Drug Burden Index and Potentially Inappropriate Medications in Community-Dwelling Older People

Drug Burden Index and PotentiallyInappropriate Medications inCommunity-Dwelling Older PeopleThe Impact of Home Medicines Review

Ronald L. Castelino,1 Sarah N. Hilmer,2,3 Beata V. Bajorek,1,2 Prasad Nishtala1

and Timothy F. Chen1

1 Faculty of Pharmacy, The University of Sydney, Sydney, New South Wales, Australia

2 Departments of Aged Care and Clinical Pharmacology, Royal North Shore Hospital, St Leonards,

Sydney, New South Wales, Australia

3 Sydney Medical School, The University of Sydney, Sydney, New South Wales, Australia

Abstract Background:A significant problem in older people (aged ‡65 years) is the useof potentially inappropriate medications (PIMs), including those with seda-

tive and anticholinergic properties. However, effective intervention strategies

have yet to be identified. The Drug Burden Index (DBI) is an evidence-based

tool that measures a person’s total exposure to medications with sedative and

anticholinergic properties and has been shown to be independently associated

with impairment in cognitive and physical function.

Objective: The main aim of the study was to investigate whether Home

Medicines Review (HMR) services by pharmacists for community-dwelling

older people would lead to an improvement in the use of medications, as

measured by a decrease in the DBI score. The study also aimed to investigate

the (i) distribution of DBI scores and PIMs among older people living in the

community, and (ii) impact of pharmacists’ recommendations on DBI scores

and PIMs.

Methods: A retrospective analysis of medication reviews was performed for

372 community-dwelling older people (aged ‡65 years) who received an

HMR service from the pharmacist. The main outcome measure was the total

DBI score at baseline and post-HMR. The data were also examined to de-

termine the extent of PIM use (2003 Beers’ criteria), and the number and

nature of pharmacists’ recommendations.

Results: Overall, medications contributing to the DBI (i.e. medications with

sedative or anticholinergic properties) and PIMs were identified in 60.5%(n= 225) and 39.8% (n = 148) of the patients, respectively. Following phar-

macist recommendations during the HMR service, medications contributing

to the DBI were identified in 51.6% (n = 192) of the patients. A statistically

ORIGINAL RESEARCH ARTICLEDrugs Aging 2010; 27 (2): 135-1481170-229X/10/0002-0135/$49.95/0

ª 2010 Adis Data Information BV. All rights reserved.

significant reduction in the sum total of DBI scores for all patients was

observed following pharmacists’ recommendations during the HMR service

(206.9 vs 157.3, p < 0.001). Pharmacists’ recommendations also led to a

decrease in the use of PIMs, which were identified in 28.2% (n = 105) of thepatients following the HMR service.

Conclusion:When the DBI is used for evaluation, pharmacists’ recommenda-

tions during HMR services, if acted upon, may effect changes in the pre-

scribing of sedative and anticholinergic medications, thereby substantially

reducing the patient’s drug burden. Future studies should focus on whether

such a decrease may translate into functional improvements. The study also

showed a positive influence of HMR services on the prescribing of PIMs.

Background

Older people often suffer from multiple co-morbidities and are subsequently prescribed mul-tiple medications, which increases the risks of ad-verse drug events (ADEs), drug-drug interactionsand drug-disease interactions.[1-4] Moreover, alter-ed pharmacokinetic and pharmacodynamic para-meters as a result of age-related changes heightenthe risk of ADEs. In addition, few data exist on therelative benefit versus risk of medication use inolder people as they are often not recruited toparticipate in clinical drug trials.[5] Hence, thequality and safety of prescribing in older peopleremains a global healthcare concern and en-deavours to improve the appropriateness of medi-cation selection for older people are warranted.

The prevalence of inappropriate prescribing inolder people has been assessed in many studiesusing various expert consensus statements, inparticular Beers’ criteria.[6-8] The original versionwas published in 1991 and was developed for theelderly nursing home population.[6] The criteriaspecifically included medications or classes ofmedications considered to be inappropriate de-pending on dose and duration of therapy, andthose considered inappropriate regardless ofthese factors.[6] The list was later refined to in-clude new medications and a severity rating, andto apply to all people aged ‡65 years.[7,8] Whilstattention has been drawn to the increasing pre-valence of potentially inappropriate medications(PIMs) in older individuals and the factors asso-

ciated with PIM use,[9] interventions and toolsthat are more inclusive than Beers’ criteria areneeded to reduce the risk of adverse reactions inolder people.[10] In recent years, concerns havebeen raised that Beers’ criteria are becoming re-dundant because of the inclusion of obsoletedrugs.[11] Additionally, there is now a greaterawareness that several other medications with thepotential to harm are not recognized in Beers’ list,whilst at the same time medications that are notavailable in all settings are included in the list andthis limits its universal application in all coun-tries.[12] However, Beers’ criteria remain a valu-able tool for initial screening for inappropriatemedications, although alternative tools have beendeveloped to assist in identifying appropriatedrug therapy.

The Drug Burden Index (DBI) is a tool thatmeasures a person’s total exposure to medica-tions that possess anticholinergic and/or sedativeproperties, using the principles of dose responseand maximal effect.[13] This index takes into ac-count the dose of the anticholinergic or sedativedrug to which the patient is exposed. A higherDBI has been associated with impaired physicalfunction in community-dwelling older people inthe US[13,14] and Australia[15] and predicts func-tional impairment over 5 years.[16] Impairment ofphysical function is an important outcome for olderpeople as it independently predicts nursing homeplacement and mortality.[17] Several studies haveshown that exposure to certain classes of medica-tions, particularly those with anticholinergic and

136 Castelino et al.

ª 2010 Adis Data Information BV. All rights reserved. Drugs Aging 2010; 27 (2)

sedative effects, poses special risks to olderpeople,[18-20] e.g. prolonged sedation and anincreased risk of falls with benzodiazepines[21] oran increased risk of cognitive impairment withanticholinergics.[22] Such medications are oftenprescribed to older patients for various ther-apeutic indications, such as antihistamines forallergy, anticholinergics for urge incontinence,antipsychotics for psychosis or behavioural pro-blems in dementia, antidepressants for depres-sion, and benzodiazepines for anxiety or sleepproblems. The use of these medications is a po-tentially modifiable risk factor, with many studiesdemonstrating that their withdrawal is withoutadverse clinical effects.[23] However, it has notbeen established whether reducing the DBI canimprove function in older people, althoughwithdrawal of medications included in the DBIhas shown promising results on some measuresof mobility in geriatric patients with a historyof falls.[24]

Several studies have addressed the prevalence ofsuboptimal prescribing in older people, emphasiz-ing the magnitude of this public health problemand the need for effective intervention strategies toprevent suboptimal prescribing and drug-relatedmorbidity.[25-28] In Australia, Home MedicinesReview (HMR), a community-based collaborativeservice provided by general practitioners (GPs)and accredited pharmacists, represents a keystrategy for achieving quality use of medicines andis a key component of Australia’s National Medi-cines Policy.[29] Pharmacists who provide anHMRservice have undergone specialized training andpractice to undertake medication management re-views, and are accredited by either the AustralianAssociation of Consultant Pharmacy (AACP) orthe Society of Hospital Pharmacists of Australia.This clinical service is Commonwealth-funded andcomprises the following steps: (i) GP referral of apatient to the patient’s preferred communitypharmacy based on standard criteria, e.g. the pa-tient taking five or more medicines or a medicinewith a narrow therapeutic index; (ii) the pharma-cist conducts an interview in the patient’s hometo ascertain a comprehensive medication profile;(iii) the pharmacist’s written report documentingHMR findings and recommendations is sent to

the GP; and (iv) the GP and patient agree on amedication management plan based on the HMRreport.[30]

Although some clinical evidence for the im-pact of medication review services exists,[2] lim-ited evidence is available regarding interventionsby pharmacists targeting specific medicationswith a high potential to cause harm in olderpeople. The DBI, which correlates well withfunctional measures,[13] may be a useful tool toassess the impact of medication review on medi-cation use.

Study Objectives

The primary purpose of this study was to in-vestigate whether pharmacists’ recommendationsduring HMR service for community-dwellingolder people would lead to an improvement in theuse of medications as measured by a decrease inthe DBI score.

The specific objectives of this study were to:1. Determine the distribution of DBI scores andPIMs in older people living in the community.2. Evaluate the impact of pharmacists’ recom-mendations on the DBI score and PIMs.

Methods

Ethics

Ethical approval for the study was granted bythe Human Ethics Committee of the Universityof Sydney (Sydney, NSW, Australia).

Study Population

This retrospective study involved the collec-tion of a purposive sample of de-identified HMRcases and reports pertaining to 372 community-dwelling older people (aged ‡65 years). Patientswere referred to the HMR service on the basis ofstandard criteria,[30] e.g. taking ‡5 regular medi-cations; taking >12 doses of medication/day; sig-nificant changes made to the medication regimenin the last 3 months; taking a medication with anarrow therapeutic index; and recent (within thelast 4 weeks) discharge from a facility/hospital.

Drug Burden Index and Potentially Inappropriate Medications 137


The HMR cases and reports were collectedfrom the AACP (n = 148) and seven individualservice providers (n = 224) in New South Wales,Australia. The reviews obtained via the AACPrepresented samples of each pharmacist’s reviewssubmitted for the purposes of reaccreditation (theprocess was changed after 2005). We obtained148 of these de-identified medication reviews by112 accredited pharmacists. Pharmacists werealso approached individually to provide access totheir de-identified medication reviews. As notedabove, seven pharmacists agreed to participateand 224 medication reviews were collected.

Data Analysis

The documented patient data included demo-graphics, diagnosed medical conditions, currentmedications, relevant pathology results and pa-tients’ interview notes. The pharmacists’ recom-mendations to the GP and the patient wereextracted from the HMR reports. All infor-mation gathered from the HMR cases andreports was thoroughly reviewed by two studyinvestigators.

The baseline demographic characteristics ofthe patients (including sex, number of medicalconditions, number of medications, and numberof sedative and anticholinergic medications) andDBI scores pre- and post-HMRwere comparablebetween the cases collected from the seven in-dividual service providers (n = 224) and the casesobtained from the AACP (n = 148). An in-dependent t-test was used to analyse numericaldata and the chi-squared (w2) test was used toevaluate categorical data. The only statisticallysignificant difference between the cases collectedfrom the individual service providers and thosecollected from the AACP using the independentsamples t-test was in mean age (SD), which was74.4 (9.1) years vs 78.4 (7.4) years, respectively(p < 0.05).

The primary outcome measure was the changein the patient’s total DBI score from pre- to post-HMR, as interpreted from the pharmacist’s medi-cation recommendations specific to sedative andanticholinergic medications. The secondary out-come measure was the number of recommenda-

tions made by the pharmacists to affect the use ofPIMs according to Beers’ criteria.

Assessment Using the Drug Burden Index (DBI)

The following data were extracted from thede-identified HMR cases and reports in order tocompute the DBI score: classification of whetherthe medication was anticholinergic or sedative;route of administration; and dose and frequencyof administration. As described in the originalstudy on the DBI,[13] medications with anticho-linergic or sedative effects were identified usingMosby’s Drug Consult[31] and the Physicians’ DeskReference,[32] and medications with both anti-cholinergic and sedative effects were classified asanticholinergic. Only exposure to anticholinergicand sedative medications used on a regular basiswas taken into account to compute the DBIscore. Medications that were used on an ‘asneeded’ basis were not taken into account whencomputing the DBI scores because the effect ofintermittent use of sedative and anticholinergicmedication use on function is not known.[13] Thedrug burden attributable to each anticholinergicor sedative medication was calculated usingequation 1:

Drug burden ¼ D

Dþ d(Eq: 1Þ

where D is the daily dose of anticholinergic orsedative medication and d is the minimumefficacious daily dose approved by the US FDA.The total drug burden was calculated as the sumof the drug burden of all anticholinergic orsedative medications that the patient was ex-posed to using a linear additive model.[33] Phar-macists’ recommendations post-HMR were thenassessed to determine if these would change theDBI score.

Assessment Using Beers’ Criteria

PIM use was identified using the 2003 Beers’criteria.[8] The 2003 Beers’ criteria identify PIMuse among older adults both independentlyof diagnoses and taking diagnoses into account.In this study, we used both criteria whereverthe complete information was available fromthe cases and reports. For assessing PIM use



independent of diagnosis, all medications (in-cluding ‘as needed’ medications) were taken intoaccount since these medications should not beused under any circumstances according to Beers’criteria. Pharmacists’ recommendations were as-sessed and compared with Beers’ criteria-basedrecommendations.

Statistical Analysis

The data were processed and analysed usingMicrosoft Excel and SPSS 17.0 for Windows�

statistical software (SPSS Inc., Chicago, IL,USA). The medical conditions from the HMRcases as recorded by the GPs were coded as perthe International Classification of Diseases (10thEdition)[34] [table I]. All medications were codedas per the Anatomical Therapeutic Chemicalclassification system[35] (table II). All significancetests were two-sided and a p-value of <0.05 wasconsidered statistically significant. Median DBIscores before and after the intervention (HMR)

were compared using a two-tailed Wilcoxonsigned-rank test (p < 0.05). Pharmacists’ recom-mendations targeting sedative and anticholi-nergic medications were categorized as those thatdecreased the DBI score, e.g. by ceasing seda-tive/anticholinergic medication or reducing thedose of sedative/anticholinergic medication; in-creased the DBI score, e.g. by adding medicationwith sedative/anticholinergic properties or in-creasing the dose of sedative/anticholinergicmedication; or did not change the DBI score,e.g. by recommending no change in the seda-tive/anticholinergic medication or dose or switch-ing of one sedative/anticholinergic medicationto another. Descriptive statistics were used topresent data on the use of PIMs.

Results

Table III shows the characteristics of thestudy population at baseline. The mean (SD) ageof the study sample was 76.1 (7.8) years and 55%

Table I. Diagnosed medical conditions categorized according to the International Classification of Diseases (10th Edition) [ICD-10][34] in the

study sample (n = 372)

Code ICD-10 No. (%)

A00–B99 Certain infectious and parasitic diseases 5 (0.2)

C00–D48 Neoplasms 34 (1.5)

D50–D89 Diseases of the blood and blood-forming organs 24 (1.1)

E00–E90 Endocrine, nutritional and metabolic diseases (e.g. diabetes mellitus, dyslipidaemia, thyroid disorders) 325 (14.4)

F00–F99 Mental and behavioural disorders (e.g. depression, anxiety, schizophrenia, psychosis, dementia) 89 (3.9)

G00–G99 Diseases of the nervous system (e.g. insomnia, peripheral neuropathy, movement disorders,

Alzheimer’s disease)

203 (9.0)

H00–H59 Diseases of the eye and adnexa (e.g. glaucoma, retinopathy, dry eyes) 111 (4.9)

H60–H95 Diseases of the ear and mastoid process 21 (0.9)

I00–I99 Diseases of the circulatory system (e.g. hypertension, coronary artery disease, heart failure, atrial fibrillation) 564 (24.9)

J00–J99 Diseases of the respiratory system (e.g. asthma, chronic obstructive airway disease) 117 (5.2)

K00–K93 Diseases of the digestive system (e.g. reflux disease, constipation, peptic ulcer disease) 226 (10.0)

L00–L99 Diseases of the skin and subcutaneous tissue 32 (1.4)

M00–M99 Diseases of the musculoskeletal system and connective tissue (e.g. osteoporosis, osteoarthritis,

pain disorders, gout)

361 (16.0)

N00–N99 Diseases of the genitourinary system (e.g. renal impairment, hyperplasia of prostate) 94 (4.2)

Q00–Q99 Congenital malformations, deformations and chromosomal abnormalities 7 (0.3)

R00–R99 Symptoms, signs and abnormal clinical and laboratory findings that are not elsewhere classified 4 (0.2)

S00–T98 Injury, poisoning and certain other consequences of external causes 17 (0.8)

V01–Y98 External causes of morbidity and mortality 3 (0.1)

Z00–Z99 Factors influencing health status and contact with health services 25 (1.1)

Total 2262 (100)



were females. The mean (SD) number of diag-nosed chronic medical conditions and regularprescription medications per patient were 6.0(3.0) and 8.7 (3.0), respectively. Table I shows thedistribution of all diagnosed medical conditionsacross the study sample. The distribution of allmedications across the study sample is presentedin table II. Only 15.6% (n = 58) of the patientswere prescribed five or fewer regular prescriptionmedications, while the majority of the patients,i.e. 68.5% (n = 255), were prescribed between sixand ten regular prescription medications. High-level polypharmacy (more than ten regular pre-scription medications) was identified in 15.9%(n = 59) of the patients.

DBI Scores

Table III shows the prevalence of sedative andanticholinergic medications contributing to theDBI at baseline. These were distributed amongst60.5% (n = 225) of the study population with 113patients (30.4%) receiving 1, 78 (21.0%) receiving2 and 34 (9.1%) receiving ‡3 medications con-tributing to the DBI. In total, 280 sedative med-ications were prescribed at baseline. These weredistributed amongst 188 (50.5%) patients, ofwhom 125 (66.5%) were receiving 1, 44 (23.4%)

were receiving 2 and 19 (10.1%) were receiv-ing ‡3 sedative medications. Overall, 110 anti-cholinergic medications were prescribed at base-line and were distributed amongst 95 (25.5%)patients, of whom 81 (85.3%) were receiving 1, 13(13.7%) were receiving 2 and 1 (1.1%) wasreceiving 3 anticholinergic medications. The pre-valence of the top five sedative and anti-cholinergic medications contributing to the DBI,which together accounted for nearly half of thetotal medications contributing to the DBI, andpharmacists’ recommendations to improve theuse of these medications, are summarized intable IV. The complete list of medications as-signed to sedative and anticholinergic groups,recommended minimum efficacious daily dose(d) and number of patients exposed to each drugare available on request from the authors.

Table V presents the median DBI scores be-fore and after the HMR service. The median DBIscores were significantly lower (i.e. significantlyless sedative and anticholinergic burden) post-HMR (Wilcoxon signed-rank test, p< 0.001). Themedian DBI exposure was reduced from 0.50 atbaseline to 0.22 post-HMR. Overall, followingpharmacists’ recommendations during the HMRservice, medications contributing to the DBIwere identified in 51.6% (n = 192) of the patients.

Table II. Medications categorized according to Anatomical Therapeutic Chemical (ATC)[35] groups in the study sample (n = 372)

Code ATC group No. (%)

A Alimentary tract and metabolism (e.g. drugs for acid-related disorders, functional bowel disorders,

diabetes mellitus; vitamins)

762 (19.1)

B Blood and blood-forming organs (e.g. antithrombotic agents, antianaemic preparations) 347 (8.7)

C Cardiovascular system (e.g. antihypertensive medications, lipid-modifying agents, diuretics) 1151 (28.8)

D Dermatologicals 97 (2.4)

G Genito-urinary system and sex hormones 69 (1.7)

H Systemic hormonal preparations 52 (1.3)

J Anti-infectives for systemic use 39 (1.0)

L Antineoplastic and immunomodulating agents 19 (0.5)

M Musculoskeletal system (e.g. anti-inflammatory products, antigout preparations, drugs for treatment

of bone disease)

270 (6.8)

N Nervous system (e.g. analgesics, psycholeptics, psychoanaleptics) 680 (17.0)

R Respiratory system (e.g. drugs for obstructive airway diseases, antihistamines for systemic use) 263 (6.6)

S Sensory organs 129 (3.2)

V Various 121 (3.0)

Total 3999 (100)



Figure 1 shows the distribution of DBI scorespre- and post-HMR service, based on theassumption that all the recommendations madeby the pharmacists were implemented by theGPs. As observed in figure 1, the number of pa-tients in the lowest categories of DBI scores (i.e.0 and >0–0.5) increased after the pharmacists’intervention. In all the other categories (i.e. pa-tients with a DBI score >0.5) the number of pa-tients in each category of DBI decreased.

Number and Types of Recommendations

Pharmacists’ recommendations to improvethe quality use of medications are summarized intable VI. Ceasing the sedative or anticholinergicmedication was the most frequently recommendedaction, pertaining to 29.3% (n= 109) of the pa-tients. The largest contributor to this recommenda-tion was cessation of medications with sedativeproperties. Overall, pharmacists’ recommenda-tions led to a decrease in the total DBI score in

26.6% (n= 99) of the patients. In 2.4% (n= 9) of thepatients, pharmacists’ recommendations led to anincrease in theDBI score post-HMR service. A fewexamples of pharmacists’ recommendations thatled to a change in the DBI score are summarized inthe Appendix.

Potentially Inappropriate Medications

Of the 372 patients, 148 (39.8%) were pre-scribed one or more PIMs. Of these 148 patients,113 (30.4%) had 1 PIM, 24 (6.5%) had 2 PIMsand 11 (3.0%) had ‡3 PIMs. The five most pre-valent PIMs included long-acting benzodiaze-pines such as diazepam (37; 21.8%), amiodarone(19; 11.2%), amitriptyline (17; 10.0%), (dextro)propoxyphene (12; 7.1%) and bisacodyl(9; 5.3%). Pharmacists’ recommendations basedon Beers’ criteria were reported for 86 (23.1%)patients. Cessation or reduction in dose was re-commended for 69 (18.5%) patients. Pharmacists’recommendations to improve the use of medica-tions based on Beers’ criteria are summarized intable VI. Overall, pharmacists’ recommendationsled to a decrease in the use of PIMs, which wereidentified in 105 (28.2%) patients.

Overall, 94 medications included in the DBIoverlapped with PIM use identified by Beers’criteria. Figure 2 shows the overlap of the medi-cations included in the DBI and identified byBeers’ criteria and the number of recommenda-tions made.

Discussion

This is the first study in which the DBI hasbeen used as a tool to specifically assess the im-pact of pharmacists’ recommendations duringmedication reviews in community-dwelling olderpeople. The results of this study indicate thatpharmacists can play an important role in im-proving the quality of prescribing in community-dwelling older people. Overall, with use of theDBI as an evidence-based tool, pharmacists’recommendations would lead to a significantdecrease in the DBI score post-HMR service. TheDBI has been shown to be a useful measure forevaluating the degree of exposure to medications

Table III. Baseline characteristics of the study sample (n= 372)

Characteristics Value

Age [mean (SD); y] 76.1 (7.8)

Sex (% female) 55.0

Chronic medical conditions per patient

[mean (SD)]

6.0 (3.0)

Medications per patient [mean (SD)] 10.7 (3.9)

regular prescription medications [mean (SD)] 8.7 (3.0)

‘as needed’ medications [mean (SD)] 2.0 (1.8)

DBI

Patients with DBI calculated [no. (%)] 225 (60.5)

DBI medications prescribeda (no. [mean (SD)]) 390 [1.05 (1.1)]

Sedative medications prescribed

(no. [mean (SD)])

280 [0.75 (1.0)]

Anticholinergic medications prescribed

(no. [mean (SD)])

110 [0.29 (0.5)]

PIMs

Patients with PIMs [no. (%)] 148 (39.8)

PIMs prescribed (no.) 196

PIMs independent of diagnosisb [no. (%)] 170 (86.7)

PIMs dependent on diagnosisc [no. (%)] 26 (13.3)

a ‘As needed’ medications excluded.

b ‘As needed’ medications included for Beers’ criteria-based

analysis.

c Includes only those PIMs for which complete information on

diagnosis was available.

DBI=Drug Burden Index; PIMs= potentially inappropriate medications.



with sedative and anticholinergic effects andlikely functional implications of an older person’smedications. Previous observational studies havedemonstrated an association between increasingDBI and impaired function after controlling forconfounding by indication as far as possible.[13,15]

Each additional unit of drug burden was asso-ciated with a negative effect on physical functionas measured by the Health ABC score and wassimilar to that of three additional physical co-morbidities. Each additional unit of drug burdenalso had a negative effect on the digit symbolsubstitution test of similar magnitude to that offour additional physical co-morbidities.[13]

In the current study, pharmacists made re-commendations to change the prescribing of medi-cations with sedative and anticholinergic effectsthat resulted in a 0.22 decrease in the median DBIscore, corresponding to approximately half of aminimum efficacious dose of a sedative or anti-cholinergic medication. The magnitude of this re-duction was similar to that seen in the previous

study on medications review in residential carehomes.[36] It is noteworthy that the majority ofpharmacists’ recommendations were to withdrawor reduce the usage of benzodiazepines, which isconsistent with the findings of a recent meta-analysis that showed that pharmacist-ledmedicationreviews were effective in reducing psychotropicdrug use in aged care homes.[37] Previous studieshave shown that the use of benzodiazepines isa well recognized cause of ADEs, includingfalls.[38,39] Studies have also associated higher dosesof benzodiazepines with cognitive decline, whereaslower doses were not associated with this adverseeffect.[40,41] Pharmacists’ proposed recommenda-tions also included withdrawal of or reduction inthe dose ofmedications with potent anticholinergiceffects and other CNS medications. Reduced ex-posure to these medications may have importanthealth outcomes in the older person. Studies haveconsistently associated anticholinergic adverseeffects with cognitive impairment.[20,42-44] A fewstudies have also attributed decline in cognitive

Table V. Drug Burden Index (DBI) scores before and after pharmacist intervention (n =372)

Parameter Baseline (pre-HMR) After pharmacist intervention (post-HMR)

Sum of all DBI scores 206.86 157.26 (49.6a)

Median DBI scores (IQR) 0.50 (0–0.89) 0.22* (0–0.66)

a Net change in the DBI after pharmacist intervention.

HMR =Home Medicines Review; IQR = interquartile range; * p< 0.001 by Wilcoxon signed-rank test.

Table IV. Sedative and anticholinergic medications most frequently responsible for the Drug Burden Index and the number of pharmacists’

recommendations made to change, cease or reduce the dose

Classification of medications Medications [no. (%)] Recommendations [no. (%)]

With sedative properties (n= 280)

Temazepam 38 (13.6) 18 (47.4)

Codeine 27 (9.6) 3 (11.1)

Sertraline 26 (9.3) 10 (38.5)

Diazepam 22 (7.9) 12 (54.5)

Tramadol 21 (7.5) 6 (28.6)

With anticholinergic properties (n =110)

Amitriptyline 17 (15.5) 9 (52.9)

Prazosin 13 (11.8) 5 (38.5)

Olanzapine 9 (8.2) 0 (0)

Oxybutynin 8 (7.3) 4 (50.0)

Prochlorperazine 8 (7.3) 4 (50.0)



function to use of antidepressants and anti-psychotics because of their anticholinergic andsedative properties.[45]

However, the implications of these results forthe quality of care of older patients should beinterpreted with caution. For example, treatmentof pain and psychiatric symptoms is an importantpart of the clinical care of a patient and thereforethe use of relevant pharmacotherapy, which mayhave a higher anticholinergic or sedative burden,is warranted. Avoidance of appropriate medica-tion because of the risk of cognitive impairmentmay lead to undertreatment. Interestingly, phar-macists’ recommendations for suboptimal painmanagement included addition of strong analge-sics such as opioids, which would result in an in-crease in the DBI score. This suggests that nosingle tool can replace individual clinical assess-ment by the healthcare team, given that a benefit-risk assessment is fundamental to decision mak-ing. The DBI tool may assist in the ‘risk’ aspectof this assessment but needs to be balancedagainst the respective ‘benefits’ of any treatmentfor the individual patient. However, usingthe lowest possible combined doses of CNSmedi-

cations, particularly when treating concurrentpain and psychiatric conditions, should beconsidered as minimizing the risk of cognitivedecline.

In the present study, we also determined theprevalence of PIM use in the study sample andthe impact of pharmacists’ recommendations.Although, the prevalence of PIM use of 39.8% inour study was comparable with the prevalence(25–40%) reported in previous studies of com-munity-dwelling older patients,[9,46,47] concernsover the suitability of Beers’ criteria outside theUS[48,49] were reinforced by the findings ofthis study. Many of the medications listed inBeers’ criteria are rarely or not used in Australia(e.g. guanadrel, trimethobenzamide). Con-versely, the DBI, which includes all drugs withanticholinergic and sedative properties as perthe registered or listed prescribing information,may be easily applied across different healthsystems. However, although exposure to anti-cholinergic and sedative drugs measured usingthe DBI is similar in community-dwelling olderpeople in the US[13] and Australia,[15] and in-creasing DBI is associated with functional

00

20

40

60

80

100

120

140

147

180

6371

86

72

46

33

158 9

5 5 3 1

160

180

200

0

>0−0

.5

>0.5

−1.0

>1.0

−1.5

>1.5

−2.0

>2.0

−2.5

>2.5

−3.0

>3.0

−3.5

DBI score

Num

ber

of p

atie

nts

Pre-HMRPost-HMR

Fig. 1. Impact of Home Medicines Review (HMR) services delivered by pharmacists on the number of patients in each Drug Burden Index(DBI) scoring category.



impairment in both countries, the tool has notbeen assessed elsewhere. Furthermore, thedesignation of certain drugs as inappropriateby Beers’ criteria is debatable, e.g. avoidanceof amiodarone in older people regardless ofthe diagnosis.[8] Amiodarone is an effectiveagent in the management of specific cardiacarrhythmias and its use may be entirely appro-priate in certain cases. This is reflected in thecurrent study, where the use of amiodaronewas high, yet none of the recommendations weredirected towards its cessation (with the excep-

tion of advice for therapeutic monitoring). Inaddition, studies evaluating the possible nega-tive outcomes related to the use of PIMs basedon Beers’ criteria have shown mixed results.Some studies have associated medications listedin Beers’ criteria with worse physical function-ing and pain,[50] higher provider and facilitycosts, and higher numbers of inpatient, out-patient and emergency department visits.[51]

Studies have also reported that the use of thesemedications had no association with revisitsto emergency departments, hospitalization,

PIMs n = 102Recommendations n = 34

DBI medications n = 296Recommendations n = 84

DBI and PIMs n = 94Recommendations n = 53a

Fig. 2. Comparison between medications contributing to Drug Burden Index (DBI) and potentially inappropriate medications (PIMs). Onlyrecommendations to reduce or cease medications are considered. a In six cases recommendations could not be classified as Beers’ criteria-based (e.g. a recommendation to reduce the dose of amitriptyline reduced the DBI but remained an inappropriate medication according toBeers’ criteria).

Table VI. Nature of pharmacists’ recommendations during the Home Medicines Review (HMR) service

Parameter Baseline

(pre-HMR)

Cease medication

[no. (%)]

Reduce dose

[no. (%)]

Increase dose

[no. (%)]

Add medication

[no. (%)]

No change

[no. (%)]

Medications contributing to DBI

sedatives 280 79 (21.2) 8 (2.2) 4 (1.1) 11 (3.0) 4a (1.1)

anticholinergics 110 30 (8.1) 4 (1.1) 3 (0.8) 1 (0.3) 5 (1.3)

PIMs prescribed according to

Beers’ criteria

196 66 (17.7)

11b (3.0)

3 (0.8) 0 5 (1.3) 1 (0.3)

a A sedative changed to another sedative (n =2), a sedative changed to an anticholinergic (n = 1) and an anticholinergic changed to a

sedative (n =1).

b Number of regular PIMs recommended to be converted to ‘as needed’.

DBI =Drug Burden Index; PIMs = potentially inappropriate medications.



functional decline or mortality, and had mini-mal impact on healthcare utilization.[50,52,53]

Although Beers’ criteria have limitations,namely being explicit in nature, it must be ac-knowledged that they may serve an importantrole in initial screening for PIMs as part ofa broader intervention to address potentiallyproblematic medications in older people. Incontrast, the DBI as a tool is more implicit innature although it takes into account only drugswith sedative and anticholinergic properties.Many of the recommendations made by thepharmacists may not reduce PIM use but mayreduce the drug burden. For example, a re-commendation to reduce the dose of amitripty-line for the treatment of neuropathic pain maynot be appropriate according to Beers’ criteriaalthough it reduces the DBI score. We believethat this approach may be more flexible than thatof Beers’ criteria and could encourage cliniciansto consider medications as a possible cause ofsymptoms in older people, thereby avoiding po-tentially problematic medications. However, theDBI is clearly not a substitute for clinical assess-ment and judgement and, hence, like other tools,it should not be used in isolation but rather inconjunction with clinical judgement temperedwith an understanding of the patient’s wishes andclinical context.

Our study does have some limitations thatshould be considered. First, we do not know towhat extent the recommendations proposed bythe pharmacists were implemented. However,previous studies of clinical pharmacist-led in-terventions have demonstrated that 41–96% ofthe recommendations are accepted and mostare acted upon by prescribers.[2,54-57] Second, inour study the DBI was used retrospectively toanalyse pharmacists’ recommendations specificto sedative and anticholinergic medications.Pharmacists in the study neither received anytraining nor were advised to particularly targetthese medications and hence the retrospectiveobservational approach avoided the possibilityof the study influencing the actual practicepharmacists employ as a part of their HMRservice. Finally, the use of the DBI as a toolhas some potential limitations. As a pharmaco-

logical model it does not take into account thedifference in pharmacokinetic and pharmaco-dynamic profiles among the drugs and the sub-jects and assumes a linear dose relationshipbetween the drug classes.[13] Despite these lim-itations, the DBI could be used by all disciplinesinvolved in the care of older people to assess theextent of the drug burden created by sedativesand anticholinergics, with the intention ofminimizing the risk of ADEs. Whether theDBI, when used as an intervention, can sig-nificantly improve function or reduce drug-related morbidity or mortality remains to beseen and future studies addressing these issuesare warranted.

Conclusions

When the DBI is used for evaluation, phar-macists’ recommendations during HMR services,if acted upon, may effect changes in the pre-scribing of sedative and anticholinergic medica-tions, thereby substantially reducing the patient’sdrug burden. Future studies should focus onwhether such a decrease may translate into func-tional improvements. The study also showed apositive influence of HMR services on the pre-scribing of PIMs.

Acknowledgements

No sources of funding were used to assist in the prepara-tion of this study. The author acknowledges the AustralianAssociation of Consultant Pharmacy and the seven accreditedpharmacists for providing access to de-identified medicationreviews used in this study. Associate Professor Sarah Hilmerholds an international patent for the Drug Burden Index. Theother authors have no conflicts of interest that are directlyrelevant to the content of this study.

Appendix

Examples of pharmacists’ recommendationsto the general practitioner that impacted on theDrug Burden Index score and potentially in-appropriate medications are shown in table AI.



References1. Castelino RL, Bajorek BV, Chen TF. Targeting suboptimal

prescribing in the elderly: a review of the impact of phar-

macy services. Ann Pharmacother 2009; 43 (6): 1096-106

2. Zermansky AG, Alldred DP, Petty DR, et al. Clinical med-

ication review by a pharmacist of elderly people living in

care homes: randomised controlled trial. Age Ageing 2006;

35 (6): 586-91

3. Goldberg RM,Mabee J, Chan L, et al. Drug-drug and drug-

disease interactions in the ED: analysis of a high-risk

population. Am J Emerg Med 1996; 14 (5): 447-50

4. Hilmer SN, Gnjidic D. The effects of polypharmacy in older

adults. Clin Pharmacol Ther 2009; 85 (1): 86-8

5. Beyth RJ, Shorr RI. Principles of drug therapy in older pa-tients: rational drug prescribing. Clin Geriatr Med 2002; 18(3): 577-92

6. Beers MH, Ouslander JG, Rollingher I, et al. Explicitcriteria for determining inappropriate medication use innursing home residents: UCLA Division of GeriatricMedicine. Arch Intern Med 1991; 151 (9): 1825-32

7. Beers MH. Explicit criteria for determining potentiallyinappropriate medication use by the elderly: an update.Arch Intern Med 1997; 157 (14): 1531-6

8. Fick DM, Cooper JW, Wade WE, et al. Updating the Beerscriteria for potentially inappropriate medication use inolder adults: results of a US consensus panel of experts.Arch Intern Med 2003; 163 (22): 2716-24

Table AI. Examples of pharmacists’ recommendations to the general practitioner that impacted on the Drug Burden Index (DBI) score and

potentially inappropriate medications (PIMs)

Medication with

sedative/anticholinergicproperties or PIM

Pharmacists’ recommendations DBI change PIM change

Amiodarone Consider monitoring thyroid function NA No change

Amitriptyline Replace amitriptyline with mirtazapine as it is less anticholinergic No change Yes

Reduce dose of amitriptyline because of its anticholinergic adverse effects Decrease No change

Cyproheptadine Cease use as it is more likely to cause sedation, disturbed coordination

and confusion

Decrease Yes

Diazepam Cease use of diazepam because of possible risk of exacerbating

confusion and cognitive impairment

Decrease Yes

Consider using on ‘as needed’ basis because of risk of cognitive

impairment, falls and depression

Decrease No change

Digoxin Reduce dose of digoxin from 250 mg to 125 mg because of low creatinine

clearance

NA Yes

Doxepin Consider SSRI because doxepin is associated with anticholinergic

adverse effects

Decrease Yes

Nitrazepam Reduce the dose of nitrazepam to half in view of morning sedation Decrease NA

Nitrofurantoin Cease use as patient complains of ‘pins and needles’ in hands and feet

and nitrofurantoin can be associated with peripheral neuropathy

NA Yes

Oxazepam Limit the use of oxazepam as regular use may increase the risk of falls

and cognitive impairment. Consider using on ‘as needed’ basis

Decrease NAa

Oxybutynin Cease oxybutynin in view of its potent anticholinergic adverse effects Decrease Yes

Prazosin Indication for prazosin is unclear – cease use because of patient’s urinary

symptoms and the fact that a-adrenoceptor antagonists are associated

with urge incontinence

Decrease NA

(Dextro)propoxyphene Cease (dextro)propoxyphene as it is not a potent analgesic, is poorly

tolerated by older people and is no better than paracetamol

(acetaminophen) for osteoarthritic pain

Decrease Yes

Tramadol Cease use of tramadol as it may potentially contribute to the serotonin

syndrome when used with citalopram

Decrease NA

Change tramadol for oxycodone as tramadol has similar adverse

effects to strong opioids but is a less potent analgesic

No change NA

Temazepam Reduce use of temazepam because of loss of effect over time and

dependence

Decrease NAb

a Only doses exceeding 60mg are considered inappropriate according to Beers’ criteria.

b Only doses exceeding 15mg are considered inappropriate according to Beers’ criteria.

NA= not applicable; SSRI = selective serotonin reuptake inhibitor.



9. Zhan C, Sangl J, Bierman AS, et al. Potentially inappropri-ate medication use in the community-dwelling elderly:findings from the 1996 Medical Expenditure Panel Survey.JAMA 2001; 286 (2): 2823-9

10. Page 2nd RL, Ruscin JM. The risk of adverse drug eventsand hospital-related morbidity and mortality among olderadults with potentially inappropriate medication use. Am JGeriatr Pharmacother 2006; 4 (4): 297-305

11. Barry PJ, O’Keefe N, O’Connor KA, et al. Inappropriateprescribing in the elderly: a comparison of the Beers criteriaand the improved prescribing in the elderly tool (IPET) inacutely ill elderly hospitalized patients. J Clin Pharm Ther2006; 31 (6): 617-26

12. Basger BJ, Chen TF, Moles RJ. Inappropriate medicationuse and prescribing indicators in elderly Australians: de-velopment of a prescribing indicators tool. Drugs Aging2008; 25 (9): 777-93

13. Hilmer SN, Mager DE, Simonsick EM, et al. A drug burdenindex to define the functional burden of medications inolder people. Arch Intern Med 2007; 167 (8): 781-7

14. Cao YJ, Mager DE, Simonsick EM, et al. Physical andcognitive performance and burden of anticholinergics,sedatives, and ACE inhibitors in older women. Clin Phar-macol Ther 2008; 83 (3): 422-9

15. Gnjidic DCR, Le Couteur DG, Abernethy DR, et al. DrugBurden Index and physical function in older Australianmen. Br J Clin Pharmacol 2009; 68 (1): 97-105

16. Hilmer SN, Mager DE, Simonsick EM, et al. Drug burdenindex score and functional decline in older people. Am JMed 2009; 122 (12): 1142-9.e1-2

17. Guralnik JM, Ferrucci L, Simonsick EM, et al. Lower-extremity function in persons over the age of 70 years as apredictor of subsequent disability. N Engl J Med 1995;332 (9): 556-61

18. Gray SL, Penninx BWJH, BloughDK, et al. Benzodiazepineuse and physical performance in community-dwelling olderwomen. J Am Geriatr Soc 2003; 51 (11): 1563-70

19. Mulsant BH, Pollock BG, Kirshner M, et al. Serum antic-holinergic activity in a community-based sample of olderadults: relationship with cognitive performance. Arch GenPsychiatry 2003; 60 (2): 198-203

20. Landi F, Russo A, Liperoti R, et al. Anticholinergic drugsand physical function among frail elderly population. ClinPharmacol Ther 2007; 81 (2): 235-41

21. Tinetti ME. Clinical practice: preventing falls in elderlypersons. N Engl J Med 2003; 348 (1): 42-9

22. Tune LE. Anticholinergic effects of medication in elderlypatients. J Clin Psychiatry 2001; 62 Suppl. 2: 11-4

23. Iyer S, Naganathan V, McLachlan AJ, et al. Medicationwithdrawal trials in people aged 65 years and older: a sys-tematic review. Drugs Aging 2008; 25 (12): 1021-31

24. van der Velde N, Stricker BHC, Pols HAP, et al. Withdrawalof fall-risk-increasing drugs in older persons: effect onmobility test outcomes. Drugs Aging 2007; 24 (8): 691-9

25. Hanlon JT, Artz MB, Pieper CF, et al. Inappropriate medi-cation use among frail elderly inpatients. Ann Pharmaco-ther 2004; 389-14

26. Hanlon JT, Schmader KE, Ruby CM, et al. Suboptimalprescribing in older inpatients and outpatients. J AmGeriatr Soc 2001; 49: 200-9

27. Hanlon JT, Lindblad CI, Gray SL. Can clinical pharmacyservices have a positive impact on drug-related problemsand health outcomes in community-based older adults?Am J Geriatr Pharmacother 2004; 2: 3-13

28. Goulding MR. Inappropriate medication prescribing forelderly ambulatory care patients. Arch Intern Med 2004;164 (3): 305-12

29. Weekes LM, Mackson JM, Fitzgerald M, et al. NationalPrescribing Service: creating an implementation arm fornational medicines policy. Br J Clin Pharmacol 2005; 59(1): 112-6

30. Department of Health and Ageing, Australian Government.Medication Management Reviews [online]. Available fromURL: http://www.health.gov.au/internet/main/publishing.nsf/Content/health-epc-dmmr-answers.htm#6 [Accessed 2009Jun 17]

31. Nissen D, editor. Mosby’s drug consult. St Louis (MO):Mosby Inc., 2004

32. Duplay D, editor. Physicians’ desk reference. 58th ed.Mont-vale (NJ): Thomson PDR, 2004

33. Tallarida RJ. Drug synergism: its detection and applica-tions. J Pharmacol Exp Ther 2001; 298 (3): 865-72

34. World Health Organization. International statistical classi-fication of disease and related health problems. 10th rev.Geneva: World Health Organization, 1992 [online]. Avail-able from URL: http://apps.who.int/classifications/apps/icd/icd10online/ [Accessed 2009 Dec 1]

35. WHO Collaborating Centre for Drug Statistics Methodol-ogy. ATC classification index with DDDs. Oslo: WorldHealth Organization, 2009 [online]. Available from URL:http://www.whocc.no/atcddd/ [Accessed 2009 Dec 1]

36. Nishtala PS, McLachlan AJ, Hannan PJ, et al. Drug BurdenIndex in aged care homes: impact of residential medicationmanagement reviews. Drugs Aging 2009; 26 (8): 677-86

37. Nishtala PS, McLachlan AJ, Bell JS, et al. Psychotropicprescribing in long-term care facilities: impact of medica-tion reviews and educational interventions. Am J GeriatrPsychiatry 2008; 16 (8): 621-32

38. Pariente A, Dartigues J-F, Benichou J, et al. Benzodiaze-pines and injurious falls in community dwelling elders.Drugs Aging 2008; 25 (1): 61-70

39. Hanlon JT, Horner RD, Schmader KE, et al. Benzodiaze-pine use and cognitive function among community-dwell-ing elderly. Clin Pharmacol Ther 1998; 64 (6): 684-92

40. Pomara N, Deptula D, Medel M, et al. Effects of diazepamon recall memory: relationship to aging, dose, and durationof treatment. Psychopharmacol Bull 1989; 25 (1): 144-8

41. Merchant RA, Li B, Yap KB. Use of drugs with anti-cholinergic effects and cognitive impairment in commu-nity-living older persons. Age Ageing 2008; 38: 105-8

42. Ancelin ML, Artero S, Portet F, et al. Non-degenerativemild cognitive impairment in elderly people and use ofanticholinergic drugs: longitudinal cohort study. BMJ2006; 332 (7539): 455-9

43. Bottiggi KA, Salazar JC, Yu L, et al. Long-term cognitiveimpact of anticholinergic medications in older adults. Am JGeriatr Psychiatry 2006; 14 (11): 980-4

44. Moore AR, O’Keeffe ST. Drug-induced cognitive impair-ment in the elderly. Drugs Aging 1999; 15 (1): 15-28



45. Fick DM, Mion LC, Beers MH, et al. Health outcomes as-sociated with potentially inappropriate medication use inolder adults. Res Nurs Health 2008; 31 (1): 42-51

46. Stuart B, Kamal-Bahl S, Briesacher B, et al. Trends in the pre-scription of inappropriate drugs for the elderly between 1995and 1999. Am J Geriatr Pharmacother 2003; 1 (2): 61-74

47. Laroche M-L, Charmes J-P, Merle L. Potentially in-appropriate medications in the elderly: a French consensuspanel list. Eur J Clinical Pharmacol 2007; 63 (8): 725-31

48. Fialova D, Topinkova E, Gambassi G, et al. Potentiallyinappropriate medication use among elderly home carepatients in Europe. JAMA 2005; 293 (11): 1348-58

49. Chin MH, Wang LC, Jin L, et al. Appropriateness of med-ication selection for older persons in an urban academicemergency department. Acad Emerg Med 1999; 6 (12):1232-42

50. Fick DM, Waller JL, Maclean JR, et al. Potentially in-appropriate medication use in a Medicare managed carepopulation: association with higher costs and utilization.J Manag Care Pharm 2001; 7 (Sep-Oct): 407-13

51. Hanlon JT, Fillenbaum GG, Kuchibhatla M, et al. Impactof inappropriate drug use on mortality and functionalstatus in representative community dwelling elders. MedCare 2002; 40 (2): 166-76

52. Fillenbaum GG, Hanlon JT, Landerman LR, et al. Impactof inappropriate drug use on health services utilization

among representative older community-dwelling residents.Am J Geriatr Pharmacother 2004; 2 (2): 92-101

53. Sorensen L, Stokes JA, Purdie DM, et al. Medicationreviews in the community: results of a randomized,controlled effectiveness trial. Br J Clin Pharmacol 2004; 58:648-64

54. Patel HR, Pruchnicki MC, Hall LE. Assessment forchronic kidney disease service in high-risk patients atcommunity health clinics. Ann Pharmacotherapy 2005;39 (1): 22-7

55. Krska J, Cromarty JA, Arris F, et al. Pharmacist-led medi-cation review in patients over 65: a randomized, controlledtrial in primary care. Age Ageing 2001; 30: 205-11

56. Sellors J, Kaczorowski J, Sellors C, et al. A randomizedcontrolled trial of a pharmacist consultation programfor family physicians and their elderly patients. CMAJ2003; 16: 917-22

57. O’Dell KM, Kucukarslan SN. Impact of the clinical phar-macist on readmission in patients with acute coronarysyndrome. Ann Pharmacother 2005; 39 (9): 1423-7

Correspondence: Ronald L. Castelino, Faculty of Pharmacy,Pharmacy Building A15, The University of Sydney, Sydney,NSW 2006, Australia.E-mail: [email protected]



Documents

Drug Burden Index and Potentially Inappropriate Medications in Community-Dwelling Older People