prevalence of potentially inappropriate medications and risk of adverse clinical outcome in a cohort...

Prevalence of potentially inappropriate medications and risk of adverse clinicaloutcome in a cohort of hospitalized elderly patients: results from the REPOSIStudy

L. Pasina*1 PharmD, C. D. Djade* Stat.D, M. Tettamanti* PhD, C. Franchi* PharmD, F. Salerno†MD, S. Corrao‡MD, A. Marengoni§MD PhD,M. Marcucci¶ MD, P. M. Mannucci** MD, A. Nobili* MD and on behalf of REPOSI Investigators*IRCCS - Istituto di Ricerche Farmacologiche “Mario Negri”, Milano, †Internal Medicine I, Policlinico IRCCS San Donato, University of Milan, Milan,‡Dipartimento Biomedico di Medicina Interna e Specialistica, University of Palermo, Palermo, §Geriatric Unit, Spedali Civili, Department of Medical andSurgery Sciences, University of Brescia, Brescia, ¶Department of Internal Medicine, University of Perugia, Perugia, and **Scientific Director, IRCCS MaggioreHospital Foundation, Milan, Italy

Received 2 October 2013, Accepted 23 April 2014

Keywords: Beers’ criteria, elderly, pharmacoepidemiology

SUMMARY

What is known and objective: Inappropriate prescribing is highlyprevalent for older people and has become a global healthcareconcern because of its association with negative health outcomesincluding ADEs, hospitalization and resource utilization. Beers’criteria are widely utilized for evaluating the appropriateness ofmedications, and an up-to-date version has recently been pub-lished. To assess the prevalence of patients exposed to PIMs athospital discharge according to the 2003 and 2012 versions ofBeers’ criteria and to evaluate the risk of adverse clinical events,re-hospitalization and all-cause mortality at 3-month follow-up.Methods: This cross-sectional study was held in 66 Italianinternal medicine and geriatric wards. The sample included1380 inpatients aged 65 years or older. Prescriptions of PIMwereanalysed at hospital discharge. We considered all patients withcomplete 3-month follow-up.Results and discussion: The prevalence of patients receiving atleast one PIM was 20�1% and 23�5% according to the 2003 and2012 versions of the Beers’ criteria, respectively. The 2012 Beers’criteria identified more patients with at least one PIM than the2003 version, although a high percentage of those patients(72�2%) were also identified by the criteria updated in 2003. Themain difference in the prevalence of patients receiving a PIMaccording to the two versions of Beers’ criteria involvedprescriptions of benzodiazepines for insomnia or agitation,chronic use of non-benzodiazepine hypnotics, prescription ofantipsychotics in people with dementia and oral iron at dosagehigher than 325 mg/day. Prescription of PIMs was not associatedwith a higher risk of adverse clinical events, re-hospitalizationand all-cause mortality at 3-month follow-up in both univariateand multivariate analysis, after adjusting for age, sex and CIRScomorbidity index.What is new and conclusions: This study found no significanteffect of inappropriate drug use according to Beers’ criteria on

health outcomes among older adults 3 month after discharge.Even though these criteria have been suggested as helpful inpromoting appropriate prescribing, reducing drug-relatedadverse events and associated healthcare costs, to date there isno clear evidence that their application can achieve objective andquantifiable improvements in clinical outcomes. A possibleexplanation is that both versions of the Beers’ criteria haveseveral recognized limitations, one of the main ones being therestricted availability of some drugs in Europe or their limitedprescription in everyday clinical practice.

WHAT IS KNOWN AND OBJECTIVE

Polypharmacy is very common among older adults and may oftenbe needed to improve symptoms, disease-related problems andquality of life.1–3 However, it may also be a major risk factor forinappropriate prescribing, poor adherence to therapies, adversedrug events (ADEs) and other adverse health outcomes.4–8 Inap-propriate prescribing is highly prevalent for older people and hasbecome a global healthcare concern because of its association withnegative health outcomes including ADEs, hospitalization andresource utilization.9 Inappropriate medications can be defined, inrelation to older people, as ‘medications or medication classes thatshould generally be avoided in persons 65 years or older becausethey are either ineffective or they pose unnecessarily high risk forolder persons and a safer alternative is available’.10 Explicit criteriafor potentially inappropriate medications (PIMs) have been devel-oped to reduce the use of drugs that involve a substantial risk ofadverse side effects in elderly patients.11 Beers’ criteria are widelyutilized for evaluating the appropriateness of medications, and anup-to-date version has recently been published.12 Despite wide-spread utilization, to date no prospective studies have shown thatclinical application of the Beers’ criteria improves such clinicaloutcomes as ADEs, morbidity, mortality or hospitalization. Arecent Cochrane review concluded that it was not clear whetherinterventions to improve appropriate polypharmacy resulted inclinically significant improvement, although they appeared bene-ficial in terms of reducing inappropriate prescribing and medica-tion-related problems.13

We examined REPOSI, a network of internal medicine andgeriatric wards created to investigate the prevalence and corre-lates with polymorbidity and polypharmacy in elderly hospital

Correspondence: Luca Pasina, Laboratory for Quality Assessment ofGeriatric Therapies and Services, Drug Information Service for theElderly, IRCCS - Istituto di Ricerche Farmacologiche “Mario Negri”,Via Giuseppe La Masa, 19, 20156 Milano, Italy. Tel.: *39 02 39014 579;fax: *39 02 39001916; e-mail: [email protected] stands for Registry of Polytherapies SIMI (Societ�a Italiana diMedicina Interna).

© 2014 John Wiley & Sons Ltd 511

Journal of Clinical Pharmacy and Therapeutics, 2014, 39, 511–515 doi: 10.1111/jcpt.12178

inpatients, to assess the prevalence of those exposed to PIMsaccording to the 2003 and 2012 versions of the Beers’ criteria andthe possible associations with the related risk of adverse clinicalevents, re-admission and all-cause mortality 3 months afterhospital discharge.

METHODS

Data collection

The Registro Politerapie SIMI (REPOSI) is a collaborative, inde-pendent, voluntary initiative of the Italian Society of InternalMedicine (SIMI) and the Istituto di Ricerche Farmacologiche MarioNegri. The registry was set-up in 2008 from a network of internalmedicine and geriatric wards to collect information on an Italiancohort of elderly hospital inpatients with multimorbidities oftenreceiving polypharmacy. The first wave of data collection wasbetween January and December 2008, the second between Januaryand December 2010. To ensure an unselected population of elderlypatients admitted to internal medicine wards, during 4-weekperiod 3 months apart (one in February, one in June, one inSeptember and one in December 2008 and similarly in 2010), thefirst ten patients admitted to the wards participating in the studywere consecutively recruited if they were 65 years old or older.Participation was voluntary, and all patients gave signed informedconsent. Data collection complied fully with Italian laws onpersonal data protection and required no ethical committeeapproval under current legal principles on patient registries. Theattending physicians completed a standardized web-based casereport form, developed for the purpose of the study by SIMI andthe Istituto di Ricerche Farmacologiche Mario Negri, includingdiagnosis at hospital admission, socio-demographic details, drugtreatment at hospital admission, during hospital stay and atdischarge.

In the second wave of REPOSI, it was decided to collectadditional information and follow-up to improve the quality ofdata: main laboratory parameters, comorbidity according to theCumulative Illness Rating Scale (CIRS),14 basic activities of dailyliving according to the Barthel Index (BI),15 cognitive impairmentaccording to the Short Blessed Test,16 depression according to theGeriatric Depression Scale (GDS),17 clinical events during hospitalstay and outcomes. All patients discharged were followed3 months after discharge with one telephone interview performedby an expert clinical monitor (an internist or a geriatrician), whocollected information on new diagnoses, new hospital admission,drug regimens, adverse events and BI. All data were collected inthe standardized web-based case report form used at hospitaladmission. To establish the prevalence of PIMs at hospitaldischarge according to the 2003 and 2012 versions of Beers’criteria, we considered all patients recruited in the second wave ofREPOSI. To evaluate the related risk of adverse clinical events, re-admission and all-cause mortality, we considered only patientswith complete 3-month follow-up, excluding patients who diedbefore discharge, were transferred to another ward or refused the3-month telephone follow-up.

Exposure to Potentially Inappropriate Medications (PIMs)

Explicit criteria for PIMs are usually developed from literaturereviews, expert opinions and consensus techniques and normallycomprise lists of drugs or drug classes and dosages that are knownto have harmful effects in older people. The advantage of explicit

criteria is that they can be applied to prescriptions with little or noclinical judgments. Beers et al. in 1991 published the first set ofexplicit criteria for determining inappropriate prescribing in olderpatients. These criteria based on consensus opinionwere updated in1997, 2003 and 2012 and can be applicable to all patients aged65 years and over irrespective of their place of residence. Toevaluate the association between PIMs exposure at hospitaldischarge and adverse clinical outcome at follow-up, we comparedthe 200318 and 201212 versions of Beers’ criteria, using the section forPIMs in older persons independently of diagnoses or conditions.

Outcomes

To assess the prevalence of PIMs, according to the 2003 and 2012versions of the Beers’ criteria, and the risk of adverse clinicalevents from discharge to follow-up date, re-hospitalization and all-cause mortality at follow-up, we considered only patients withcomplete follow-up data eligible for analysis. Adverse clinicalevents were defined as any new acute clinical problem, includingany adverse drug reactions, that occurred in hospital or fromdischarge to follow-up date.19 Information on re-hospitalizationand survival of the patients was obtained after 3 months. To checkwhether adverse clinical events were associated with PIMs, weinvestigated the cause of death, re-admission or adverse clinicalevent for each patient.

Statistical analysis

Analysis of variance was used to study the relationshipbetween potentially inappropriate medication and length ofhospital stay, and logistic regression between PIMs and inci-dence of adverse clinical events, re-admission and mortality.Multivariate analyses were adjusted for age, sex and CIRScomorbidity index as possible confounders for adverse clinicaloutcomes. Analyses were carried out with JMP Pro 10 (SASInstitute Inc.)

RESULTS

The initial study sample comprised 1380 subjects, and follow-updata were available for 844 (61�2%). Of the 1380 inpatientsrecruited, data were not available for 536 (38%). Table 1 summa-rizes their socio-demographic characteristics, which were verysimilar to those of patients without follow-up data (mean age79�3 years; female 49�4%; CIRS severity index (mean � SD)1�7 � 0�3; CIRS comorbidity index (mean � SD) 3�0 � 1�8). The2012 Beers’ criteria identified more patients with at least one PIMthan the 2003 version, although a high percentage of those patients(72�2%) were also identified by the criteria updated in 2003. Mainsocio-demographic characteristics of patients identified by Beers’criteria and of those without a PIM at discharge are described inTable 2. According to our results, the main difference in theprevalence of patients receiving a PIM according to the twoversions of Beers’ criteria involved prescriptions of benzodiaze-pines (any type) for insomnia or agitation, chronic use (>90 days)of non-benzodiazepine hypnotics, prescription of antipsychotics inpeople with dementia and oral iron at dosage higher than 325 mg/day (Table 3). Prescription of the drugs listed in the Beers’ criteriawas not associated with a higher risk of adverse clinical events, re-hospitalization and all-cause mortality at 3-month follow-up inboth univariate and multivariate analysis (adjusted for age, sexand CIRS comorbidity index) (Table 4).

© 2014 John Wiley & Sons Ltd Journal of Clinical Pharmacy and Therapeutics, 2014, 39, 511–515512

Potentially inappropriate medications L. Pasina et al.

To check whether PIMs were associated with adverse clinicaloutcome in patients with higher comorbidity, we considereddifferent cut-offs for the CIRS comorbidity index, obtaining similarresults. Finally, among patients who reported an adverse clinicaloutcome at the 3-month follow-up, only one was potentiallyrelated to a PIM listed in both versions of Beers’ criteria, althoughassociated to a severe drug–drug interaction (DDI): one patientreceiving amiodarone and bisoprolol died for cardiac arrest. Forother five patients, who were not receiving a PIM, the cause of re-admission was classified as related to their potentially severe DDIat discharge.

DISCUSSION

In the present study, the use of PIMs according to the Beers’criteria was not associated with adverse health outcomes amongpreviously hospitalized older adults, confirming the findings ofother hospital-based studies.20,21 The Beers’ criteria are widelyemployed for evaluating PIMs in the USA and have been adoptedby the American Geriatrics Society. In Europe, several largeepidemiological studies used the Beers’ criteria to quantify theprevalence of inappropriate prescribing in older people in primary,secondary and long-term care. PIMs have been identified in14–66% of hospitalized older adults.20,22,23 Despite widespread

utilization of these criteria, no randomized controlled studies haveprospectively shown that their clinical application to prescriptionsto older patients improves such clinical outcomes as ADEs,morbidity, mortality, hospitalization and cost.

Some observational studies found an association between Beers’criteria and such serious problems as delirium, gastrointestinalbleeding, falls and fracture,24–27 whereas others found no signif-icant association with ADEs or all-cause mortality,20,21 and theirusefulness in Europe is still debated.

An Italian study of older patients admitted to a geriatric unitfound that among hospital admission related to ADEs that weredefinitely or possibly avoidable, only one-fifth (<1% of alladmissions) had received an inappropriate medication or a drugnot indicated for the diagnosed disease according to the Beers’criteria28. Similarly, a US-based study showed no significant causallink between PIMs according to the Beers’ criteria and ADEsidentified from emergency department visits in a very largenumber of older patients (177,504).29,30

The present study is first to examine the prevalence of PIMsaccording to the 2012 version of Beers’ criteria in hospitalizedelderly patients. The prevalence with the 2003 version was similarto results of other Italian studies,27,28 but there was a slightlyhigher prevalence of patients with at least a PIM using the 2012version. However, there was a broad overlap of patients identifiedby both versions (72%). Neither versions of Beers’ criteria wasassociated with the risk of adverse clinical events, re-hospitaliza-tion and all-cause mortality at 3-month follow-up, confirming theobservational studies that found no significant association withADEs or all-cause mortality. One case of death was classified aspotentially related to a PIM listed in both versions of Beers’ criteriaat the 3-month follow-up, although associated to a severe DDI. In arecent study31 conducted on the same cohort of patients andfocused on DDIs and related risk of adverse outcomes, we foundthat five patients were re-admitted within 3 months after dis-charge for a severe DDI. In all these cases, patients were notreceiving PIMs. A possible explanation of the lack of associationbetween PIMs and related risk of adverse clinical outcome, assuggested in other recent studies,32,33 is that both versions of theBeers’ criteria have several recognized limitations, one of the mainones being their poor transferability outside North America,because they include several medications not available in Europeor rarely prescribed in everyday clinical practice. Concerns overthe suitability of Beers’ criteria for use outside the USA arereinforced by two studies that compared STOPP with Beers’

Table 1. Main socio-demographic characteristics of patients with3-month follow-up data

Patients with follow-up (844)

Age (mean � SD), years 78�8 (7�4)Female, n (%) 432 (51�2)Number of drugs (mean � SD) 6�3 (2�8)Adverse clinical events, n (%) 82 (9�7)Re-hospitalization, n (%) 145 (17�2)Death, n (%) 66 (7�8)Cumulative Illness Rating Scale (CIRS)Diagnosis (mean � SD) 6�7 (3�0)Severity index (mean � SD) 1�7 (0�3)Comorbidity index (mean � SD) 3�0 (1�8)Potentially inappropriate medication (PIM)Beers 2003, n (%) 170 (20�1)Beers 2012 (%) 198 (23�5)

Table 2. Main socio-demographic characteristics of patients identified by Beers’ criteria at discharge

Beers 2003 (170) Beers 2012 (198) Patients without PIMs (619)

Age (mean � SD) 80�5 (7�3) 79�7 (7�1) 78�4 (7�4)Female, n (%) 82 (48�2) 106 (53�5) 315 (50�9)Adverse clinical events, n (%) 14 (8�2) 18 (9�1) 63 (10�2)Re-hospitalization, n (%) 25 (14�7) 30 (15�2) 112 (18�1)Died, n (%) 11 (6�5) 15 (7�6) 48 (7�8)Number of drugsa (mean � SD) 6�7 (2�6) 6�5 (2�6) 5�9 (2�7)Cumulative Illness Rating Scale (CIRS)Severity index (mean � SD) 1�8 (0�3) 1�8 (0�3) 1�7 (0�3)Comorbidity index (mean � SD) 3�7 (2�0) 3�5 (1�9) 2�9 (1�8)

aExcluding potentially inappropriate medications.



criteria and showed that STOPP, unlike Beers, were significantlyassociated with avoidable ADEs in older people that require orcontribute to urgent hospital admission.32,33 The designation ofcertain drugs as inappropriate by Beers’ criteria is debatable, forexample avoidance of amiodarone and doxazosin in older peopleregardless of the diagnosis: amiodarone may be the only agent foreffective control of arrhythmia, and although not often the firstchoice, it may be entirely appropriate in specific cases; similarly,doxazosin may be appropriate in patients with resistant hyper-tension and nitrofurantoin may be the only antimicrobial to which

an infecting pathogen is sensitive and would therefore beappropriate to prescribe.

Limitations of the present study include the small number ofpatients with 3-month follow-up data, which were not availablefor about 38% of those recruited in the second wave of REPOSIand may introduce a bias in the results; we also have not clearinformation on medication adherence, because we could notassess whether a drug prescribed at hospital discharge was reallytaken; the lack of data on adverse clinical outcome and cause ofdeath for some patients who died after discharge limited thepossibility of checking the association between PIMs and ADEs,tending to underestimate the real risk of adverse clinicaloutcome. Again, the study was not designed to specificallyinvestigate the adverse drug events of PIMs according to Beers’criteria and we could not assess whether or not a patient’sclinical responses, laboratory tests or clinical management wereused to reduce the risk of adverse outcomes. Although thetelephone follow-up interview has been performed by an expertclinical monitor, no specific information to establish the causalrelationship between a drug and an adverse drug event has beencollected, as suggested by Naranjo probability scale.34 To quan-tify the clinical relevance of PIMs at discharge, a prospectivedesign would be necessary, including longer follow-up afterdischarge and specific collection of drug-related problems.Another limitation includes the generalizability of the results,which refer to a hospital population and cannot be extrapolatedto elderly subjects in different settings.

WHAT IS NEW AND CONCLUSION

This study found no significant effect of inappropriate drug useaccording to Beers’ criteria on health outcomes among previously

Table 3. Prevalence of potentially inappropriate medications (PIMs) among the 844 patients discharged

Beers 2003 Patients (%) Beers 2012 Patients (%)

Ticlopidine 54 (6�5) Ticlopidine 54 (6�4)Amiodarone 43 (5�1) Antiarrhythmic drugs 47 (5�6)Doxazosin 29 (3�4) Alpha1 Blockers 29 (3�4)Clonidine 22 (2�6) Alpha blockers, central 22 (2�6)Iron > 325 mg/day 11 (1�3) Benzodiazepinesd 21 (2�5)Chlorpropamide 9 (1�1) Sulfonylureas 16 (1�9)Amitriptyline 4 (0�5) Antipsychotics in demented people 13 (1�5)First-generation antihistamines 4 (0�5) Barbiturates 10 (1�2)Digoxin > 125 mg/daya 3 (0�4) Non-benzodiazepine hypnoticse 4 (0�5)Indomethacin 2 (0�2) Tricyclic antidepressants 4 (0�5)Muscle relaxantsb 2 (0�2) First-generation antihistamines 4 (0�5)Nitrofurantoin 2 (0�2) Indomethacin or Ketorolac 4 (0�5)Ketorolac 2 (0�2) Digoxin > 125 mg/daya 3 (0�4)Flurazepam 1 (0�1) Nitrofurantoin 2 (0�2)High-dose short-acting benzodiazepinesc 1 (0�1) Long-term use of non-coxib NSAIDS without gastroprotective agent 1 (0�1)Fluoxetine 1 (0�1) Androgens 1 (0�1)Barbiturates (except phenobarbital) 1 (0�1) Megestrol 1 (0�1)

Metoclopramide 1 (0�1)Mineral oil, oral 1 (0�1)

aExcept for patients with atrial arrythmias.bIncluding methocarbamol, carisoprodol, oxybutynin, chlorzoxazone, metaxalone and cyclobenzaprine.cDaily doses exceeding: lorazepam, 3 mg; oxazepam, 60 mg; alprazolam, 2 mg; temazepam, 15 mg; zolpidem, 5 mg; triazolam, 0�25 mg.dExcept for treating seizure disorders, rapid-eye-movement sleep disorders, benzodiazepine withdrawal, ethanol withdrawal,severe generalized anxiety disorder, periprocedural anaesthesia, end-of-life care.eAccording to information reported by patients.

Table 4. PIMs and adverse clinical outcomes at 3-month follow-up

Beers 2003 Beers 2012

OR (95%, CI) P value OR (95%, CI) P value

Univariate analysisAdverse clinicalevents

0�78 (0�43–1�43) 0�42 0�91 (0�52–1�57) 0�72

Re-hospitalization 0�77 (0�48–1�24) 0�28 0�82 (0�53–1�27) 0�37All-causemortality

0�78 (0�40–1�52) 0�45 0�96 (0�53–1�74) 0�88

Multivariate analysisAdverse clinicalevents

0�75 (0�39–1�34) 0�34 0�88 (0�49–1�51) 0�65

Re-hospitalization 0�72 (0�43–1�15) 0�17 0�77 (0�48–1�19) 0�24All-causemortality

0�60 (0�29–1�16) 0�13 0�84 (0�44–1�52) 0�58

Adjusted for age, sex and CIRS comorbidity index.



hospitalized older adults. Even though these criteria have beensuggested as helpful in promoting appropriate prescribing,reducing drug-related adverse events and associated healthcarecosts, to date there is no clear evidence that their application canachieve objective and quantifiable improvements in clinicaloutcomes. A possible explanation is that both versions of theBeers’ criteria have several recognized limitations, one of the mainones being the restricted availability of some drugs in Europe ortheir limited prescription in everyday clinical practice.

ACKNOWLEDGEMENTS

REPOSI is a network of Italian internal medicine hospital wardswhich, voluntarily and without any financial support, agreed toparticipate in data collection during the four index weeks. We aregrateful to J.D. Baggott for editorial assistance.

CONFLICT OF INTEREST

The authors have no conflict of interest.

REFERENCES

1. Veehof LJG, Stewart RE, Haaijer-RaskampFM et al. The development of polypharma-cy. A longitudinal study. Fam Pract,2000;17:261–267.

2. Gurwitz JH. Polypharmacy. A new para-digm for quality drug therapy in the elderly?Arch Intern Med, 2004;164:1957–1959.

3. Hajjar ER, Cafiero AC, Hanlon JT. Poly-pharmacy in elderly patients. Am J GeriatrPharmacother, 2007;5:345–351.

4. Hilmer SN, Gnjidic D. The effects of poly-pharmacy in older adults. Clin PharmacolTher, 2009;85:86–98.

5. Steinman MA, Landefeld CS, Rosenthal GEet al. Polypharmacy and prescribing qualityin older people. J Am Geriatr Soc,2006;54:1516–1523.

6. Koh Y, Kutty FBM, Li SC. Drug-relatedproblems in hospitalized patients on poly-pharmacy: the influence of age and gender.Ther Clin Risk Manag, 2005;1:39–48.

7. Lund BC, Carnahan RM, Egge JA, Chrischil-les EA, Kaboli PJ. Inappropriate prescribingpredicts adverse drug events in older adults.Ann Pharmacother, 2010;44:957–963.

8. Nobili A, Garattini S, Mannucci PM. Multi-ple diseases and polypharmacy in theelderly: challenges for the internist of thethird millennium. J Comorbidity, 2011;1:28–44.

9. O’Connor MN, Gallagher P, O’Mahony D.Inappropriate prescribing: criteria, detectionand prevention. Drugs Aging, 2012;29:437–452.

10. Beers MH, Ouslander JG, Rollingher I et al.Explicit criteria for determining inappropri-ate medication use in nursing home resi-dents. Arch Intern Med, 1991;151:1825–1832.

11. Chang CB, Chan DC. Comparison of pub-lished explicit criteria for potentially inap-propriate medications in older adults. DrugsAging, 2010;27:947–957.

12. American Geriatrics Society 2012 Beers Crite-ria Update Expert Panel. American GeriatricsSociety updated Beers Criteria for potentiallyinappropriate medication use in older adults.J Am Geriatr Soc, 2012;60:616–631.

13. Patterson SM, Hughes C, Kerse N, CardwellCR, Bradley MC Interventions to improve

the appropriate use of polypharmacy forolder people. Cochrane Database Syst Rev,2012;5:CD008165.

14. Parmelee PA, Thuras PD, Katz IR et al.Validation of the cumulative illness ratingscale in a geriatric residential population. JAm Geriatr Soc, 1995;43:130–137.

15. Shah S, Vanclay F, Cooper B. Improving thesensitivity of the Barthel Index for strokerehabilitation. J Clin Epidemiol, 1989;42:703–709.

16. Katzman R, Brown T, Fuld P et al. Valida-tion of a Short Orientation-Memory-Con-centration Test of cognitive impairment. AmJ Psychiatry, 1983;140:734–739.

17. Hickie C, Snowdon J Depression scales forthe elderly: GDS, Gilleard, Zung. Clin Ger-ontol, 1987;6:51–53.

18. Fick DM, Cooper JW,WadeWE et al.Updat-ing the Beers criteria for potentially inappro-priate medication use in older adults: resultsof a US consensus panel of experts. ArchIntern Med, 2003;163:2716–2724.

19. Bernardini B, Meinecke C, Zaccarini C et al.Adverse clinical events in dependent long-term nursing home residents. J Am GeriatrSoc, 1993;41:105–111.

20. Laroche ML, Charmes JP, Nouaille Y et al. Isinappropriate medication use a major causeof adverse drug reactions in the elderly? Br JClin Pharmacol, 2007;63:177–186.

21. Onder G, Landi F, Liperoti R et al. Impact ofinappropriate drug use among hospitalizedolder adults. Eur J Clin Pharmacol,2005;61:453–459.

22. Onder G, Landi F, Cesari M et al. Inappro-priate medication use among hospitalizedolder adults in Italy: results from the ItalianGroup of Pharmacoepidemiology in theElderly. Eur J Clin Pharmacol, 2003;59:157–162.

23. Corsonello A, Pedone C, Lattanzio F et al.Potentially inappropriate medications andfunctional decline in elderly hospitalizedpatients. J Am Geriatr Soc, 2009;57:1007–1014.

24. Jano E, Aparasu RR. Healthcare outcomesassociated with Beers’ Criteria: a systematicreview. Ann Pharmacother, 2007;41:438–447.

25. Chang CM, Liu PY, Yang YH et al. Use ofthe Beers Criteria to predict adverse drugreactions among first-visit elderly outpa-tients. Pharmacotherapy, 2005;25:831–838.

26. Chrischilles EA, VanGilder R, Wright Ket al. Inappropriate medication use as a riskfactor for self-reported adverse drug effectsin older adults. J Am Geriatr Soc,2009;57:1000–1006.

27. Passarelli MC, Jacob-Filho W, Figueras A.Adverse drug reactions in an elderly hospi-talised population: inappropriate prescrip-tion is a leading cause. Drugs Aging,2005;22:767–777.

28. Franceschi M, Scarcelli C, Niro V et al.Prevalence, clinical features and avoidabil-ity of adverse drug reactions as cause ofadmission to a geriatric unit: a prospectivestudy of 1756 patients. Drug Saf,2008;31:545–556.

29. Budnitz DS, Shehab N, Kegler SR et al.Medication use leading to emergencydepartment visits for adverse drug eventsin older adults. Ann Intern Med,2007;147:755–765.

30. Budnitz DS, Lovegrove MC, Shehab N et al.Emergency hospitalizations for adversedrug events in older Americans. N Engl JMed, 2011;21:2002–2012.

31. Pasina L, Djade CD, Nobili A et al. Drug-drug interactions in a cohort of hospitalizedelderly patients. Pharmacoepidemiol Drug Saf,2013;22:1054–1060.

32. Hamilton H, Gallagher P, Ryan C et al.Potentially inappropriate medicationsdefined by STOPP criteria and the risk ofadverse drug events in older hospitalizedpatients.Arch InternMed, 2011;171:1013–1019.

33. Gallagher P, O’Mahony D. STOPP (Screen-ing Tool of Older Persons’ potentially inap-propriate Prescriptions): application toacutely ill elderly patients and comparisonwith Beers’ criteria. Age Ageing,2008;37:673–679.

34. Naranjo CA, Busto U, Sellers EM. A methodfor estimating the probability of adversedrug reactions. Clin Pharmacol Ther,1981;30:239–245.



prevalence of potentially inappropriate medications and risk of adverse clinical outcome in a cohort...

Documents