should be allowed for participants to understand the implications of theresults for their work and discuss possible actions to improve AMD management. Discuss theresults in a nonjudgmental way and ask participants for suggestions to explain unexpectedresults. Make sure that everyone participates by going around the group one-by-one askingeach for suggestions to improve performance. Record the suggestions for the report tomanagement.

12. Prepare to report to the administrative level. Invite all officials involved in drug use issuesincluding prescribers, nurses, pharmacy staff, and senior management to meet and discuss thestudy. Prepare and circulate the summary report with tables and graphs before the meeting.Hospital management should receive the written report of results as well as therecommendations of hospital staff for remedial actions.


39/76


13. Review with the administrative group the purpose of the study and present overall results . Ifsimilar studies have been conducted elsewhere in the country or outside, compare results andconduct a discussion of similarities and differences. Explore the possibility of making thestudy an ongoing monitoring activity of the hospital within routine supervision activities.Record conclusions and decisions and define follow-up actions.

IX. HOW MUCH DOES AN ASSESSMENT COST?

The cost of a typical two-week assessment is primarily determined by staff time. The only otherinputs are the forms, which cost less than US$5. If three people conduct the study, the cost isessentially 1.5 person-months of effort. A trial study of this methodology took three people oneweek for data collection in each hospital and one week to process information and write a reportfor each hospital.

REFERENCES

1. World Health Organization (WHO). Nairobi conference of experts on the rational use ofdrugs (1985 Nairobi Conference). Geneva: WHO. 1987.

2. World Health Organization (WHO). How to investigate drug use in health facilities: selecteddrug use indicators. WHO/DAP/93.1. Action Programme on Essential Drugs. Geneva:WHO. 1993.

3. Program and Abstracts from the First International Conference on Improving Use ofMedicines (ICIUM). Chaing Mai, Thailand. 1997.

4. Beringer, PM, A Wong-Beringer, and JP Rho. Economic aspects of antibacterialadverse effects. Pharmacoeconomics . 1998; 13:3549.

5. Liss, RH, and FR Batchelor. Economic evaluations of antibiotic use and resistancea perspective: report of Task Force 6." Review of Infectious Diseases . 1987; 9 Suppl 3:S297 312.

6. Urassa, W, E Lyamuya, and F Mhalu. Recent trends on bacterial resistance to antibiotics. East African Medical Journal . 1997; 74:12933.

7. Centers for Disease Control. Staphylococcus aureus with reduced susceptibility tovancomycinUnited States, 1997. Morbidity and Mortality Weekly Report ; 1997. (46)7656.

8. World Health Organization. Control of antibiotic-resistant bacteria: memorandumfrom a WHO meeting. WHO Scientific Working Group on Antimicrobial Resistance.

American Journal of Hospital Pharmacy . 1984; 41:132937.


40/76


9. Levy, SB, JP Burke, and CK Wallace. Antibiotic use and antibiotic resistance worldwide.Report of a study sponsored by the Fogarty International Center of the National Institutes ofHealth, 19831986. Review of Infectious Diseases . 1987; 9 Suppl 3:S231-316.

10. Kunin, CM. Resistance to antimicrobial drugsa worldwide calamity. Annals of Internal Medicine . 1993; 118:55761.

11. Institute of Medicine (IOM). The costs of antimicrobial drug resistance in Antimicrobialresistance: issues and options. Workshop report . Washington, DC: IOM. 1998.

12. World Health Organization (WHO). International nonproprietary names (INN) for pharmaceutical substances. Cumulative list No. 9 . Geneva: WHO. 1996.


41/76

Ann ex A. Detail ed I nstructions and Sample Data F orms 3

ANNEX A.

DETAILED INSTRUCTIONS AND SAMPLE DATA FORMS


42/76

H ow to I nvestigate Anti micr obial Dr ug Use in H ospital s36


43/76


Instructions for completing Instrument 1: Basic Information

This form is used for the following indicator:

Indicator 1. Existence of standard treatment guidelines (STGs) and a list of officially

sanctioned antimicrobial drugs in a formulary listData collection summary:

Information for this form is collected from hospital management (e.g., director, chairman ofDTC, finance manager) during interviews at the beginning of the study.

Instructions:

1. Name of Unit: Record the name of hospital and/or service.

2. Data Collector: Record the name of the person collecting the data.

3. Date: Record the date on which the data are collected.

4. Numbered Questions: Answer each question with either Y (yes) or N (no) or thespecific information requested. For question 2, the date of the last meeting of the DTCmust be taken from the minutes of the meeting. If there are no minutes, there was nomeeting. Be sure to ask for and collect copies of the formulary list and standardtreatment guidelines (STGs) or treatment protocol for pneumonia, if these exist.

Note: All spaces should be filled in on this data collection form. Enter N/A if data for a particular item is not available.


44/76


es

no minutes

es

no

no

no

es

4

1,745239

$5,371

$3,100

18

Instrument 1 SAMPLEBasic Information

Name of unit: General Hospital Data collector: John Thomas Date: March 2, 2000

1. Does the hospital have a Drug and Therapeutics Committee? _____________

2. If affirmative, when was the last meeting? ___________ [Review minutes, if any.]

3. Does the hospital have a formulary list or official list of drugs authorized for acquisition bythe hospital? _____________

4. If yes, how many antimicrobial drugs are on the formulary list? _________[Request a copy of the list.]

5. Are all of the drugs on the formulary list identified by generic name (INN)? _____________

6. Does the hospital have standard treatment guidelines (STGs) or protocols for the most prevalent conditions? _____________ For pneumonia? _____________ [Request a copyof the list.]

7. How many treatments have had protocols developed? _____________

8. Does the hospital laboratory routinely perform antimicrobial drug sensitivity tests(antibiograms, cultures)? _____________

9. How many discharges did the hospital have during the last calendar year? _____________

10. How many surgical interventions were performed during the last calendar year? __________

11. Does the hospital have protocols or norms for surgical prophylaxis with antimicrobialdrugs? _____________

12. How much did the hospital spend on drugs last year? _____________

13. How much was budgeted or allotted for drugs by the hospital or Ministry of Health last year? _____________

Instrument 1 is used to collect information for Indicator 1.

no


45/76


Instructions for completing Instrument 2: Form to Record Antimicrobial Drug Treatments

This form is used for the following indicators:

Indicator 5. Percentage of hospitalizations with one or more antimicrobial drugs prescribed

Indicator 6. Average number of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed

Indicator 7. Percentage of antimicrobial drugs prescribed consistent with the hospitalformulary list

Indicator 8. Average cost of antimicrobial drugs prescribed per hospitalization withantimicrobial drugs prescribed

Indicator 9. Average duration of prescribed antimicrobial drug treatment

Indicator 11. Percentage of patients with pneumonia who are prescribed antimicrobial drugsin accordance with standard treatment guidelines

Indicator 12. Percentage of antimicrobial drugs prescribed by generic name

Indicator 13. Percentage of doses of prescribed antimicrobial drugs actually administered

Indicator 14. Average duration of hospital stay of patients who receive antimicrobial drugs

Indicator 15. Number of antimicrobial drug sensitivity tests reported per hospital admissionwith antimicrobials prescribed

Data collection summary:

Information for this form is collected from the hospital clinical histories (CH) sampled. EachCH number is entered in column 1 and the CH is examined to determine whether an AMDwas prescribed. If not, enter N in column 3. If affirmative, enter the name of the AMD incolumn 6 exactly as written in the prescription on the clinical history. If more than oneAMD was prescribed during the hospital stay, enter each one on a new line of the form.Then, for each AMD prescribed, enter the remaining prescription information.

Instructions:

1. Name of Unit: Record the name of hospital and/or service.

2. Data Collector: Record the name of the person collecting the data.

3. Date: Record the date on which the data are collected.


46/76


4. Numbered Columns:

Column 1 Enter the serial number of the clinical history (CH).At the end of the column total the number of clinical records entered.

Column 2 Was the patient diagnosed with pneumonia? Enter Y (yes) or N (no).

At the end of the column total the number of Ys.Column 3 Was an antimicrobial prescribed? Enter Y or N. At the end of the columntotal the number of Ys.

Column 4 Enter the number of days spent in hospital between entry and discharge

Documents

How to Invesitigate Antimicrobial Drug Use in Hospitals