- Email: [email protected]
Update on drug-related problems in the elderly
The American Journal of Geriatric Pharmacotherapy
J.T. Hahn
et al.
Update on Drug-Related Problems in the Elderly Joseph T. Hanlon, PharmD, MS,‘” Teresa C. McCarthy, MD, MS4
Catherine I. Lindblad, PharmD,lf3 Emily R. Hajjar, PharmD,‘a
and
‘Institute for the Study of Geriatric Pharmacotherapy and Department of Experimental and Clinical Pharmacology College of Pharmacy, and *Division of Health Services Research and Policy, School of Public Health, University of Minnesota, 3Geriatric Research, Education and Clinical Center; Veterans Affairs Medical Center, and 4Program in Aging and Geriatric Medicine, School of Medicine, University of Minnesota, Minneapolis
ABSTRACT Background: Although
as defined by the Beers criteria. Another study examined whether inappropriate drug use, as defined by the Beers or DUR criteria, was associated with death or a decline in functional status; it found that only use of drugs defined as inappropriate by DUR criteria (particularly those drugs associated with drug-drug or drug-disease interactions) was associated with a decline in the ability to perform basic self-care. Three studies, 1 from the United States, 1 from the United Kingdom, and 1 from Canada, described consensus development of quality indicators for drug use in the elderly, including drugs to avoid, maximum daily dose, drug duplication, limits on duration of use, drug-drug and drug-disease interactions, need for drug monitoring, underuse of necessary drugs to treat or prevent common problems, and inappropriate drug-administration technique. Conclusions: Drug-related problems are common in elderly patients. Data from recently published studies provide guidance to practitioners and directions for lkure research. (Am / Geriaw Pharmacother. 2003;1:3843) Copyright 0 2003 Excerpta Medica, Inc. Key words: elderly, prescribing, adverse drug reactions.
INTRODUCTION Although pharmacotherapy
of MEDLINE (2002-March ing drug-related problems, and adverse drug events.
pharmacotherapy for the elderly can treat diseases and improve well-being, its benefits can be compromised by drug-related problems. Objective: This article reviews recent publications concerning drug-related problems in the elderly, as well as articles describing the development of 3 sets of quality indicators for medication use in the elderly. Methods: Relevant articles were identified through a search of MEDLINE (2002-March 2003) for articles on drug-related problems, inappropriate prescribing, and adverse drug events in the elderly. Results: The review included 7 articles published in 2002 and 2003. A study in ambulatory elderly persons reported that -5.0% of patients had 21 adverse drug event within the previous year. Another study found that -20.0% of ambulatory elderly persons used 21 inappropriate drug, as defined by drug utilization review (DUR) criteria, with drug-disease interactions and duration of use being the most common drug-related problems. A third study involving elderly individuals in assisted living facilities found that 16.0% used 21 inappropriate drug,
for the elderly can treat diseases and improve well-being, its benefits can be compromised by drug-related problems. Strand et al1 identied 8 types of drug-related problems: (1) untreated indication, (2) drug use without an indication, (3) improper drug selection, (4) subtherapeutic dosage, (5) overdosage, (6) medication error/noncompliance, (7) drug interactions, and (8) adverse drug reactions. The present article reviews 4 recent studies examining the epidemiology of drug-related problems in the elderly,2-5 as well as 3 articles describing the development of separate sets of quality indicators for medication use in the elderly.@ These publications were identified through a search
38
September2003
Volume
1
??
Number 1
2003) for articles concerninappropriate prescribing,
EPIDEMIOLOGY OF DRUG-RELATED PROBLEMS IN THE ELDERLY Gurwitz et al2 evaluated the incidence of adverse drug events (ADEs) over a l-year period in >30,000 elderly ambulatory patients enrolled in a multispecialty clinic in central Massachusetts. Using a prospective cohort design, they screened voluntary reports, computer-generated Accepted for publication July 9, 2003. Copyright 0 2003 Excerpta Medica, Inc.
1543-5946/03/$19.00
J. 2: Hanlon et al.
laboratory and medical records signals, and emergency department and hospital discharge summaries for potential ADEs. Written case reports of the potential ADEs were prepared by clinical pharmacists for evaluation by pairs of physicians using implicit review methods rather than a standardized algorithm. They identified 1523 ADEs, for an incidence of 50.1/1000 person-years. Overall, 11 ADEs were fatal, 136 life threatening, and 431 serious; the remainder were described as significant. Four hundred twenty-one (27.6%) ADEs were judged to have been preventable. Among the preventable ADEs, problem areas included monitoring (60.8%), prescribing (58.4%), adherence (21.10/o), and dispensing (~2.0%). The most common medication classes involved in preventable ADEs included cardiovascular agents, diuretics, nonopioid analgesics, hypoglycemics, and anticoagulants. This study had a number of strengths, including, but not limited to, its large sample size, use of electronic screening for potential ADEs, and consensus evaluation of ADEs by physicians using case reports abstracted from patients’ charts by clinical pharmacists2 It has, however, several limitations. One concern is a potential underdetection of “true” ADEs. Table I shows the rates of ADEs from this and 2 other studies in the ambulatory elderly.9J0 The rate of -5% reported by Gurwitz et al is 2 to 7 times lower than the rates reported in the other 2 studies. This difference may be due in part to the fact
Table 1. Adverse
Author
that the latter 2 studies used a semistructured interview to determine whether participants had experienced any adverse effects, unwanted reactions, or other problems with their medications over the follow-up period. Patient self-report appears to be a more sensitive method for detecting ADEs in ambulatory patients than record review.“>12 Another possible reason for the difference is that Gurwitz et al did not screen handwritten notes or records of medication discontinuations (this can be done electronically), both sensitive and reliable approaches to screening for potential ADEs.12 Another limitation to the study by Gurwitz et al2 was discussed in an accompanying editorial, which noted a need for further refinement of the definitions of serious ADEs and preventable ADEs.13 The US Food and Drug Administration defines a serious ADE as any event due to a drug that causes death, requires or prolongs hospitalization, is permanently disabling, is life threatening, or requires intervention to prevent permanent impairment or damage. l4 Finally, it is difficult to reconcile the low proportion of ADEs rated as preventable (27.6%) by Gurwitz et al with the finding of previous studies that 95% of ADEs are type A reactions; that is, common, predictable, less serious, dose-related extensions of a drug’s usual pharmacologic effects and thus, by definition, potentially preventable.15 Two studies examined the prevalence of inappropriate prescribing in the elderly and another examined the
drug events (ADEs) in elderly outpatients
Setting
Gurwitz et al2
Multispecialty group practice in central Massachusetts
Chrischilles et al9
2 Rural Iowa counties
Hanlon et allo
Veterans Affairs general medicine clinic
Sample Size
Means of Detecting ADEs
The American Journal of Geriatric Pharmacotherapy
in selected studies.
Period of Observation
% with 21 ADE
Incidence per 1000 Person-Years
Evaluation of ADEs
No. of ADEs
Health care providers, hospital and emergency department discharge summaries, electronic medical records
Consensus evaluation by pair of physician reviewers
1523
1 year
5
50
3170
Self-report
Pharmacist evaluation
521
1 year
IO
164
167
Self-report
Pharmacist evaluation
80
1 year
35
479
>30,000
I
39
The American Journal of Geriatric Pharmacotherapv
J.T Hanion et al.
association between inappropriate prescribing and health outcomes.3-5 The first study applied drug utilization review (DUR) criteria to medications reported as being taken by respondents at the second and third in-person interviews in the Duke University Established Populations for Epidemiologic Studies of the Elderly ( EPESE).3 Information about sociodemographic characteristics, health status, access to health care, and drug use in the previous 2 weeks was collected during an in-home interview. Drug use was coded for 8 drugs or therapeutic classes-digoxin, calcium channel blockers, angiotensin-converting enzyme inhibitors, histamineareceptor antagonists, nonsteroidal anti-inflammatory drugs (NSAIDs), benzodiazepines, antipsychotics, and antidepressants-and for appropriateness by applying the DUR criteria for maximum daily dose, duplication, drug-drug interactions, duration of therapy, and drug-disease interactions. It was found that 21.0% of participants in the second in-person survey and 19.2% of participants in the third in-person survey who used 21 agent from the 8 drug categories had 21 element flagged as inappropriate. Benzodiazepines and NSAIDs were the therapeutic classes associated with the highest frequency of problems. The most common problems involved drugdisease interactions and duration of use. Two factorswhite race and arthritis-were found to have important associations with inappropriate drug use in both the cross-sectional and longitudinal multivariable analyses. The investigators concluded that inappropriate drug use is common among the community-dwelling elderly and persists over time. The Duke University EPESE is a rich data set that includes detailed information on demographic variables, health status, and access to health care. Drug-utilization information was collected in an optimal fashion through the use of structured interviews and visual verification of each medication reported as taken. The findings of the Duke University EPESE appear to have face validity. With limited geriatric-specific training during the years in which the second in-person interview of the Duke University EPESE was conducted (1989-1990), few clinicians would have identified these drugs (benzodiazepines and NSAIDs) as problematic. It is not surprising that by the third round of in-person interviews in the Duke University EPESE, use of these drugs had decreased, reflecting clinicians’ growing knowledge about older patients. The finding of an association between white race and arthritis and failure to meet the DUR criteria3 are consistent with clinical realities. White patients are more likely to be prescribed benzodi-
40
I
azepines than are those of other racial groups,16 and arthritis is frequently managed with NSAID therapy. The finding that patients with a prescription for inappropriate drugs at the second in-person interview were more likely to have a prescription for inappropriate drugs at the third in-person interview should raise questions about patient characteristics that were not collected in the study (eg, personality factors, compliance with physician recommendations), as well as questions about characteristics of the physicians prescribing these medications (eg, level of training, practice type, methods of monitoring prescription refills). Overall, this study confirmed what is known about potentially dangerous drug use in older patients during the era these data were collected. The results support the DUR process as an effective tool for identifying populations at risk for inappropriate prescribing. The second study was designed to identify the prevalence of inappropriate medication use in elderly persons living in residential care/assisted living facilities in Florida, Maryland, New Jersey, and North Carolina.4 The authors conducted a secondary analysis of data collected in the facilities by trained nurses as part of the Collaborative Studies of Long-Term Care. Inappropriate medication use in the elderly was defined using a modification of the consensus criteria reported by Beers17 that incorporated input from Dr. Beers. These authors found that 16.0% of patients were receiving 21 inappropriate medication. The most common inappropriate medications were oxybutynin, propoxyphene, amitriptyline, diphenhydrarnine, and ticlopidine. Cross-sectional multivariable analyses revealed an increased risk of inappropriate drug use in those who received a higher number of medications and at those facilities with moderate rates of turnover of licensed practical nurses. High monthly rates, moderate or severe dementia, smaller number of beds, and weekly physician visits were protective against inappropriate prescribing. The investigators concluded that inappropriate prescribing was common among residents of assisted living facilities and that regular physician visits to such facilities and attention to such highrisk groups as those taking multiple medications may improve prescribing. The finding that drugs to be avoided in the elderly were commonly prescribed for patients in assisted living facilities4 is no surprise. However, the rate of 16.0% is lower than that reported in another study (25 .0%).18 This may be due in part to exclusion of the Beers maximum daily dose criteria for short-acting benzodiazepines and digoxin from the Collaborative Studies of Long-
J. T Hanlon et al.
Term Care. One might expect use of these agents to be more prevalent in this population than many or most of the inappropriate medications identified in the study. It is also notable that facility characteristics were more common risk factors for inappropriate prescribing than were patient attributes. The third study used longitudinal multivariable analyses to assess the association between inappropriate drug use, as defined by both Beers and DUR criteria, and mortality or decline in functional statu~.~ The study used data on sociodemographic characteristics, health status, access to health care, and drug use for 3234 subjects from the Duke University EPESE. Death was identified based on the National Death Index. Change in 4 measures of fimctional status-basic self-care, intermediate self-care, complex self-care, and physical function-was determined over 3 years from baseline. No significant association was found between mortality and inappropriate drug use identified by either set of criteria. Based on the Beers criteria, there was no significant association between inappropriate drug use and decline in functional status. Based on the DUR criteria, however, the relationship between inappropriate drug use and the ability to perform basic self-care was significant (P < 0.05) and was more pronounced for drugs associated with problems of drug-drug or drug-disease interactions. The investigators concluded that identifying the impact of inappropriate drug use may depend on the criteria applied. This study compared the predictive validity of 2 criteria for inappropriate drug use in the elderly: DUR and drugs to avoid (Beers criteria).5 Mortality and functional status are clinically and epidemiologically relevant outcomes, and the population was appropriately geriatricthe majority were aged ~75 years, female, in fair to poor health, and taking multiple medications. The findings suggest, not surprisingly, that the relationship between drug use and relevant outcomes is highly patient specific. Thus, criteria of the drugs-to-avoid type may have limited usefulness. Criteria that incorporate more detailed characteristics of patients’ drug use and medical status would provide a better tool for identifying potentially harmful drug use in the elderly
MEASURING THE QUALITY OF MEDICATION USE IN THE ELDERLY Three recent articles-l from the US,6 1 from the United Kingdom,7 and 1 from Canada8-described the consensus development of new sets of quality indicators for drug use in older persons. They shared the common goal of decreasing the incidence of preventable drug-
The American Journal of Geriatric Pharmacothevapy
related morbidity (PDRM). As defined in these studies, a PDRM must have been recognizable as drug related, the treatment failure must have been foreseeable, and the causes must have been both identifiable and controllable. The US article described a list of drug-quality indicators and their relation to PDRM based on data compiled from a review of the literature and agreed by consensus of an expert panel using a 2-round Delphi technique.6 The 7-member panel included physicians, physicianadministrators, and a geriatric-specialty clinical pharmacist. Clinical indicators of PDRM were presented as associations between particular patterns of care and adverse therapeutic outcomes, For classification as a clinical indicator of PDRM, a scenario had to be well referenced, occur fairly commonly in the geriatric population, result in serious to moderate adverse outcomes, and be searchable in the electronic medical records of the health care system cooperating in the study. Only those scenarios chosen by a majority of panel members were included in the final list. Overall, the panel reached consensus on 52 clinical indicators of PDRM, including 20 drug-disease interactions, required monitoring for 22 drug classes, 1 drug-drug interaction, 4 drugs to avoid, 1 drug with a limit on duration of use, 1 drug duplication, and 3 instances of underuse of necessary medications.6 Quality indicators not covered by other groups included avoidance of typical antipsychotic agents and laboratory test monitoring for aminoglycosides, angiotensin-converting enzyme inhibitors, antidiabetic agents, antiepileptic drugs, heparin, lithium, levothyroxine, NSAIDs, theophylline, ticlopidine, troglitazone, and concomitant trimethoprim/sulfamethoxazole and methotrexate. Also unique were the criteria for underuse of necessary drugs (ie, inhaled steroids in asthmatic patients, treatment for the prevention of osteoporosis in patients taking chronic corticosteroids) and drug-disease interactions (ie, imipramine and diabetic bladder atony, and use of nitrofurantoin, nalidixic acid, or methenamine in patients with chronic renal failure j. The UK investigation also developed quality indicators for the prevention of PDRM.7 Starting with the 52 indicators developed by the US group,6 the UK group deleted 20 criteria that did not apply to or were of low prevalence in outpatient elderly persons in the UK health care system. They then conducted a questionnaire survey of a panel of 5 general practitioners and 10 pharmacists directly involved in medication review in general practice. The 2-round Delphi survey used a scale from
The American journal
of Geriatric Pharmacotherapy
J.T. Hanh
et al.
1 (definitely PDRM) to 7 (definitely not PDRM) and considered 53 individual criteria. By the end of the second round, the panel had identified 29 quality indicators. These indicators addressed 9 drug-disease interactions, 6 drug-drug interactions, 2 drugs to avoid, 3 necessary medications for the prevention of medical problems, and 9 drugs requiring monitoring. They developed unique criteria for drug-disease interactions (ie, metoclopramide and Parkinson’s disease), laboratory monitoring (ie, statins and liver function tests), and appropriate drug-administration techniques (ie, use of a spacer with corticosteroid inhalers). In the Canadian for medication
effort
to develop
quality indicators
use in the elderly,* 2 clinical pharmacists
reviewed the 52 US indicator& for relevance. Fifty-eight quality indicators were developed based on a 2-stage Delphi survey of 6 geriatricians and 6 clinical pharmacologists. A modified nominal group technique was then used to obtain feedback from 12 general practitioners, which led to the deletion of 6 quality indicators and a final set of 52 PDRMs. The clinical indicators included 7 drug-disease interactions, 4 drug-drug interactions, 9 drugs to avoid, 2 drug duplications, 10 conditions in which necessary drugs were underused, 13 drugs requiring laboratory monitoring, 2 drugs with limits to duration of use, and 5 drugs with maximum dosage limits. The group developed unique criteria for underuse (ie, influenza and pneumococcal vaccination, regularly scheduled analgesics for chronic pain), maximum initial dosing of individual drugs (ie, levodopa/carbidopa and risperidone), drugs to avoid (ie, multiple psychotropic agents), drug duration (ie, risperidone for delirium after hospitalization), and drug-disease interactions (ie, neuroleptics and Parkinson’s disease). The new clinical indicators for PDRM developed by these 3 investigative groups built on indicators compiled by earlier groups. 17~19-22It is notable that the recent groups did not consider the Centers for Medicare & Medicaid Services DUR guidelines concerning drugtherapy monitoring indicators for long-term care facilities22 or the Study of Clinically Relevant Indicators for Pharmacologic Therapy. 2o It is also interesting that each of these studies used a different method for reaching consensus. Moreover, only 4 drug-disease interactions, 3 drugs requiring monitoring, 2 drug-drug interactions, 1 drug duplication, and 1 drug class to avoid appeared on all 3 lists (Table II). Finally, these groups did not provide the evidence on which their recommendations for baseline laboratory monitoring of drugs and timing of follow-up monitoring were based.
42
I
Table II. Quality indicators appearing in each of 3 sets of consensus guidelines for drug use in older personsm Drugs
to avoid
Long half-life Drug-disease
benzodiazepines interactions
in patients with COPD in patients with heart failure NSAlDs and steroids in patients with peptic ulcer disease Beta-blockers NSAlDs
Drug-drug interactions Antibiotics and warfarin NSAlDs and warfarin Drug duplication 22 NSAlDs Required monitoring INR for warfarin Thyroid function tests in patients taking levothyroxine Electrolytes in patients taking ACE inhibitors COPD = chronic obstructive pulmonary disease; NSAlDs = nonsteroidal anti-inflammatory drugs; INR = international normalized ratio; ACE = angiotensin-converting enzyme.
CONCLUSIONS Drug-related problems, including ADEs and inappropriate drug use, are common in elderly patients. Although new sets of quality indicators for prescribed medications have been developed by groups in 3 countries, there is minimal overlap between the sets. Future research should examine whether the identified negative outcomes are associated with failure to comply with quality indicators for the prescribed agents. Studies are also needed to develop prescribing guidelines that take into account directions for use, indications, and effectiveness of therapy in the frail elderly. Finally, it would be helpful to be able to identify those at high risk for ADEs a priori and to test interventions to reduce suboptimal drug use and ADEs.
REFERENCES 1. Strand LM, Morley PC, Cipolle RJ, et al. Drug-related problems: Their structure and function. DICP
1990;24:
1093-1097. 2. Gurwitz JH, Field TS, Harrold LR, et al. Incidence and preventability of adverse drug events among older persons in the ambulatory setting. JAMA. 2003;289:1107-1116. 3. Hanlon JT, Schrnader KE, Boult C, et al. Use of inappropriate prescription drugs by older people. JAm Geriatr Sot.
2002;50:26-34.
J.T Hahn
4. Sloane PD, Zimmerman S, Brown LC, et al. Inappropriate medication prescribing in residential care/assisted living facilities. JAm
Geriatr Sot. 2002;50:1001-1011.
et al.
The American Journal of Geriatric Pharmacotherapy
14. Kessler DA. Introducing MEDWatch. A new approach to reporting medication and device adverse effects and product problems. JAMA.
1993;269:2765-2768.
5. Hanlon JT, Fillenbaum GG, Kuchibhatla M, et al. Impact
15. Venulet J, ten Ham M. Methods for monitoring and doc-
of inappropriate drug use on mortality and functional status in representative community dwelling elders. Med
umenting adverse drug reactions. Int J Clin Pharmacol Ther. 1996;34:112-129.
Care. 2002;40:166-176. 6. MacKirmon NJ, Hepler CD. Preventable drug-related mor-
16. Blazer D, Hybels C, Simon& E, Hanlon JT. Sedative, hypnotic and anti-anxiety medication use in an aging
bidity in older adults: Indicator development. J Managed
cohort over ten years: A racial comparison. JAm Gtiatr Sot.
Care Barr.
2000;48:1073-1079.
2002;8:365-371.
7. Morris CJ, Canttill JA, Hepler CD, Noyce IX
Preventing
17. Beers MH.
Explicit criteria for determining
potentially
drug-related morbidity~etermining valid indicators. Int JQuaLity Health Care. 2002;14:183-198.
inappropriate medication use by the elderly. An update.
8. Robertson HA, MacKinnon NJ. Development of a list of consensus-approved clinical indicators of preventable
18. Spore DL, Mor V, Larrat I?, et al. Inappropriate drug pre-
drug-related morbidity in older adults. Clin Ther. 2002;
scriptions for elderly residents of board and care facilities.
Am J Ptlblic Health. 1997;87:404-409. 19. Knight EL, Avorn J. Quality indicators for appropriate
24:1595-1613. 9. Chrischilles
Arch Intern Med. 1997;157:1531-1536.
EA, Segar ET,
Wallace BB.
Self-reported
adverse drug reactions and related resource use. A study of community-dwelling
persons 65 years of age and older.
medication use in vulnerable elders. Ann Intern Med. 2001; 135:703-710. 20. Study of Clinically Relevant Indicators for Pharmacologic Therapy (SCRIPT).
Ann Intern Med. 1992;117:634-640. 10. Hanlon JT, Schmader KE, Koronkowsld MJ, et al. Adverse
drug events in high risk older outpatients. J Am Geriati
Available at: http://scriptproject.org/
home.html. Accessed May 23,2003. 21. McLeod PJ, Huang AB, Tamblyn BM, Gayton DC. Defining inappropriate practices in prescribing for elderly
Sot. 1997;45:945-948. 11. Gandhi TK, Weingart SN, Borus J, et al. Adverse drug events in ambulatory care. NE& 1564.
JMed. 2003;348:1556-
people: A national consensus panel. CM@
1997;156:385-
391. 22. Appendix N. Surveyor procedures for pharmaceutical serin long-term
care facilities. In: State
methods for detecting potential adverse drug events in frail elderly inpatients and outpatients. Am J Health Syst
Operations Manual, Provider Health and Human Services,
Certification. Dept of Health Care Financing
Pharm. 2001;58:1622-1626.
Administration.
12. Hanlon JT, Maher B, Lindblad C, et al. Comparison of
13. Classen D. Medication
safety: Moving from illusion to
reality. JA.MA. 2003;289:1154-1156.
Editorial.
vice requirements
Transmittal no. 242,
September
1990.
Available at: http://www.ascp.com/public/pr/2OOl/pdfs/ App_N.pdf Accessed May 23,2003.
Address correspondence to: Joseph T. Hanlon, PharmD, MS, Institute for the Study of Geriatric Pharmacotherapy, College of Pharmacy, 7-l 15 Weaver-Densford Hall, University of Minnesota, 308 Harvard Street SE, Minneapolis, MN 55455. E-mail: [email protected]
I
43
J.T. Hahn
et al.
Update on Drug-Related Problems in the Elderly Joseph T. Hanlon, PharmD, MS,‘” Teresa C. McCarthy, MD, MS4
Catherine I. Lindblad, PharmD,lf3 Emily R. Hajjar, PharmD,‘a
and
‘Institute for the Study of Geriatric Pharmacotherapy and Department of Experimental and Clinical Pharmacology College of Pharmacy, and *Division of Health Services Research and Policy, School of Public Health, University of Minnesota, 3Geriatric Research, Education and Clinical Center; Veterans Affairs Medical Center, and 4Program in Aging and Geriatric Medicine, School of Medicine, University of Minnesota, Minneapolis
ABSTRACT Background: Although
as defined by the Beers criteria. Another study examined whether inappropriate drug use, as defined by the Beers or DUR criteria, was associated with death or a decline in functional status; it found that only use of drugs defined as inappropriate by DUR criteria (particularly those drugs associated with drug-drug or drug-disease interactions) was associated with a decline in the ability to perform basic self-care. Three studies, 1 from the United States, 1 from the United Kingdom, and 1 from Canada, described consensus development of quality indicators for drug use in the elderly, including drugs to avoid, maximum daily dose, drug duplication, limits on duration of use, drug-drug and drug-disease interactions, need for drug monitoring, underuse of necessary drugs to treat or prevent common problems, and inappropriate drug-administration technique. Conclusions: Drug-related problems are common in elderly patients. Data from recently published studies provide guidance to practitioners and directions for lkure research. (Am / Geriaw Pharmacother. 2003;1:3843) Copyright 0 2003 Excerpta Medica, Inc. Key words: elderly, prescribing, adverse drug reactions.
INTRODUCTION Although pharmacotherapy
of MEDLINE (2002-March ing drug-related problems, and adverse drug events.
pharmacotherapy for the elderly can treat diseases and improve well-being, its benefits can be compromised by drug-related problems. Objective: This article reviews recent publications concerning drug-related problems in the elderly, as well as articles describing the development of 3 sets of quality indicators for medication use in the elderly. Methods: Relevant articles were identified through a search of MEDLINE (2002-March 2003) for articles on drug-related problems, inappropriate prescribing, and adverse drug events in the elderly. Results: The review included 7 articles published in 2002 and 2003. A study in ambulatory elderly persons reported that -5.0% of patients had 21 adverse drug event within the previous year. Another study found that -20.0% of ambulatory elderly persons used 21 inappropriate drug, as defined by drug utilization review (DUR) criteria, with drug-disease interactions and duration of use being the most common drug-related problems. A third study involving elderly individuals in assisted living facilities found that 16.0% used 21 inappropriate drug,
for the elderly can treat diseases and improve well-being, its benefits can be compromised by drug-related problems. Strand et al1 identied 8 types of drug-related problems: (1) untreated indication, (2) drug use without an indication, (3) improper drug selection, (4) subtherapeutic dosage, (5) overdosage, (6) medication error/noncompliance, (7) drug interactions, and (8) adverse drug reactions. The present article reviews 4 recent studies examining the epidemiology of drug-related problems in the elderly,2-5 as well as 3 articles describing the development of separate sets of quality indicators for medication use in the elderly.@ These publications were identified through a search
38
September2003
Volume
1
??
Number 1
2003) for articles concerninappropriate prescribing,
EPIDEMIOLOGY OF DRUG-RELATED PROBLEMS IN THE ELDERLY Gurwitz et al2 evaluated the incidence of adverse drug events (ADEs) over a l-year period in >30,000 elderly ambulatory patients enrolled in a multispecialty clinic in central Massachusetts. Using a prospective cohort design, they screened voluntary reports, computer-generated Accepted for publication July 9, 2003. Copyright 0 2003 Excerpta Medica, Inc.
1543-5946/03/$19.00
J. 2: Hanlon et al.
laboratory and medical records signals, and emergency department and hospital discharge summaries for potential ADEs. Written case reports of the potential ADEs were prepared by clinical pharmacists for evaluation by pairs of physicians using implicit review methods rather than a standardized algorithm. They identified 1523 ADEs, for an incidence of 50.1/1000 person-years. Overall, 11 ADEs were fatal, 136 life threatening, and 431 serious; the remainder were described as significant. Four hundred twenty-one (27.6%) ADEs were judged to have been preventable. Among the preventable ADEs, problem areas included monitoring (60.8%), prescribing (58.4%), adherence (21.10/o), and dispensing (~2.0%). The most common medication classes involved in preventable ADEs included cardiovascular agents, diuretics, nonopioid analgesics, hypoglycemics, and anticoagulants. This study had a number of strengths, including, but not limited to, its large sample size, use of electronic screening for potential ADEs, and consensus evaluation of ADEs by physicians using case reports abstracted from patients’ charts by clinical pharmacists2 It has, however, several limitations. One concern is a potential underdetection of “true” ADEs. Table I shows the rates of ADEs from this and 2 other studies in the ambulatory elderly.9J0 The rate of -5% reported by Gurwitz et al is 2 to 7 times lower than the rates reported in the other 2 studies. This difference may be due in part to the fact
Table 1. Adverse
Author
that the latter 2 studies used a semistructured interview to determine whether participants had experienced any adverse effects, unwanted reactions, or other problems with their medications over the follow-up period. Patient self-report appears to be a more sensitive method for detecting ADEs in ambulatory patients than record review.“>12 Another possible reason for the difference is that Gurwitz et al did not screen handwritten notes or records of medication discontinuations (this can be done electronically), both sensitive and reliable approaches to screening for potential ADEs.12 Another limitation to the study by Gurwitz et al2 was discussed in an accompanying editorial, which noted a need for further refinement of the definitions of serious ADEs and preventable ADEs.13 The US Food and Drug Administration defines a serious ADE as any event due to a drug that causes death, requires or prolongs hospitalization, is permanently disabling, is life threatening, or requires intervention to prevent permanent impairment or damage. l4 Finally, it is difficult to reconcile the low proportion of ADEs rated as preventable (27.6%) by Gurwitz et al with the finding of previous studies that 95% of ADEs are type A reactions; that is, common, predictable, less serious, dose-related extensions of a drug’s usual pharmacologic effects and thus, by definition, potentially preventable.15 Two studies examined the prevalence of inappropriate prescribing in the elderly and another examined the
drug events (ADEs) in elderly outpatients
Setting
Gurwitz et al2
Multispecialty group practice in central Massachusetts
Chrischilles et al9
2 Rural Iowa counties
Hanlon et allo
Veterans Affairs general medicine clinic
Sample Size
Means of Detecting ADEs
The American Journal of Geriatric Pharmacotherapy
in selected studies.
Period of Observation
% with 21 ADE
Incidence per 1000 Person-Years
Evaluation of ADEs
No. of ADEs
Health care providers, hospital and emergency department discharge summaries, electronic medical records
Consensus evaluation by pair of physician reviewers
1523
1 year
5
50
3170
Self-report
Pharmacist evaluation
521
1 year
IO
164
167
Self-report
Pharmacist evaluation
80
1 year
35
479
>30,000
I
39
The American Journal of Geriatric Pharmacotherapv
J.T Hanion et al.
association between inappropriate prescribing and health outcomes.3-5 The first study applied drug utilization review (DUR) criteria to medications reported as being taken by respondents at the second and third in-person interviews in the Duke University Established Populations for Epidemiologic Studies of the Elderly ( EPESE).3 Information about sociodemographic characteristics, health status, access to health care, and drug use in the previous 2 weeks was collected during an in-home interview. Drug use was coded for 8 drugs or therapeutic classes-digoxin, calcium channel blockers, angiotensin-converting enzyme inhibitors, histamineareceptor antagonists, nonsteroidal anti-inflammatory drugs (NSAIDs), benzodiazepines, antipsychotics, and antidepressants-and for appropriateness by applying the DUR criteria for maximum daily dose, duplication, drug-drug interactions, duration of therapy, and drug-disease interactions. It was found that 21.0% of participants in the second in-person survey and 19.2% of participants in the third in-person survey who used 21 agent from the 8 drug categories had 21 element flagged as inappropriate. Benzodiazepines and NSAIDs were the therapeutic classes associated with the highest frequency of problems. The most common problems involved drugdisease interactions and duration of use. Two factorswhite race and arthritis-were found to have important associations with inappropriate drug use in both the cross-sectional and longitudinal multivariable analyses. The investigators concluded that inappropriate drug use is common among the community-dwelling elderly and persists over time. The Duke University EPESE is a rich data set that includes detailed information on demographic variables, health status, and access to health care. Drug-utilization information was collected in an optimal fashion through the use of structured interviews and visual verification of each medication reported as taken. The findings of the Duke University EPESE appear to have face validity. With limited geriatric-specific training during the years in which the second in-person interview of the Duke University EPESE was conducted (1989-1990), few clinicians would have identified these drugs (benzodiazepines and NSAIDs) as problematic. It is not surprising that by the third round of in-person interviews in the Duke University EPESE, use of these drugs had decreased, reflecting clinicians’ growing knowledge about older patients. The finding of an association between white race and arthritis and failure to meet the DUR criteria3 are consistent with clinical realities. White patients are more likely to be prescribed benzodi-
40
I
azepines than are those of other racial groups,16 and arthritis is frequently managed with NSAID therapy. The finding that patients with a prescription for inappropriate drugs at the second in-person interview were more likely to have a prescription for inappropriate drugs at the third in-person interview should raise questions about patient characteristics that were not collected in the study (eg, personality factors, compliance with physician recommendations), as well as questions about characteristics of the physicians prescribing these medications (eg, level of training, practice type, methods of monitoring prescription refills). Overall, this study confirmed what is known about potentially dangerous drug use in older patients during the era these data were collected. The results support the DUR process as an effective tool for identifying populations at risk for inappropriate prescribing. The second study was designed to identify the prevalence of inappropriate medication use in elderly persons living in residential care/assisted living facilities in Florida, Maryland, New Jersey, and North Carolina.4 The authors conducted a secondary analysis of data collected in the facilities by trained nurses as part of the Collaborative Studies of Long-Term Care. Inappropriate medication use in the elderly was defined using a modification of the consensus criteria reported by Beers17 that incorporated input from Dr. Beers. These authors found that 16.0% of patients were receiving 21 inappropriate medication. The most common inappropriate medications were oxybutynin, propoxyphene, amitriptyline, diphenhydrarnine, and ticlopidine. Cross-sectional multivariable analyses revealed an increased risk of inappropriate drug use in those who received a higher number of medications and at those facilities with moderate rates of turnover of licensed practical nurses. High monthly rates, moderate or severe dementia, smaller number of beds, and weekly physician visits were protective against inappropriate prescribing. The investigators concluded that inappropriate prescribing was common among residents of assisted living facilities and that regular physician visits to such facilities and attention to such highrisk groups as those taking multiple medications may improve prescribing. The finding that drugs to be avoided in the elderly were commonly prescribed for patients in assisted living facilities4 is no surprise. However, the rate of 16.0% is lower than that reported in another study (25 .0%).18 This may be due in part to exclusion of the Beers maximum daily dose criteria for short-acting benzodiazepines and digoxin from the Collaborative Studies of Long-
J. T Hanlon et al.
Term Care. One might expect use of these agents to be more prevalent in this population than many or most of the inappropriate medications identified in the study. It is also notable that facility characteristics were more common risk factors for inappropriate prescribing than were patient attributes. The third study used longitudinal multivariable analyses to assess the association between inappropriate drug use, as defined by both Beers and DUR criteria, and mortality or decline in functional statu~.~ The study used data on sociodemographic characteristics, health status, access to health care, and drug use for 3234 subjects from the Duke University EPESE. Death was identified based on the National Death Index. Change in 4 measures of fimctional status-basic self-care, intermediate self-care, complex self-care, and physical function-was determined over 3 years from baseline. No significant association was found between mortality and inappropriate drug use identified by either set of criteria. Based on the Beers criteria, there was no significant association between inappropriate drug use and decline in functional status. Based on the DUR criteria, however, the relationship between inappropriate drug use and the ability to perform basic self-care was significant (P < 0.05) and was more pronounced for drugs associated with problems of drug-drug or drug-disease interactions. The investigators concluded that identifying the impact of inappropriate drug use may depend on the criteria applied. This study compared the predictive validity of 2 criteria for inappropriate drug use in the elderly: DUR and drugs to avoid (Beers criteria).5 Mortality and functional status are clinically and epidemiologically relevant outcomes, and the population was appropriately geriatricthe majority were aged ~75 years, female, in fair to poor health, and taking multiple medications. The findings suggest, not surprisingly, that the relationship between drug use and relevant outcomes is highly patient specific. Thus, criteria of the drugs-to-avoid type may have limited usefulness. Criteria that incorporate more detailed characteristics of patients’ drug use and medical status would provide a better tool for identifying potentially harmful drug use in the elderly
MEASURING THE QUALITY OF MEDICATION USE IN THE ELDERLY Three recent articles-l from the US,6 1 from the United Kingdom,7 and 1 from Canada8-described the consensus development of new sets of quality indicators for drug use in older persons. They shared the common goal of decreasing the incidence of preventable drug-
The American Journal of Geriatric Pharmacothevapy
related morbidity (PDRM). As defined in these studies, a PDRM must have been recognizable as drug related, the treatment failure must have been foreseeable, and the causes must have been both identifiable and controllable. The US article described a list of drug-quality indicators and their relation to PDRM based on data compiled from a review of the literature and agreed by consensus of an expert panel using a 2-round Delphi technique.6 The 7-member panel included physicians, physicianadministrators, and a geriatric-specialty clinical pharmacist. Clinical indicators of PDRM were presented as associations between particular patterns of care and adverse therapeutic outcomes, For classification as a clinical indicator of PDRM, a scenario had to be well referenced, occur fairly commonly in the geriatric population, result in serious to moderate adverse outcomes, and be searchable in the electronic medical records of the health care system cooperating in the study. Only those scenarios chosen by a majority of panel members were included in the final list. Overall, the panel reached consensus on 52 clinical indicators of PDRM, including 20 drug-disease interactions, required monitoring for 22 drug classes, 1 drug-drug interaction, 4 drugs to avoid, 1 drug with a limit on duration of use, 1 drug duplication, and 3 instances of underuse of necessary medications.6 Quality indicators not covered by other groups included avoidance of typical antipsychotic agents and laboratory test monitoring for aminoglycosides, angiotensin-converting enzyme inhibitors, antidiabetic agents, antiepileptic drugs, heparin, lithium, levothyroxine, NSAIDs, theophylline, ticlopidine, troglitazone, and concomitant trimethoprim/sulfamethoxazole and methotrexate. Also unique were the criteria for underuse of necessary drugs (ie, inhaled steroids in asthmatic patients, treatment for the prevention of osteoporosis in patients taking chronic corticosteroids) and drug-disease interactions (ie, imipramine and diabetic bladder atony, and use of nitrofurantoin, nalidixic acid, or methenamine in patients with chronic renal failure j. The UK investigation also developed quality indicators for the prevention of PDRM.7 Starting with the 52 indicators developed by the US group,6 the UK group deleted 20 criteria that did not apply to or were of low prevalence in outpatient elderly persons in the UK health care system. They then conducted a questionnaire survey of a panel of 5 general practitioners and 10 pharmacists directly involved in medication review in general practice. The 2-round Delphi survey used a scale from
The American journal
of Geriatric Pharmacotherapy
J.T. Hanh
et al.
1 (definitely PDRM) to 7 (definitely not PDRM) and considered 53 individual criteria. By the end of the second round, the panel had identified 29 quality indicators. These indicators addressed 9 drug-disease interactions, 6 drug-drug interactions, 2 drugs to avoid, 3 necessary medications for the prevention of medical problems, and 9 drugs requiring monitoring. They developed unique criteria for drug-disease interactions (ie, metoclopramide and Parkinson’s disease), laboratory monitoring (ie, statins and liver function tests), and appropriate drug-administration techniques (ie, use of a spacer with corticosteroid inhalers). In the Canadian for medication
effort
to develop
quality indicators
use in the elderly,* 2 clinical pharmacists
reviewed the 52 US indicator& for relevance. Fifty-eight quality indicators were developed based on a 2-stage Delphi survey of 6 geriatricians and 6 clinical pharmacologists. A modified nominal group technique was then used to obtain feedback from 12 general practitioners, which led to the deletion of 6 quality indicators and a final set of 52 PDRMs. The clinical indicators included 7 drug-disease interactions, 4 drug-drug interactions, 9 drugs to avoid, 2 drug duplications, 10 conditions in which necessary drugs were underused, 13 drugs requiring laboratory monitoring, 2 drugs with limits to duration of use, and 5 drugs with maximum dosage limits. The group developed unique criteria for underuse (ie, influenza and pneumococcal vaccination, regularly scheduled analgesics for chronic pain), maximum initial dosing of individual drugs (ie, levodopa/carbidopa and risperidone), drugs to avoid (ie, multiple psychotropic agents), drug duration (ie, risperidone for delirium after hospitalization), and drug-disease interactions (ie, neuroleptics and Parkinson’s disease). The new clinical indicators for PDRM developed by these 3 investigative groups built on indicators compiled by earlier groups. 17~19-22It is notable that the recent groups did not consider the Centers for Medicare & Medicaid Services DUR guidelines concerning drugtherapy monitoring indicators for long-term care facilities22 or the Study of Clinically Relevant Indicators for Pharmacologic Therapy. 2o It is also interesting that each of these studies used a different method for reaching consensus. Moreover, only 4 drug-disease interactions, 3 drugs requiring monitoring, 2 drug-drug interactions, 1 drug duplication, and 1 drug class to avoid appeared on all 3 lists (Table II). Finally, these groups did not provide the evidence on which their recommendations for baseline laboratory monitoring of drugs and timing of follow-up monitoring were based.
42
I
Table II. Quality indicators appearing in each of 3 sets of consensus guidelines for drug use in older personsm Drugs
to avoid
Long half-life Drug-disease
benzodiazepines interactions
in patients with COPD in patients with heart failure NSAlDs and steroids in patients with peptic ulcer disease Beta-blockers NSAlDs
Drug-drug interactions Antibiotics and warfarin NSAlDs and warfarin Drug duplication 22 NSAlDs Required monitoring INR for warfarin Thyroid function tests in patients taking levothyroxine Electrolytes in patients taking ACE inhibitors COPD = chronic obstructive pulmonary disease; NSAlDs = nonsteroidal anti-inflammatory drugs; INR = international normalized ratio; ACE = angiotensin-converting enzyme.
CONCLUSIONS Drug-related problems, including ADEs and inappropriate drug use, are common in elderly patients. Although new sets of quality indicators for prescribed medications have been developed by groups in 3 countries, there is minimal overlap between the sets. Future research should examine whether the identified negative outcomes are associated with failure to comply with quality indicators for the prescribed agents. Studies are also needed to develop prescribing guidelines that take into account directions for use, indications, and effectiveness of therapy in the frail elderly. Finally, it would be helpful to be able to identify those at high risk for ADEs a priori and to test interventions to reduce suboptimal drug use and ADEs.
REFERENCES 1. Strand LM, Morley PC, Cipolle RJ, et al. Drug-related problems: Their structure and function. DICP
1990;24:
1093-1097. 2. Gurwitz JH, Field TS, Harrold LR, et al. Incidence and preventability of adverse drug events among older persons in the ambulatory setting. JAMA. 2003;289:1107-1116. 3. Hanlon JT, Schrnader KE, Boult C, et al. Use of inappropriate prescription drugs by older people. JAm Geriatr Sot.
2002;50:26-34.
J.T Hahn
4. Sloane PD, Zimmerman S, Brown LC, et al. Inappropriate medication prescribing in residential care/assisted living facilities. JAm
Geriatr Sot. 2002;50:1001-1011.
et al.
The American Journal of Geriatric Pharmacotherapy
14. Kessler DA. Introducing MEDWatch. A new approach to reporting medication and device adverse effects and product problems. JAMA.
1993;269:2765-2768.
5. Hanlon JT, Fillenbaum GG, Kuchibhatla M, et al. Impact
15. Venulet J, ten Ham M. Methods for monitoring and doc-
of inappropriate drug use on mortality and functional status in representative community dwelling elders. Med
umenting adverse drug reactions. Int J Clin Pharmacol Ther. 1996;34:112-129.
Care. 2002;40:166-176. 6. MacKirmon NJ, Hepler CD. Preventable drug-related mor-
16. Blazer D, Hybels C, Simon& E, Hanlon JT. Sedative, hypnotic and anti-anxiety medication use in an aging
bidity in older adults: Indicator development. J Managed
cohort over ten years: A racial comparison. JAm Gtiatr Sot.
Care Barr.
2000;48:1073-1079.
2002;8:365-371.
7. Morris CJ, Canttill JA, Hepler CD, Noyce IX
Preventing
17. Beers MH.
Explicit criteria for determining
potentially
drug-related morbidity~etermining valid indicators. Int JQuaLity Health Care. 2002;14:183-198.
inappropriate medication use by the elderly. An update.
8. Robertson HA, MacKinnon NJ. Development of a list of consensus-approved clinical indicators of preventable
18. Spore DL, Mor V, Larrat I?, et al. Inappropriate drug pre-
drug-related morbidity in older adults. Clin Ther. 2002;
scriptions for elderly residents of board and care facilities.
Am J Ptlblic Health. 1997;87:404-409. 19. Knight EL, Avorn J. Quality indicators for appropriate
24:1595-1613. 9. Chrischilles
Arch Intern Med. 1997;157:1531-1536.
EA, Segar ET,
Wallace BB.
Self-reported
adverse drug reactions and related resource use. A study of community-dwelling
persons 65 years of age and older.
medication use in vulnerable elders. Ann Intern Med. 2001; 135:703-710. 20. Study of Clinically Relevant Indicators for Pharmacologic Therapy (SCRIPT).
Ann Intern Med. 1992;117:634-640. 10. Hanlon JT, Schmader KE, Koronkowsld MJ, et al. Adverse
drug events in high risk older outpatients. J Am Geriati
Available at: http://scriptproject.org/
home.html. Accessed May 23,2003. 21. McLeod PJ, Huang AB, Tamblyn BM, Gayton DC. Defining inappropriate practices in prescribing for elderly
Sot. 1997;45:945-948. 11. Gandhi TK, Weingart SN, Borus J, et al. Adverse drug events in ambulatory care. NE& 1564.
JMed. 2003;348:1556-
people: A national consensus panel. CM@
1997;156:385-
391. 22. Appendix N. Surveyor procedures for pharmaceutical serin long-term
care facilities. In: State
methods for detecting potential adverse drug events in frail elderly inpatients and outpatients. Am J Health Syst
Operations Manual, Provider Health and Human Services,
Certification. Dept of Health Care Financing
Pharm. 2001;58:1622-1626.
Administration.
12. Hanlon JT, Maher B, Lindblad C, et al. Comparison of
13. Classen D. Medication
safety: Moving from illusion to
reality. JA.MA. 2003;289:1154-1156.
Editorial.
vice requirements
Transmittal no. 242,
September
1990.
Available at: http://www.ascp.com/public/pr/2OOl/pdfs/ App_N.pdf Accessed May 23,2003.
Address correspondence to: Joseph T. Hanlon, PharmD, MS, Institute for the Study of Geriatric Pharmacotherapy, College of Pharmacy, 7-l 15 Weaver-Densford Hall, University of Minnesota, 308 Harvard Street SE, Minneapolis, MN 55455. E-mail: [email protected]
I
43