using published criteria to develop a list of potentially inappropriate medications for elderly...

ORIGINAL REPORT

Using published criteria to develop a list of potentially inappropriatemedications for elderly patients in Taiwan†

Chirn-Bin Chang1, Shu-Yu Yang2, Hsiu-Yun Lai3, Ru-Shu Wu2, Hsing-Cheng Liu4, Hsiu-Ying Hsu2,Shinn-Jang Hwang5,6{ and Ding-Cheng Chan7,8*

1Department of Internal Medicine, Changhua Christian Hospital, Changhua City, Changhua County, Taiwan2Department of Pharmacy, Taipei City Hospital, Taipei, Taiwan3Department of Family Medicine, Yuanshan Veteran Hospital, Yuanshan Township, Yilan County, Taiwan4Taipei City Psychiatry Center, Taipei City Hospital, Taipei, Taiwan5Department of Family Medicine, Taipei Veteran General Hospital, Taipei, Taiwan6National Yang Ming University School of Medicine, Taipei, Taiwan7Department of Geriatrics and Gerontology, National Taiwan University, Taipei, Taiwan8Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan

ABSTRACTBackground Explicit criteria for potentially inappropriate medications (PIMs) developed from other regions were often difficult to applyto a specific territory without significant modifications.Purpose To describe a process of developing a country-specific explicit PIM criteria from quality review of several published PIM criteria,followed by consensus among regional experts in Taiwan.Methods After a review of the literature, we selected seven sets of published PIM criteria. Medications/medication classes listed in at leastthree of the seven sets of criteria were selected as preliminary core PIMs. We asked a group of 21 experts from various specialties to rate howappropriate they found inclusion of each medication/medication class in final PIM criteria after two rounds of modified Delphi methods.Results Table 1 of the instrument included 24 medication/medication classes to be generally avoided in older adults irrespective of co-morbidities, and Table 2 included 12 chronic conditions with six medication/medication classes that patients with these conditions shouldavoid. The Taiwan criteria contained only half the number of statements that were included in the Beers criteria (36 vs 68 statements) butdetected nearly 70–75% as many PIMs in older patients with polypharmacy in a secondary data analysis. Features included straight-forward statement arrangements, suggestions of alternatives, and clear definitions of long-acting benzodiazepine and anticholinergic drugsfor Table 1 PIMs.Conclusion A user-friendly instrument was developed to detect PIMs for Taiwanese older adults. Further prospective studies are needed tovalidate its use in clinical and research settings. Copyright © 2012 John Wiley & Sons, Ltd.

key words—potentially inappropriate medications; explicit criteria; aged; pharmacoepidemiology

Received 4 November 2011; Revised 13 February 2012; Accepted 11 March 2012

INTRODUCTION

Pharmacodynamic and pharmacokinetic propertiesof drugs change with the patients’ aging process asdemonstrated by delayed drug metabolism and elimina-tion.1 Older adults also have a higher number of chronic

conditions than do younger people and usually takemultiple medications. Therefore, the incidence ofadverse drug reactions (ADRs) increases significantlywith age.2 Some ADRs are preventable if the medica-tions with a high risk/benefit ratio are avoided. Drugswith a high risk/benefit ratio, uncertain therapeuticeffects, or with safer alternatives for older adults aretermed potentially inappropriate medications (PIMs).3

The Beers criteria, first published in 1991,4 weredeveloped to discourage the use of PIMs in older adults.This latest version of the Beers criteria, published in

2003,5 has been widely used in many countries as an

*Correspondence to: D.-C. Chan, Department of Geriatrics and Gerontology,National Taiwan University, Hospital, No. 7 Zhongshan South Road, Taipei,Taiwan, 100. E–mail: [email protected]; [email protected]†This manuscript has not been published or presented elsewhere in part or inentirety and is not under consideration by another journal.{Alternative corresponding author, equal contribution.

Copyright © 2012 John Wiley & Sons, Ltd.

pharmacoepidemiology and drug safety (2012)Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/pds.3274

Table 1. 2010 Criteria for potentially inappropriate medication use in persons aged≥ 65 years (PIM-TAIWAN): independent of diagnoses or conditions

Classification/medicationsATCcodes Concern Alternative therapy*

Averageof Likertscales

Drugs for peptic ulcerand gastro-esophagealreflux disease (GERD)

A02B

Cimetidine A02BA01 Central nervous system (CNS) adverse effects includingconfusion.

PPI or other H2 receptor antagonists 3.7

AntispasmodicsAntispasmodics-gastrointestinal

A03B Gastrointestinal (GI) antispasmodic drugs are highlyanticholinergic and have uncertain effectiveness. Thesedrugs should be avoided (especially for long-term use).

Use only when strongly indicated,avoid long-term use

Belladonna alkaloid A03BA/A03BB

4.2

Hyoscyamine products A03BA03 4.2Antispasmodics-urologicals

G04B

Oxybutynin G04BD04 This antispasmodic drug is poorly tolerated by elderlypatients because these cause anticholinergic adverseeffects, sedation, and weakness. Additionally, theireffectiveness at doses tolerated by elderly patients isquestionable.

Suggest behavioral therapy, use onlywhen strongly indicated

4.2

Oral hypoglycemic drugs A10Chlorpropamide A10BB02 It has a prolonged half-life in elderly patients and could

cause prolonged hypoglycemia. Additionally, it is the onlyoral hypoglycemic drug that causes inappropriateantidiuretic hormone hypersecretion (SIADH).

Short to intermediate acting secondgeneration sulfonylurea or other oralanti-hyperglycemic drugs

4.8

Antithrombotic agents B01ATiclopidine B01AC05 Has been shown to be no better than aspirin in preventing

clotting and may be considerably more toxic.Aspirin or clopidogrel 3.7

Digitalis glycosides C01AADigoxin> 0.125mg/d C01AA05 Decreased renal clearance may lead to increased risk of

toxic effects. (should not exceed> 0.125 mg/d except forarrhythmia)

- 3.7

Anti-hypertensives C02AClonidine C02AC01 Potential for orthostatic hypotension and central nervous

system (CNS) adverse effectsAnother class of antihypertensivedrugs

3.8

Methyldopa C02AB01 May cause bradycardia and exacerbate depression inelderly patients.

4.0

Reserpine C02AA02 May induce depression, impotence, sedation, andorthostatic hypotension

4.2

Anti-inflammatory andanti-rheumatic product,non-steroids

M01A 1. Acetaminophen;2. Other NSAID (avoid long-termuse)

Ketorolac M01AB15 Immediate and long-term use should be avoided in olderpersons, a significant number have asymptomaticgastrointestinal (GI) pathologic conditions.

4.2

Indometacin(indomethacin)

M01AB01 Of all available non-steroidal anti-inflammatory drugs, thisdrug produces the most central nervous system (CNS)adverse effects.

4.2

Phenylbutazone M01AA01 Severe hematological adverse effects. 4.3Piroxicam M01AC01 Have the potential to produce confusion, gastrointestinal

(GI) bleeding, renal failure, high blood pressure, and heartfailure.

4.0

Muscle relaxants M03 Most muscle relaxants are poorly tolerated by elderlypatients because these cause anticholinergic adverseeffects, sedation, and weakness. Additionally, theireffectiveness at doses tolerated by elderly patients isquestionable.

Use only when strongly indicated,avoid long-term use

Baclofen M03BX01 3.9Carisoprodol M03BA02 3.9Chlormezanone M03BB02 3.6

Chlorphenesin carbamate M03 3.7Chlorzoxazone M03BB03 3.9

(Continues)

c.-b. chang et al.

Copyright © 2012 John Wiley & Sons, Ltd. Pharmacoepidemiology and Drug Safety, (2012)DOI: 10.1002/pds

Table 1. (Continued)



Cyclobenzaprine M03BX08 4.0Dantrolene M03CA01 4.0Methocarbamol M03BA03 3.9Orphenadrine (citrate) M03BC01 3.8Phenprobamate M03BA01 3.9Pridinol M03BX03 3.9Tizanidine M03BX02 3.9Tolperisone M03BX04 3.8Analgesics-opioids N02APethidine (Merperidine) N02AB02 Not an effective oral analgesic in doses commonly used.

May cause confusion and has many disadvantages to othernarcotic drugs.

Other narcotic drugs 4.2

Propoxyphene N02AC04 Offers few analgesic advantages over acetaminophen, yethas the adverse effects of other narcotic drugs.

4.1

Antipsychotics N05AAntipsychotics-lowpotency

Anticholinergic effects, second choice of drugs. Atypical antipsychotics, exceptclozapine

Chlorpromazine N05AA01 4.0Levomepromazine N05AA02 3.9Loxapine N05AH01 3.7Thioridazine N05AC02 3.9Antipsychotics-clozapineClozapine N05AH02 May lower seizure thresholds and increase risk of

agranulocytosisOther atypical antipsychotics 3.9

Anxiolytics, hypnoticsand sedatives

N03/N05

Anxiolytics, hypnotics andsedatives-long-actingbenzodiazepine

These drugs have a long half-life in elderly patients (oftenseveral days), producing prolonged sedation andincreasing the risk of falls and fractures.

1. Use short-acting benzodiazepineonly when strongly indicated2. Lowest effective dose, avoid long-termuse

Clonazepam N03AE01 3.8Chlordiazepoxide N05BA02 4.1Clobazam N05BA09 4.0Diazepam N05BA01 4.2Fludiazepam N05BA17 3.8Flunitrazepam N05CD03 4.1Flurazepam N05CD01 4.1Nitrazepam N05CD02 4.0Nordazepam N05BA16 4.0Oxazolam N05BA 3.6Potassium clorazepate N05BA05 3.9Anxiolytics-carbamateMeprobamate N05BC01 This is a highly addictive and sedating anxiolytic. Those

using meprobamate for prolonged periods may becomeaddicted and may need to be withdrawn slowly.


4.2

Hypnotics and sedatives-benzodiazepineTriazolam> 0.25mg/day N05CD05 Risk of falls and fracture. ≦ 0.25mg/day, or other short-acting

benzodiazepine4.0

Hypnotics and sedatives-barbiturate

N05CA Are highly addictive and cause more adverse effects thanmost sedative or hypnotic drugs in elderly patients.


Amobarbital N05CA02 4.7Pentobarbital N05CA01 4.7Secobarbital N05CA06 4.8Antidepressants N06ATricyclic antidepressants Because of its strong anticholinergic and sedation

properties, tricyclic antidepressants are rarely the choicefor elderly patients.

Antidepressants other than TCA

Amitriptyline N06AA09 4.3Clomipramine N06AA04 4.4

(Continues)

potentially inappropriate medication—taiwan criteria


indicator of geriatric healthcare quality.6,7 However,opinions differ as to the inappropriateness of certainmedications in the Beers criteria, such as amiodarone,doxazosin, amitriptyline, and fluoxetine.8,9 Additionally,the Beers criteria do not clearly list anticholinergicmedications with different grade of anticholinergicactivity. The statements contained in the Beers criteriaare not arranged in any particular order, making itdifficult to apply them in routine clinical practice.In addition, some of the medications in the Beerscriteria are not widely available outside the USA. Toovercome these shortcomings, other explicit criteriahave been developed in Canada,10,11 France,8 Ireland,12

Norway,13 Thailand,14 Italy,15 and Germany.16

When we used six published criteria to analyzesecondary data from a Taiwanese cohort of older adultswith multiple prescriptions, we found that only 50–89%of the medications listed in the published criteria are

available in Taiwan.17 PIM prevalence in Taiwanmay be underestimated if significant portions of listedmedications are not available.18–21 Therefore, criteriadeveloped for other countries are probably not suitablefor Taiwan without significant modifications.Establishing a new set of criteria may be time

consuming, particularly during the literature reviewprocess, because there are a limited number of clinicaltrials of drug in older adults. The authors of the 2011German criteria performed a comprehensive literaturesearch but added only four new medications that werenot mentioned in previously published criteria.16

This concept is supported by a recent systemic reviewof PIM criteria that most of them were also developedbased mainly on existing criteria with various degreesof literature reviews.22 Because many published PIMcriteria are available as references, core PIMs could beidentified from medications simultaneously listed in

Table 1. (Continued)



Dosulepin N06AA16 4.3Doxepin N06AA12 4.2Imipramine N06AA02 3.6Melitracen N06AA14 4.1Antihistamine R06First-generationantihistamine

These antihistamines may have potent anticholinergicproperties.

Second-generation antihistamine

Alimemazine R06AD01 4.1Azatadine R06AX09 4.0Brompheniramine R06AB01 4.0Buclizine R06AE01 4.1Carbinoxamine R06AA08 4.1Chlorcyclizine R06AE04 4.0Chlorpheniramine R06AB04 4.2Chlorphenoxamine R06AA06 4.0Clemastine R06AA04 4.1Cyclizine R06AE03 4.0Cyproheptadine R06AX02 3.9Dexchlorpheniramine R06AB02 4.0Diphenhydramine R06AA02 4.1Diphenylpyraline R06AA07 3.9Doxylamine R06AA09 4.0Homochlorcyclizine HCL(Homoginin)

R06AE 4.0

Ketotifen R06AX17 3.6Mebhydrolin R06AX15 4.0Meclozine (meclizine) R06AE05 3.9Mepyramine R06AC01 4.0Mequitazine R06AD07 4.0Oxatomide R06AE06 3.8Phenindamine R06AX04 4.0Pheniramine R06AB05 3.8Promethazine R06AD02 4.1Tripelennamine R06AC04 4.1Triprolidine R06AX07 3.8

*Alternative therapy was suggested by four pharmacists not from the consensus of 21 experts.

c.-b. chang et al.


several PIM criteria. These core PIMs are accepted bymost experts worldwide and are probably applicable tothe Taiwanese population. This approach is likely toresult in the inclusion of fewer statements in the corePIM list than were included in the reference PIMcriteria. In general, criteria with fewer statements areeasier to use. However, if the core PIM list includesmany highly prevalent PIMs, prevalence estimatesmade using the core PIM list will not be much lowerthan those made using the reference PIM criteria. Oneexample showed that inpatient PIM prevalence wasestimated at 22% using the improved prescribing inthe elderly tool (IPET, 14 statements), whereas theprevalence was estimated at 34% when the Beerscriteria (68 statements) were used.23

The aim of this study is to describe a step-by-stepprocess of developing the user-friendly PIM-Taiwancriteria from several published PIM criteria based onmedications available in Taiwan.

METHODS

Establishment of preliminary list

As a first step for developing PIM criteria in Taiwan, weidentified 13 sets of explicit criteria published in Englishlanguage from a search of literature in the PubMed

database published between 1991 and 2009. Afteromitting duplicated criteria, we published a paper de-scribing seven sets of criteria in 2010.24 These sets werethe 2003 version of the Beers criteria,5 the McLeod,10

the Rancourt,11 the Laroche,8 the Screen Tool of OlderPerson’s Prescription (STOPP),12 the Winit Watjana,14

and the Norwegian General Practice criteria.13

A research team was convened including threegeriatricians, one psychiatrist and four clinical pharma-cists to create a list of preliminary PIMs from theseseven published criteria. Any medication/medicationclass appeared in at least three of these seven sets ofcriteria were selected as candidate PIMs. Criteria createdby Italian authors 15 were published after we created thepreliminary PIM lists in 2010. The panelists decided notto include the Italian criteria because more than 90% ofthe medications were also appeared in the in the 2003Beers criteria.The availability of medications in Taiwan listed in

the candidate PIM criteria was confirmed using themedication database from Bureau of National HealthInsurance. We identified generic names and theAnatomical Therapeutic Chemical (ATC) classificationsystem code for each medication. When a medicationclass, rather than an individual medication, was listedas a PIM, we identified all available medications

Table 2. 2010 Criteria for potentially inappropriate medication use in persons≥ 65 years of age in Taiwan (PIM-TAIWAN): considering diagnoses orconditions

Disease or condition Medication classification ConcernAverage ofLikert scales

Blood clotting disorders orreceiving anticoagulant therapy

Anti-inflammatory and anti-rheumatic products,non-steroids (NSAID), acetylsalicylic acid

May prolong clotting time or inhibit plateletaggregation, resulting in an increased potential forbleeding

3.8

Chronic constipation Anticholinergic May exacerbate constipation 4.2Chronic kidney disease Anti-inflammatory and anti-rheumatic products,

non-steroids (NSAID)May worsen renal failure, may cause salt andwater retention

4.3

Chronic obstructive pulmonarydisease

Benzodiazepines May induce respiratory depression. 3.7

Non-selective Beta blocking agents May exacerbate or cause respiratory depression 4.3Cognitive impairment,dementia

Anticholinergics, Concern due to CNS-altering effects 4.1

Benzodiazepines 4.0Glaucoma Anticholinergics May aggravate glaucoma 4.6Heart failure Anti-inflammatory and anti-rheumatic products,

non-steroids (NSAID)May cause salt and water retention and worsen heartfailure

3.6

Peptic ulcer disease Anti-inflammatory and anti-rheumatic products,non-steroids (NSAID), (COX II inhibitorswere excluded)

May exacerbate existing ulcers or produce new/additional ulcers

4.5

Acetylsalicylic acid 4.4Sleep apnea syndrome Benzodiazepines May suppress breathing 4.1Syncope or falls Benzodiazepines May produce ataxia, impaired psychomotor

function, syncope, and additional falls4.3

Urinary incontinence Alpha-adrenoreceptor antagonists May produce polyuria and worsening of incontinence 4.0Benzodiazepines May induce drowsiness, dizziness, and urine

incontinence3.6

Urinary retention Anticholinergics May decrease urinary flow, leading to urinary retention 4.6



belonging to that class. Most published criteria consid-ered anticholinergics and long-acting benzodiazepinesas PIMs. However, clear definitions for these twomedication classes were often lacking. We definedbenzodiazepines with half-life greater than 24 hours aslong-acting. We included medications with anticholin-ergic activities grades 2 to 3 as PIMs, based on a listpublished in a previous study.25

The preliminary list of PIM-Taiwan contained twoparts. Part 1 included medication/medication classes(86 individual drugs) that should be generally avoidedin older adults, and part 2 contained medication classes(109 individual drugs) that should be avoided by olderadults with any of 12 medical conditions.

Delphi method

A group of 21 experts from different specialties (geriatri-cians, neurologists, psychiatrists, cardiologists, pulmo-nologists, gastroenterologist, urologists, and clinicalpharmacists) were invited to develop consensus, usingthe Delphi methods, from the preliminary PIM-Taiwanlist.26 Briefly, this method is a communicationwith structured questionnaire between a panel of expertsto reach a consensus decision.26 Most invited non-geriatrician experts have extensive experiences on caringfor older adults and certain degrees of exposures togeriatric prescribing principles. The experts were askedto rank each statement for the degree of inappropriate-ness using a five-point Likert scale that ranged from fivepoints (strongly agree) to one point (strongly disagree).The first round of the questionnaires was scored by eachexpert around October 2010. PIMs with mean Likertscale ≧ 3.5 were scored again around December 2010.For the second round, selective and non-selective betablockers were rated separately for patients with chronicobstructive pulmonary diseases, based on suggestionsfrom several experts. Experts also suggested separatingaspirin from non-steroid anti-inflammatory drugs(NSAIDs) for its unique protective effect on thecardiovascular system. The final PIM list was generatedfrom medications/medication classes with mean Likertscale scores ≧ 3.5 after the second round rating. Fourclinical pharmacists suggested alternative therapy forpart 1 PIMs to be finalized by the research team.However, these alternative therapies did not go throughthe Delphi process for approval from the expert panel.

Applying PIM-Taiwan criteria for secondarydataset analysis

Using a method described earlier,17 we analyzedsecondary data from the Medication Safety ReviewClinic-Taiwan study (MSRC-Taiwan) to compare the

ability of the PIM-Taiwan criteria in detect major PIMsagainst the Beers Criteria. Briefly, the MSRC-Taiwanstudy enrolled 193 older adults prescribed with a meanof nine chronic medications. Comprehensive geriatricassessments and medication review were performed toidentify and solve drug-related problems. In the baselinedataset, detailed information on chronic conditionsand medications was available to allow us to applyboth the Beers criteria and PIM-Taiwan criteria todetect PIMs.17 Comparisons were made between theprevalence estimates derived from both instruments.Time for the manual PIM identification from the datasetusing both sets of criteria was estimated. ATC codeswere not available from the dataset.

RESULTS

The final PIM-Taiwan criteria included 36 statements.Table 1 listed 24 medication/medication classes(83 individual drugs) generally to be avoided in olderadults irrespective of co-morbidities. Table 2 listed 12chronic conditions with six medication/medicationclasses (99 individual drugs) to be avoided by indivi-duals with any of 12 chronic conditions. There is sig-nificant overlap of individual drugs in both tables,particularly anticholinergic and benzodiazepines.Overall, 151 different drugs were included in the tables.The ATC codes, medication/medication classes, andreasons for inclusion are provided in each table. Alterna-tive therapies for PIMs are included in Table 1. Whenno alternative medication could be suggested, werecommended that these medications should be usedonly with strong indications and for short periods.Whenan entire medication class was considered as PIMs, allmedications available in the class were listed in Table 1.Except for aspirin, Table 2 only included medicationclasses that should be avoided in patients with certainchronic conditions. The individual medicationscontained in each class are listed in Appendix 1, ratherthan in Table 2. The average of Likert scales rangesfrom 3.6 to 4.8 in the final lists of PIM-Taiwan.The MSRC-Taiwan study was conducted at the

National Taiwan University Hospital and its affiliatedBei-Hu Branch. The population include in this studywas aged 76.2 in averages, and 53.4% was malegender. Although all medications in the PIM-Taiwancriteria are available in Taiwanese market, only 47%of them are included in the NTUH drug formulary(Table 3). Sixty-nine percent of the Beers drugs wereavailable in Taiwan and 43% in the NTUH drugformulary. The PIM-Taiwan criteria included halfthe number of statements as in the Beers criteria

c.-b. chang et al.


(36 vs 68). Prevalence of PIM at person level was 40%with the PIM-Taiwan criteria and 55% with the Beerscriteria. Prevalence of PIM at drug level was 7.3% withthe PIM-Taiwan criteria and 10.3% with the Beerscriteria. Overall, the new instrument detected 73% (40/55)PIMs at personal level and 71% (7.3/10.3) at drug levelas with the Beers criteria. Time for manual identificationPIMs in the 193 patients’ dataset were estimated. Onaverage, it takes less time to complete the secondarydata analysis using the PIM-Taiwan versus Beerscriteria (71.5 vs 133.5minutes) (Table 3).

DISCUSSION

This is the first PIM criteria set established in a Chinese-speaking region. Medications in the PIM-Taiwancriteria were selected from the most common drugsincluded in published PIM criteria from three differentregions including the North America, Europe, and Asia.This increased the generalizability of the PIM-Taiwanfor international comparison studies. Our method ofdeveloping PIM criteria is easily replicable by otherresearch groups.To make the instrument user-friendly, we adapted

the following strategies: (i) the core-PIM conceptsignificantly reduces the number of statements whilepreserving the ability to detect important PIMs; (ii) thearrangements of the PIMs in Table 1 is straight-forward,based on pharmacological properties of medications;(iii) clear definitions of long-acting benzodiazepineand anticholinergic medications with moderate or highanticholinergic activity are included; (iv) suggestionsfor alternative therapies for Table 1 PIMs areprovided, decreasing the time required to find substitutemedications when PIMs are identified; (v) the reasonfor inclusion of each PIM is provided, making theinstrument self-explanatory; and (vi) the ATC code foreach medication is provided to help pharmacistsidentify medications at their institutions. We performeda secondary dataset analysis showing that this newlydeveloped instrument can save significant administration

time with preserved 70–75% PIM detection ratecompared with the Beers criteria in older adults withpolypharmacy. The majority of PIMs that were notidentified in PIM-Taiwan criteria were theophylline,amiodarone, bisacodyl, dipyridamole, and calciumchannel blockers used among older adults withchronic constipation. These medications were consid-ered reasonable prescriptions for certain conditions inour local clinic practice. Also, their average Likertscales were low during our two-round Delphi method.We considered that the gap of 27% identification ratewas reasonable.The experts who participated in the establishment of

most published PIM criteria were general practitioners,geriatricians, psychiatrists, and clinical pharmacists.24

In this study, we invited many experts from otherspecialties such as cardiology, neurology, and urologyto reflect the multidisciplinary perspectives.With a multidisciplinary approach, the rating of some

PIMs in the preliminary list can be more heterogeneous,particularly for medications commonly prescribed incertain specialties. In Table 1, dipyridamole was ratedlower than expected because of its potential role insecondary prevention in coronary artery, cerebrovascular,and chronic kidney diseases.27 Furthermore, theophyl-line and aminophylline are still often prescribed to treatchronic obstructive pulmonary disease in Taiwan.Because of their low cost and ease of use in older adults,many experts were reluctant to include them as PIMs. InTable 2, aspirin prescribed for patients with chronickidney disease was not considered as inappropriateas well.27

Recently, Hamilton et al. 28 showed in a Europeanprospective study that prescription of PIMs fromthe STOPP criteria to hospitalized older adults wereassociated with ADRs, but PIMs from the Beers criteriawere not. This suggests that country-specific PIMcriteria are necessary to enhance the quality ofpharmacotherapy during daily practices. In a recentpublication, Dr Budnitz and his colleagues demon-strated that side effects of four medications (antiplate-lets, warfarin, insulin, and oral hypoglycemic agents)

Table 3. Comparison of basic characteristic between the Beers criteria and the PIM-Taiwan criteria

CriteriaYear of

establishment

Numberof

statements

Number ofspecific drugnames listed

N (%) ofavailabledrugs inTaiwan

N (%) ofavailabledrugs atNTUH

Prevalence (%) ofPIM (person level

analysis)

Prevalence (%)of PIM (druglevel analysis)

Estimated time(minutes) for

identification PIM

Beers 2003 68 103 71/103(69) 44/103(43) 106/193(55) 177/1713(10.3) 133.5Taiwan 2011 36 151 151/151(100) 71/151(47) 78/193 (40) 125/1713(7.3) 71.5

PIM, potentially inappropriate medication.NTUH, National Taiwan University Hospital.



were major causes of older patients’ hospitalization.29

From established medical evidence, warfarin andantiplatelet drugs were major medications for strokeprevention in patients with atrial fibrillation andtreatment of cardiovascular diseases. Similarly, insulinand hypoglycemic agents were essential for hyperglyce-mic control. No medication is with better benefit/riskratio than these medications currently. Therefore, thesemedications are not defined as PIMs, and they werenot included in most current established PIM criteria.Although PIMs in this study accounted the minor partsof side effects, it is preventable if clinicians prescribemedications with better benefit/risk ratio.We are currently designing studies to prospectively

validate the PIM-Taiwan criteria in both inpatient andoutpatient settings using a computer-assisted prescribingalert system. Previous studies30 have shown thatcomputer alert systems were able to reduce PIMprescribing in daily practice. Authors suggested thatconcerns for prescribing PIMs and alternative therapiesshould be implanted in the alert system to increasephysician compliance with the alert messages. ThePIM-Taiwan criteria include all of the features necessaryfor integration with a computer alert system.The disadvantage of this tool was that PIMs were

derived from existing criteria. New findings from recentclinical trials were not reviewed. We recognized thislimitation, and new PIMs can be added to the nextversion. Also, a set of criteria was published by Germanauthor in 2011,16 after we finalized the PIM-Taiwancriteria. The panelists decided to include medicationslisted in the German criteria when in future versions ofthe PIM-Taiwan criteria. We were only able to validatethis tool to a specific population with relatively morechronic medications compared with the general olderpopulation. The power was relatively low, and furtherstudies are needed to determine the performance thistool in general older population. Our criteria only listedcommission errors. Omission errors are also animportant issue that should be addressed in futurecriteria.In conclusion, a set of country-specific PIM-Taiwan

criteria was developed trough a systemic methodeasily replicable by other groups. The instrument isuser-friendly and performed relatively well in a popula-tion of older patients with polypharmacy against theBeers Criteria. Further studies are needed to validateits use in clinical and research settings.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

KEY POINTS• Potentially inappropriate medication (PIM) criteriadeveloped from other countries are probablynot suitable to be applied to a different region(e.g., Taiwan) without significant modifications.However, the process for developing a new set ofcriteria maybe time consuming.

• This study described a step-by-step process todevelop a new set of criteria from core PIMsidentified from seven existing sets of criteria anddrugs available in local markets with suggestedalternative therapies.

• This newly developed PIM-Taiwan criteriacontained only half the number of statements(36 vs 68) that were included in the Beers criteriabut detected nearly 70–75% asmany PIMs in olderpatients with polypharmacy in a secondary dataanalysis with significant shorter administrationtime (71.5 vs 133.5 minutes)

ACKNOWLEDGEMENTS

The authors thank for Dr Wen-Chen Ouyang (JiananMental Hospital, Taiwan), Dr Cheng-Sheng Chen(Kaohsiung Medical University Chung-Ho MemorialHospital, Taiwan), Dr Pei-Ning Wang (Taipei VeteransGeneral Hospital, Taiwan), Dr Chaur-Jong Hu (TaipeiMedical University Shuang Ho Hospital, Taiwan),Dr Jen-Hau Chen (National Taiwan University Hospital,Taiwan), Dr Chih-Hsueh Lin (China Medical UniversityHospital, taiwan), Dr Chin-Ying Chen (National TaiwanUniversity Hospital, Taiwan), Dr Feng-Hwa Lu(National Cheng Kung University Hospital, Taiwan),Dr Chia-Ming Chang (National Cheng Kung UniversityHospital, Taiwan), Dr Chuen-Den Tseng (NationalTaiwan University Hospital, Taiwan), Dr Fu-TienChiang (National Taiwan University Hospital, Taiwan),Dr Ming-Shyan Huang (Kaohsiung Medical UniversityChung-Ho Memorial Hospital, Taiwan), Dr Pei-MingYang (National Taiwan University Hospital, Taiwan),Dr Jin-De Chen (National Taiwan University Hospital,Bei-Hu Branch, Taiwan), Dr Diahn-Warng Perng(Taipei Veterans General Hospital, Taiwan), and DrChang-Chi Chang (Taipei City Hospital, ZhongxiaoBranch, Taiwan) for the participation in modified Delphiprocesses.The authors also thank the following pharmacists:

Hsiang-Wen Lin (China Medical University Hospital,Taiwan), Chen-Chen Chu (National Taiwan UniversityHospital, Taiwan), Ting-Ying Wang (Tri-ServiceGeneral Hospital, Taiwan), Wan-Ying Huang (Tai-LinPharmacy, Taiwan), Yi-Ping Hsiang (Kaohsiung Chang

c.-b. chang et al.


Gung Memorial Hospital, Taiwan), and Kuei-Ju Cheng(Wan Fang Hospital, Taiwan) for the participationin modified Delphi processes and complement ofalternative suggestions. This study is support by aninternal research grant from the Taipei City Hospital.

REFERENCES

1. Merle L, Laroche ML, Dantoine T, Charmes JP. Predicting and preventingadverse drug reactions in the very old. Drugs Aging 2005; 22: 375–92.

2. Cresswell KM, Fernando B, McKinstry B, Sheikh A. Adverse drug events in theelderly. Br Med Bull 2007; 83: 259–74.

3. O’Mahony D, Gallagher PF. Inappropriate prescribing in the older population:need for new criteria. Age Ageing 2008; 37: 138–41.

4. Beers MH, Ouslander JG, Rollingher I, Reuben DB, Brooks J, Beck JC.Explicit criteria for determining inappropriate medication use in nursing homeresidents. UCLA Division of Geriatric Medicine. Arch Intern Med 1991; 151:1825–32.

5. Fick DM, Cooper JW, Wade WE, Waller JL, Maclean JR, Beers MH. Updatingthe Beers criteria for potentially inappropriate medication use in olderadults: results of a US consensus panel of experts. Arch Intern Med 2003; 163:2716–24.

6. Wenger NS, Shekelle PG. Assessing care of vulnerable elders: ACOVE projectoverview. Ann Intern Med 2001; 135: 642–6.

7. Pugh MJ, Hanlon JT, Zeber JE, Bierman A, Cornell J, Berlowitz DR. Assessingpotentially inappropriate prescribing in the elderly Veterans Affairs populationusing the HEDIS 2006 quality measure. J Manag Care Pharm 2006; 12: 537–45.

8. Laroche ML, Charmes JP, Merle L. Potentially inappropriate medications in theelderly: a French consensus panel list. Eur J Clin Pharmacol 2007; 63: 725–31.

9. O’Mahony D, Gallagher P, Ryan C, Byrne S, Hamilton H, Barry P, O’ConnorM, Kennedy J. STOPP & START criteria: A new approach to detectingpotentially inappropriate prescribing in old age. Eur Geriatr Med 2010; 1: 7.

10. McLeod PJ, Huang AR, Tamblyn RM, Gayton DC. Defining inappropriatepractices in prescribing for elderly people: a national consensus panel. CMAJ1997; 156: 385–91.

11. Rancourt C, Moisan J, Baillargeon L, Verreault R, Laurin D, Gregoire JP.Potentially inappropriate prescriptions for older patients in long-term care.BMC Geriatr 2004; 4: 9.

12. Gallagher P, Ryan C, Byrne S, Kennedy J, O’Mahony D. STOPP (ScreeningTool of Older Person’s Prescriptions) and START (Screening Tool to AlertDoctors to Right Treatment). Consensus validation. Int J Clin Pharm Th 2008;46: 72–83.

13. Rognstad S, Brekke M, Fetveit A, Spigset O, Wyller TB, Straand J. TheNorwegian General Practice (NORGEP) criteria for assessing potentiallyinappropriate prescriptions to elderly patients. A modified Delphi study. ScandJ Prim Health Care 2009; 27: 153–9.

14. Winit-Watjana W, Sakulrat P, Kespichayawattana J. Criteria for high-riskmedication use in Thai older patients. Arch Gerontol Geriat 2008; 47: 35–51.

15. Maio V, Del Canale S, Abouzaid S. Using explicit criteria to evaluate the qualityof prescribing in elderly Italian outpatients: a cohort study. J Clin Pharm Ther2010; 35: 219–29.

16. Holt S, Schmiedl S, Thurmann PA. Potentially inappropriate medications in theelderly: the PRISCUS list. Dtsch Arztebl Int 2010; 107: 543–51.

17. Chang CB, Chen JH, Wen CJ, et al. Potentially inappropriate medications ingeriatric outpatients with polypharmacy: application of six sets of publishedexplicit criteria. Br J Clin Pharmacol 2011; 72: 482–9.

18. Chang CM, Liu PY, Yang YH, Yang YC, Wu CF, Lu FH. Use of the Beerscriteria to predict adverse drug reactions among first-visit elderly outpatients.Pharmacotherapy 2005; 25: 831–8.

19. Chen YC, Hwang SJ, Lai HY, et al. Potentially inappropriate medication foremergency department visits by elderly patients in Taiwan. PharmacoepidemiolDrug Saf 2009; 18: 53–61.

20. Lai HY, Hwang SJ, Chen YC, Chen TJ, Lin MH, Chen LK. Prevalence of theprescribing of potentially inappropriate medications at ambulatory care visitsby elderly patients covered by the Taiwanese National Health Insuranceprogram. Clin Ther 2009; 31: 1859–70.

21. Lin HY, Liao CC, Cheng SH, Wang PC, Hsueh YS. Association of potentiallyinappropriate medication use with adverse outcomes in ambulatory elderlypatients with chronic diseases: experience in a Taiwanese medical setting. DrugsAging 2008; 25: 49–59.

22. Dimitrow MS, Airaksinen MS, Kivela SL, Lyles A, Leikola SN. Comparisonof prescribing criteria to evaluate the appropriateness of drug treatment inindividuals aged 65 and older: a systematic review. J Am Geriatr Soc 2011;59: 1521–30.

23. Barry PJ, O’Keefe N, O’Connor KA, O’Mahony D. Inappropriate prescribing inthe elderly: a comparison of the Beers criteria and the improved prescribing inthe elderly tool (IPET) in acutely ill elderly hospitalized patients. J Clin PharmTher 2006; 31: 617–26.

24. Chang CB, Chan DC. Comparison of published explicit criteria for potentiallyinappropriate medications in older adults. Drugs Aging 2010; 27: 947–57.

25. Rudolph JL, Salow MJ, Angelini MC, McGlinchey RE. The anticholinergic riskscale and anticholinergic adverse effects in older persons. Arch Intern Med 2008;168: 508–13.

26. Day J, Bobeva M. A generic tookit for the successful management of Delphistudies. Electron J Business Res Methods 2005; 3: 130–16.

27. Notaro LA, Usman MH, Burke JF, Siddiqui A, Superdock KR, Ezekowitz MD.Secondary prevention in concurrent coronary artery, cerebrovascular, andchronic kidney disease: focus on pharmacological therapy. Cardiovasc Ther2009; 27: 199–215.

28. Hamilton H, Gallagher P, Ryan C, Byrne S, O’Mahony D. PotentiallyInappropriate Medications Defined by STOPP Criteria and the Risk of AdverseDrug Events in Older Hospitalized Patients. Arch Intern Med; 171: 1013–9.

29. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizationsfor adverse drug events in older Americans. N Engl J Med 2011; 365: 2002–12.

30. Hume AL, Quilliam BJ, Goldman R, Eaton C, Lapane KL. Alternatives topotentially inappropriate medications for use in e-prescribing software: triggersand treatment algorithms. BMJ Qual Saf 2011; 20: 875–84.



APPENDIX

MEDICATIONS INCLUDED IN EACH MEDICATION CLASS IN TABLE 2

AnticholinergicsCimetidine A02BA01Atropine products A03BA01Dicyclomine (Dicycloverine) hydrochloride A03AA07Hyoscyamine products A03BA03Loperamide hydrochloride A07DA03Oxybutynin chloride G04BD04Tolterodine tartrate G04BD07Baclofen M03BX01Carisoprodol M03BA02Cyclobenzaprine hydrochloride M03BX08Tizanidine hydrochloride M03BX02Amantadine hydrochloride N04BB01Chlorpromazine hydrochloride N05AA01Fluphenazine hydrochloride N05AB02Perphenazine N05AB03Prochlorperazine maleate N05AB04Thioridazine hydrochloride N05AC02Trifluoperazine hydrochloride N05AB06Clozapine N05AH02Olanzapine N05AH03Hydroxyzine hydrochloride and hydroxyzine pamoate N05BB01Amitriptyline hydrochloride N06AA09Imipramine hydrochloride N06AA02Pseudoephedrine hydrochloride R01BA02Chlorpheniramine maleate R06AB04Cyproheptadine hydrochloride R06AX02Diphenhydramine hydrochloride R06AA02Meclizine hydrochloride R06AE05Promethazine hydrochloride R06AD02Triprolidine hydrochloride R06AX07Cetirizine hydrochloride R06AE07Loratadine R06AX13Non-selective Beta blocking agentsCarvedilol C07AG02Labetalol C07AG01Alprenolol C07AA01Carteolol C07AA15Pindolol C07AA03Propranolol C07AA05Sotalol C07AA07Alpha-adrenoreceptor antagonistsBunazosin NADoxazosin C02CA04Prazosin C02CA01Alfuzosin G04CA01Tamsulosin G04CA02Terazosin G04CA03Anti-inflammatory and anti-rheumatic products, non-steroids (NSAID)Aceclofenac M01AB16Acemetacin M01AB11Alclofenac M01AB06Benzydamine M01AX07Celecoxib M01AH01Diclofenac M01AB05Etodolac M01AB08Etoricoxib M01AH05Fenbufen M01AE05Fenoprofen M01AE04Flufenamic acid M01AG03Flurbiprofen M01AE09Ibuprofen M01AE01Indometacin M01AB01Ketoprofen M01AE03

(Continues)

c.-b. chang et al.


Ketorolac M01AB15Lornoxicam M01AC05Meclofenamic acid M01AG04Mefenamic acid M01AG01Meloxicam M01AC06Mepirizole NANabumetone M01Ax01Naproxen M01AE02Niflumic acid M01AX02Nimesulide M01Ax17Phenylbutazone M01AA01Piroxicam M01AC01Sulindac M01AB02Tenoxicam M01AC02Tiaprofenic acid M01AE11Tiaramide NATolmetin M01AB03BenzodiazepineAlprazolam N05BA12Bromazepam N05BA08Brotizolam N05CD09Chlordiazepoxide N05BA02Clobazam N05BA09Clonazepam N03AE01Diazepam N05BA01Estazolam N05CD04Fludiazepam N05BA17Flunitrazepam N05CD03Flurazepam N05CD01Lorazepam N05BA06Lormetazepam N05CD06Medazepam N05BA03Midazolam N05CD08Nimetazepam NANitrazepam N05CD02Nordazepam N05BA16Oxazepam N05BA04Oxazolam N05BA04Potassium clorazepate N05BA05Triazolam N05CD05



using published criteria to develop a list of potentially inappropriate medications for elderly...

Documents