Jgs50253

Inappropriate Medication Prescribing in Residential
Care/Assisted Living Facilities
Philip D. Sloane, MD, MPH,* Sheryl Zimmerman, PhD,* Lori C. Brown, PharmD,‡Timothy J. Ives, PharmD, MPH,† and Joan F. Walsh, PhD* OBJECTIVES: To identify the extent to which inappro-
estimating equations, IPM use was associated with the priately prescribed medications (IPMs) are administered to number of medications received, smaller facility bed size, older patients in residential care/assisted living (RC/AL) moderate licensed practical nurse turnover, absence of de- facilities and to describe facility and resident factors asso- mentia, low monthly fees, and absence of weekly physi- ciated with receipt of one or more IPMs.
DESIGN: Cross-sectional study of a stratified, representa-
CONCLUSIONS: IPMs remain a problem in long-term
tive sample of 193 facilities in four states.
care, but rates in these RC/AL settings compare favorably SETTING: We identified representative geographic re-
with those reported for other frail older populations, sug- gions within Florida, New Jersey, North Carolina, and gesting that use of medications with severe adverse effects Maryland and drew from within them a stratified random may be waning. Regular physician facility visits may im- sample of 193 RC/AL facilities. Three subtypes of facilities prove prescribing, as will attention to high-risk groups were included in the sample: small homes (Ͻ16 beds), such as individuals on multiple medications. J Am Geriatr
larger “new-model” homes, and larger “traditional” homes.
Soc 50:1001–1011, 2002.
PARTICIPANTS: Within each larger home, a random
Key words: medications; assisted living; long-term care
sample of residents aged 65 and older was approached for
consent; in smaller homes all residents were approached.
The overall enrollment rate was 92%; 2,078 residents
were enrolled.
MEASUREMENTS: Questionnaires and on-site observa-
Adverse drug events are the most common medical er- tions were used to gather data on facility administration ror occurring in the United States today.1 Although and staffing and resident characteristics. All prescription older persons represent less than one-fifth of the U.S. pop- and nonprescription medications taken at least 4 of the 7 ulation, they use more than one-third of all prescription days before data collection were taken from medication medications dispensed.2 As many as 18% of all outpatient administration records and coded for analysis. IPM desig- visits involve drug complications,3 which are implicated in nation was based on modification of a list developed by 6% to 21% of older outpatient visits.4,5 Between 18% and Beers et al. and currently used by nursing home surveyors.
24% of admissions of hospitalized older patients are at- RESULTS: The majority of RC/AL patients were taking
tributable to adverse drug events. Persons aged 65 and five or more medications; 16.0% of these patients were re- older are particularly susceptible to adverse drug events ceiving IPMs. The most common IPMs were oxybutynin, because of high rates of medication use and physiological propoxyphene, diphenhydramine, ticlopidine, doxepin, and changes associated with aging. These factors are accentu- dipyridamole. In multivariate analyses, using generalized ated in long-term care facilities, where polypharmacy iscommon and the reported rates of adverse drug events areas high as 67% to 74%.8,9 Medication selection is an important factor influencing the likelihood of adverse drug events. Advances in therapeu- From the *Sheps Center for Health Services Research, and †School of Phar- tics require that physicians update their prescribing prac- macy and Department of Family Medicine, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; and ‡Kerr tices when safer, superior alternatives to existing products Drug EPCC, Greensboro, North Carolina.
become available. In addition, changes in patient medical Financial support was provided by National Institute on Aging Grants R01 status over time can cause medications that have been used chronically to become unsafe or ineffective. In clinical prac- Address correspondence to Philip D. Sloane, MD, MPH, Sheps Center for tice, drug selection involves a variety of biomedical and psy- Health Services Research, University of North Carolina at Chapel Hill, 725 chosocial factors. For example, because of the limited fi- Airport Road, CB 7590, Chapel Hill, NC 27599. E-mail: psloane@med. unc.edu nances of many geriatric patients, physicians often choose 2002 by the American Geriatrics Society SLOANE ET AL.
older, less-expensive medications rather than newer, more- gion fall within 30% of the state mean on eight demo- targeted alternatives.10 Thus, medication prescribing in- graphic and health services measures (per capita income; volves a process whereby the physician considers multiple percentage of population aged Ն65, nonwhite, employed, possible agents and selects one based on the individual pa- and below poverty level; ratio per 1,000 persons aged Ն65 tient’s needs, medical status, and available resources.
to primary care physicians, hospital beds, nursing home Nevertheless, certain medications are rarely if ever in- beds). Across all four states, only one variable deviated dicated for older persons, because they are ineffective or from the 30% limit in one state, and examination of that because safer, effective alternatives exist. For example, al- region indicated adequate dispersion of counties over high pha-methyldopa and reserpine were once acceptable antihy- pertensives, but newer agents have supplanted them. In the Within each of the four regions, all licensed RC/AL fa- early 1990s, the term inappropriate medications was intro- cilities were identified using state licensure lists. The study duced by Beers et al. to describe such drugs whose use in definition of RC/AL included any licensed facility, not li- older persons was no longer recommended.11,12 In 1999, the censed as a nursing home, that provided room and board, Health Care Financing Administration (HCFA) incorpo- 24-hour supervision, and assistance with activities of daily rated a modification of this “potentially inappropriate med- living. To maximize efficiency in enrolling older subjects, ication” list into nursing home survey criteria.13 Investiga- facilities that primarily served persons with mental retar- tors have used modifications of the Beers list to report on dation or developmental disabilities were excluded, as potentially inappropriate medication use in skilled nursing were facilities with fewer than 16 beds that housed fewer facilities14 and by community-dwelling older people15 and than four residents aged 65 and older and facilities with homebound managed care plan participants.16 16 or more beds that housed fewer than 10 residents aged One healthcare sector in which inappropriately pre- scribed medicine (IPM) use has received little attention is Eligible facilities were then divided into three strata to residential care/assisted living (RC/AL). Spore et al. showed adequately sample the range of facility types: small homes that between 20% and 25% of residents in a 10-state sam- (Ͻ16 beds); new-model large facilities (Ն16 beds; built af- ple of board and care homes had at least one inappropri- ter January 1, 1987, and fulfilling at least one of the fol- ate prescription, raising concern about drug interactions lowing criteria: at least two different private pay rates,based on resident’s service needs; 20% or more of resident and adverse effects.17 This is of special concern because the population requiring assistance in transfer; 25% or more number of RC/AL facilities is growing rapidly; by 2005, it of resident population incontinent daily; and a registered is estimated that more persons will be housed in RC/AL fa- nurse (RN) or a licensed practical nurse (LPN) on duty cilities than in nursing homes.18 Compounding this, RC/AL around the clock); and traditional large homes (Ն16 beds facilities serve primarily older persons, a population that is and not meeting the criteria for new-model facilities). One typified by multiple disease states, polypharmacy, altered hundred thirteen small facilities, 40 new-model, and 40 pharmacokinetics, and a high prevalence of drug-related traditional facilities were enrolled in the four study states.
adverse events.19,20 In addition, the staff in RC/AL facilities The facility refusal rate was 41%, but 44 nonrespondents who administer medications are generally not nurses, and (90% of those selected for sampling) completed a tele- many have little or no training in medication administra- phone survey, and participating and nonparticipating fa- tion and effects. Finally, compared with nursing homes, cilities were found to not differ in age; size; occupancy; orresident age, race, or ethnicity. Nonparticipating RC/AL RC/AL facilities have less oversight by registered nurses facilities differed from participating facilities in three of 42 items queried (hours worked by owners, number of rate The analysis reported in this paper was performed to determine the prevalence of IPM use among a representa- Within each study home, a representative sample of tive sample of more than 2,000 RC/AL residents in four residents was enrolled as follows. In small homes, all resi- states. In this study, the prevalence and distribution of IPMs dents aged 65 and older were approached for participa- are described, and resident and facility factors are identified tion. In large facilities, random sampling of residents aged that are associated with receipt of one or more IPMs.
65 and older was used to achieve target sample sizes(range 17–23 depending on stratum and state). Informed consent was obtained from participating staff and resi-dents; proxies gave written consent for cognitively im- paired residents, and the residents were required to assent Data reported in this paper were collected as part of the to in-person data collection. Two thousand seventy-eight Collaborative Studies of Long-Term Care (CS-LTC), a RC/AL residents were enrolled: 665 in small facilities, 765 multistate study of RC/ALs. The study was conducted in in new model facilities, and 648 in traditional facilities.
four states: Florida, Maryland, New Jersey, and North The subject refusal rate was 8%. The Institutional Review Carolina. The study states were chosen because each had a Boards of the University of North Carolina at Chapel Hill well-developed RC/AL industry, and the four represented and the University of Maryland at Baltimore approved a range of state regulatory approaches to the development subject enrollment and data collection procedures. Further of newer assisted living models. To increase efficiency of details of the methods of the Collaborative Studies of data collection, a purposive sample of counties (a sam- Long-Term Care are published elsewhere.21 pling region) was selected within each state. Criteria forselection of a sampling region included that it contain at Measures, Data Collection, Coding, and Analysis
least 15% of the state’s residential care facilities, that it On-site interviewers, the majority of whom were RNs, contain urban/suburban and rural areas, and that the re- conducted baseline data collection between October 1997 INAPPROPRIATE MEDICATIONS IN RESIDENTIAL CARE
and November 1998. Questionnaires on the structure and mines were excluded because they are considered at times process of care were administered to facility staff, and ad- to be appropriate treatment for allergies; (4) oral iron ditional facility data were gathered using direct observa- preparations and digoxin were excluded because their pre- tion. Resident data were gathered using a combination of scription’s appropriateness on Beer’s list depended on dos- record review, patient or proxy interview, and direct ob- age and the CS-LTC data did not include drug dosages; (5) short-acting benzodiazepines (oxazepam, triazolam, loraze- Data on the following facility factors were obtained pam, temazepam, alprazolam, and zolpidem) were ex- during interviews with facility administrators: bed size, cluded because they are widely used for agitation in de- profit/nonprofit status, affiliations with other facilities, the mentia and for short-term treatment of insomnia in some existence of separate levels of care in the same building, ser- older people; they have a short half-life, leading to few vices provided, monthly fees, staff hours actually worked in problems with hangover and accumulation; and dosage the previous week by nurses (LPNs and RNs) and nursing was considered in the original criteria; and (6) trazodone assistants, and turnover rates of LPNs, RNs, and nursing was excluded because it is appropriately used in low doses assistants. Data on the following resident characteristics for sedation and agitation. The final list of IPMs used in were gathered by interviews with the residents and the staff the study, and recommended alternative medications, is caregivers who knew the residents best: age, race, gender, frequency of visits from friends/family, presence of moder- Staff of the Cecil G. Sheps Center for Health Services ate/severe dementia, payment source (e.g., Medicaid), and Research Data coded, entered, cleaned, and analyzed data dependency in six activities of daily living (eating, toileting, using standard procedures and analytical software pack- transferring, locomotion, dressing and bathing). Moderate ages. To identify factors associated with resident’s receiv- or severe dementia was determined to be present if the ing one or more drugs on the IPM list, associations were Mini-Mental State Examination (MMSE)22 score was below studied between facility and resident variables known to 17, (when the MMSE was not available) if the Minimum affect care provision in long-term care settings. The fol- Data Set Cognition Scale23 score was greater than 3, or (in lowing facility factors were studied as predictors of IPM rare cases where neither was available) if a physician or use: facility type (small facilities were the reference group), nurse noted one or more of the following diagnoses on the bed size, square of bed size, profit/nonprofit status, nurs- medical record: Alzheimer’s disease, senile dementia, senile ing home affiliation, whether a physician visited at least dementia of the Alzheimer’s type, organic brain syndrome, weekly, whether nursing services were provided at least cerebral arteriosclerosis, multiinfarct dementia, subcortical weekly, minimum monthly fee, nurse/resident ratio, nurs- dementia, Binswanger’s disease, Pick’s disease, Creutzfeldt- ing assistant/resident ratio, RN turnover, LPN turnover, Jakob disease, Huntington’s disease, Lewy body disease, or and nursing assistant turnover. The latter six variables were trichotomized into low (Ͻ25th percentile), moderate As part of each subject’s on-site data collection, re- (25–75th percentile), and high (Ͼ75th percentile) and en- search staff reviewed the Medication Administration Rec- tered into the regression as dummy variables, with the ord and wrote down the names of all prescription and lowest percentile as the reference group. The following nonprescription medications that had been administered resident factors were included in the regression: dementia to the subject on at least 4 of the previous 7 days. Informa- (moderate or severe), race, gender, age (as four categories), tion on dosage was not gathered. Drug names from the frequency of family/friend visits (moderate ϭ 2–6 days/2 data collection forms were entered verbatim into data en- weeks; high ϭ Ն7 days/2 weeks), Medicaid or state assis- try fields. The resulting data files were cleaned and coded tance (yes/no), dependency in activities of daily living using an existing program to correct misspellings and to (moderate ϭ assistance needed with one or two; heavy ϭ code for recognized drugs using the American Hospital assistance needed with three or more), and number of Formulary Service system.24 A pharmacist (LCB) and a medications (moderate ϭ 4–8; high ϭ Ն9). Bivariate asso- geriatrician (PDS) reviewed each remaining uncoded drug ciations were studied using t tests (for continuous vari- to determine what medication was represented and to as- ables) or chi-square tests (for categorical variables).
sign a code. Of the 2,078 RC/AL residents in the study To identify the relative contribution of resident and fa- sample, 64 (3.1%) had data containing one or more medi- cility factors to the likelihood of a resident receiving one or cations that could not be coded because of illegibility or more IPMs, multivariate regression was performed, using misspellings. This paper reports on the 2,014 subjects for generalized estimating equations to control for the cluster- whom complete medication data were available.
ing effects of the study sample. Analyses were performed us- IPMs were coded using an updated version of the list ing PROC GENMOD in Statistical Analysis Systems, which developed by Beers et al.11,12 (Dr. Beers was consulted accounts for intrafacility correlation while weighting each about revising the list during the course of our analyses subject equally.25 The dichotomous dependent variable was (March 21, 2000); he encouraged revision to reflect whether a given resident’s medication list included one or changes in pharmacotherapy.) For purposes of these anal- more IPMs. The regression analysis excluded 29 residents yses, the following medications were excluded from the who were on no medications and 64 residents for which Beers list:12 (1) flecainide, phenylbutazone and cyclande- one or more medications could not be coded.
late were not included because they are no longer mar-keted; (2) haloperidol and thioridazine were excluded be-cause they may be appropriate for some indications, even though they may cause sedation, extrapyramidal effects, Table 1 presents the characteristics of the 193 study facili- and sedation in some patients; (3) nonsedating antihista- ties and 2,014 subjects used in these analyses. The major- SLOANE ET AL.
ity of facilities (58.5%) were small, but the number of sub- 4.0% in traditional homes. The most common categories jects was relatively equally divided between the three of medications received across all strata were cardiovascu- facility types (31.8% in small, 37.0% in new-model, and lar drugs (received overall by 53% of subjects); diuretics/ 31.1% in traditional). Facility staffing data indicate a de- potassium (40%); laxatives/antacids (37%); vitamins/min- pendency on unlicensed staff, with few nursing hours per erals (37%); pain medications such as nonsteroidal anti- week (0.6 hours per resident for RNs and 1.2 hours per inflammatory medications, aspirin, and acetaminophen resident for LPNs). Residents tended to be female, very old (over half were Ն85), and impaired in at least one activity Three hundred sixty-nine of the 11,649 prescriptions of daily living and to pay privately.
in the sample (3.2%) were IPMs. Across the three types of Most study subjects had taken at least one medication homes, the percentage of prescriptions that involved IPMs on 4 or more of the 7 days before data collection. The was similar: 3.3% in small homes, 3.2% in new-model mean number of medications taken regularly was 5.1 in homes, and 2.9% in traditional homes. Three hundred small homes, 6.1 in new-model homes, and 5.6 in tradi- twenty-two (16.0%) of the 2,014 study subjects received tional homes. Few residents were on no medications: at least one IPM, ranging from 15.5% in traditional 6.0% in small homes, 3.4% in new-model homes, and homes to 16.9% in new-model homes (Table 2). The most Table 1. Descriptive Characteristics of the Study Sample
Staffing ratio (expressed as weekly hours per resident—census at data collection) Turnover (per 6 months: n left/n current FTE) *Excludes observations that had missing data for specific variables.
†Fewer than 16 beds.
RC/AL ϭ residential care/assisted living; ADL ϭ activities of daily living; FTE ϭ full time equivalent.
INAPPROPRIATE MEDICATIONS IN RESIDENTIAL CARE
common IPMs were oxybutynin, propoxyphene, amitrip- older people received a potentially inappropriate drug. Al- tyline, ticlopidine, doxepin, and dipyridamole (Table 3).
though any generalization based on these results must be Table 4 displays the results of the generalized estimat- made with caution, these findings are reassuring in that ing equation regression of facility and resident factors on they suggest that RC/AL settings do not have higher rates the probability of having one or more IPMs. The facility than other settings, in spite of employing fewer nurses.
factors that were independently associated with an in- As with all studies of medication use in long-term creased probability of a resident being on an IPM included care, potential sources of error exist. One is the potential smaller bed size, low minimum monthly fees for residents, that over-the-counter, complementary/alternative, and other moderate LPN turnover, and absence of a weekly physi- medications were brought in by families and administered cian visit. Resident factors associated with an increased to residents without being recorded on the medication ad- probability of being on an IPM were number of medica- ministration record. No estimate of the extent of this miss- tions received and absence of moderate/severe dementia.
ing data is available, but data collectors who interviewedresidents as part of the study did not feel that this consti- DISCUSSION
tuted a problem. Another source of missing data is the The data presented here indicate that polypharmacy is 3.1% of medication that could not be coded because of il- prevalent in RC/AL facilities. The majority of residents in legibility or misspellings; this is higher than the 1.6% rate this sample were taking at least five medications, and use of missing data reported by Hanlon et al. in a community of 10 or more medications was not unusual. Furthermore, study.26 In addition, this study assessed appropriateness in according to an updated version of Beer’s “potentially in- terms of efficacy only, because data on indication, dose, appropriate” medication list, 16.0% of study subjects duration, comorbid diseases, and potential interactions were receiving one or more IPMs, and between 2.9% and with other medications were not collected. Thus, this 3.3% of RC/AL medications fell into the IPM category.
study likely underestimated the presence of “inappropri- Although these numbers are not unusual for a group of ate” medications in RC/AL facilities. Finally, the cross-sec- older patients, the normalcy of such numbers should not tional design limits the ability to derive causal inferences.
Can further reductions be achieved? Multivariate Strict comparison of these findings with other pub- analyses to identify facility and resident factors associated lished studies is not possible for a number of reasons.
with IPMs provide insight into potential intervention tar- First, other published studies have used somewhat differ- gets. As noted in Table 4, the strongest predictor of IPM ent criteria for inappropriateness; in this study, modifica- use is the number of medications a patient receives. This tion was made to reflect prescribing practices at the time finding is intuitive, but it emphasizes that quality monitor- of the study and to account for the absence of dosage in- ing efforts should concentrate on patients with the longest formation. Second, the potential for secular trends to in- medication lists. The finding that frequent physician visits fluence results exists because of differences in the dates of were associated with fewer inappropriate prescriptions data collection across studies. Finally, the extent to which suggests that quality may be improved by encouraging nonprescription drugs were included in published stud- stronger linkages between RC/AL facilities and physicians ies is unclear, but the rate of potentially inappropriate who make regular visits. The association between moder- medication use from this study appears lower than that ate LPN turnover and high levels of inappropriate pre- previously reported in board and care homes17 and in scriptions suggests that minimal levels of nursing oversight homebound older people,16,26 suggesting that use of these may be inadequate, but the lack of a consistent association medications is decreasing or that the RC/AL facilities sur- between other nursing turnover, or of nurse staffing levels, veyed expose their residents less frequently to these drugs.
and inappropriate medication use suggests the need for Thus, when analyses are adjusted to remove medications further research in this area. Finally, increased oversight not included in this study’s list, the findings of Golden et by consultant pharmacists, although not assessed in this al.16 showed that 8% of prescriptions used by homebound study, may be able to further reduce IPMs. In nursing older people were inappropriate, of Beers et al.14 that homes, consultant pharmacists review medications monthly, 4.9% of nursing home prescriptions were inappropriate; but for RC/AL facilities the timing and extent of pharmacy and of Wilcox et al.15 that 23.5% of community-dwelling review varies and is generally less frequent.
Table 2. Frequency and Number of Inappropriately Prescribed Medications (IPMs), by Facility Type
RC/AL ϭ residential care/assisted living.
SLOANE ET AL.
Table 3. Number of Study Subjects (N ؍ 2,014) Regularly
Table 4. Facility and Resident Factors Influencing Likelihood
Receiving Each Inappropriately Prescribed Medication (IPM),
of a Residential Care/Assisted Living Resident Having One or
by Facility Type
More Inappropriately Prescribed Medications
Note: The following medications on the “potentially inappropriate” list were not received by any study subjects: belladonna, buprenorphine, butorphanol, chlor-propamide, chlorzoxazone, dezocine, isoxsuprine, meperidine, meprobamate, me- thocarbamol, minoxidil, nalbuphine, and reserpine.
RC/AL ϭ residential care / assisted living.
Other associations noted in the multivariate analyses are more difficult to interpret. Facility size and monthly fee may have indirect effects as proxies for resources such as nursing oversight. The fact that increases in the mini- mum monthly fee are associated with parallel decreases in the likelihood of IPMs suggests that some kind of socio- economic effect is present, a finding that was also identi-fied by Wilcox et al.15 The finding of an independent rela- Note: Analysis was performed using Generalized Estimating Equations (GEE).
tionship between dementia and absence of IPMs is puzzling; *Odds ratios are adjusted for all other variables in the model. The scale parameterfor GEE was computed as the square root of the normalized Pearson’s chi-square.
perhaps it arose because some common IPMs are used to RNϭ registered nurse; LPN ϭ licensed practical nurse; PCA ϭ patient care assis- treat conditions or symptoms rarely voiced by persons with tant; ADL ϭ activities of daily living.
dementia. For example, this may be true of propoxyphenebecause persons with dementia tend to request and receivefewer pain medications than persons with similar condi- known. In the CS-LTC study, observations of patient som- nolence at a standard time in the midafternoon were not The extent to which these “potentially inappropriate” correlated with IPM use (␹2 ϭ 1.35, P ϭ .24). Other po- medications result in adverse resident outcomes is un- tential outcomes, such as hospitalization, morbidity, mor- INAPPROPRIATE MEDICATIONS IN RESIDENTIAL CARE
tality, and disability will be studied in a longitudinal fol- 8. Cooper JW. Probable adverse drug reactions in a rural geriatric nursing low-up of the cohort, but, given the distribution of the use home population: A four-year study. J Am Geriatr Soc 1996;44:194–197.
9. Cooper JW. Adverse drug reactions and interactions in a nursing home. Nurs of individual medications in the study (Table 3), it is un- likely that this list constitutes a strong predictor of adverse 10. Beers MH, Fingold SF, Ouslander JG et al. Characteristics and quality of pre- scribing by doctors practicing in nursing homes. J Am Geriatr Soc 1993;41: Physicians’ prescribing patterns are changing, and 11. Beers MH, Ouslander JG, Rollinger I et al. Explicit criteria for determining many of yesterday’s “inappropriate” medications (those inappropriate medication use in nursing home residents. Arch Intern Med with the most severe adverse drug events) have been elimi- nated from practice through manufacturer’s withdrawal, 12. Beers MH. Explicit crieria for determining potentially inappropriate medica- regulatory efforts, or voluntary changes in physician pre- tion use by the elderly: An update. Arch Intern Med 1997;157:1531–1536.
13. Eans TL. New HCFA drug-prescribing criteria for nursing homes and sug- scribing. In addition, the pharmaceutical industry has re- gested alternate prescribing to avoid care deficiencies. Ann Long-Term Care sponded in some cases to toxic drug effects by reformulat- ing their products. For example, one of the most frequently 14. Beers MH, Ouslander JG, Fingold SF et al. Inappropriate medication pre- used “inappropriate” medications on this study’s list is scribing in skilled nursing facilities. Ann Intern Med 1992;117:684–689.
15. Wilcox SM, Himmelstein DU, Woolhandler S. Inappropriate drug prescrib- oxybutynin. Since the time of data collection, oxybutynin ing for community dwelling elderly. JAMA 1994;272:292–296.
has been reformulated as a sustained-release product, with 16. Golden AG, Preston RA, Barnett SD et al. Inappropriate medication prescrib- evidence of reduced adverse effects.29,30 ing in homebound older adults. J Am Geriatr Soc 1999;47:948–953.
Given the possibility that the most clearly “inappro- 17. Spore DL, Mor V, Larrat P et al. Inappropriate drug prescriptions for elderly residents of board and care facilities. Am J Public Health 1997;87:404–409.
priate” medications are prescribed less frequently, regula- 18. Meyer H. The bottom line on assisted living. Hosp Health Network 1998;7: tory efforts may need to refocus, as is expected when a quality improvement effort is successful.31 Thus, facility 19. Williams L, Lowenthal DT. Drug therapy in the elderly. South Med J 1992; medical directors, consultant pharmacists, and HCFA sur- 20. Blanchette K. New Directions for State Long-term Care Services, Volume III: veyors may achieve little clinical improvement by trying to Supportive Housing. Washington DC: Public Policy Institute, American As- eliminate medications whose appropriateness depends on dose, duration, or indication. Instead, a broader examina- 21. Zimmerman SI, Sloane PD, Eckert JK et al. Overview of the collaborative tion of medication use may be needed and other indicators studies of long-term care. In: Zimmerman SI, Sloane PD, Eckert JK, eds. As-sisted Living: Long-Term Care in Transition. Baltimore, MD: Johns Hopkins of prescribing problems sought. Measures that may be helpful include patient-adjusted doses (to account for pa- 22. Folstein MF, Folstein S, McHugh PR. “Mini-Mental State”. A practical tient factors, such as weight and renal function); frequency method for grading the cognitive state of patients for the clinician. J Psychiatr of administration; presence of pharmacotherapy without 23. Hartmaier SL, Sloane PD, Guess HA et al. The MDS Cognition Scale: A valid indication; the presence of possible adverse effects; and use instrument for identifying and staging nursing home residents with dementia of over-the-counter medications, social drugs (e.g., alco- using the minimum data set. J Am Geriatr Soc 1994;42:1173–1179.
hol), and complementary/alternative therapies. In addi- 24. McEvoy JK. AHFS Drug Information 2001. Bethesda, MD: American Soci- tion, underprescribing may be emerging as a significant ety of Health-System Pharmacists, 2001.
25. Stokes ME, Davis CS, Koch GG. Categorical Data Analysis Using the SAS problem as more therapeutic options are available for a System. Cary, NC: SAS Institute Inc., 2000.
variety of chronic conditions. The most critical issues in 26. Hanlon JT, Fillenbaum, GG, Schmader KE et al. Inappropriate drug use pharmacotherapeutics in long-term care change over time, among community-dwelling elderly. Pharmacotherapy 2000;20:575–582.
and monitoring efforts must account for these changes.
27. Cariaga J, Burgio L, Flynn W et al. A controlled study of disruptive vocaliza- tions among geriatric residents in nursing homes. J Am Geriatr Soc 1991;39:501–507.
ACKNOWLEDGMENTS
28. Sengstaken EA, King SA. The problems of pain and its detection among geri- The Collaborative Studies of Long-Term Care involves a atric nursing home residents. J Am Geriatr Soc 1993;41:541–544.
team of investigators, data collectors, and support staff 29. Gleason DM, Susset J, White C et al. Evaluation of a new once-daily formu- lation of oxbutynin for the treatment of urinary urge incontinence. Ditropan from the University of North Carolina at Chapel Hill and XL Study Group. Urology 1999;54:420–423.
the University of Maryland at Baltimore. Special apprecia- 30. Goldenberg MM. An extended-release formulation of oxybutynin chloride tion is extended to the staff and residents of the cooperat- for the treatment of overactive urinary bladder. Clin Ther 1999;21:634–642.
31. Schnelle JF, Ouslander JG, Osterweil D et al. Total quality management: Ad- ministrative and clinical applications in nursing homes. J Am Geriatr Soc1993;41:1258–1266.
32. Hass WK, Easton JD, Adams HP Jr et al. A randomized trial comparing REFERENCES
ticlopidine hydrochloride with aspirin for the prevention of stroke in high- 1. Kohn LT, Corrigan JM, Donaldson MS, eds. To Err Is Human: Building a risk patients. Ticlopidine Aspirin Study Group. New Engl J Med 1989;321: Safer Health System. Washington, DC: National Academy Press, 2000.
2. Mueller C, Schur C, O’Connell J. Prescription drug spending: The impact of 33. Gent M, Glakely JA, Ellis DJ et al. The Canadian American Ticlopidine age and chronic disease status. Am J Public Health 1997;87:1626–1629.
Study (CATS) in thromboembolic stroke. Lancet 1989;1:1215–1220.
3. Gandhi TK, Burstin HR, Cook EF et al. Drug complications in outpatients. J 34. Steinhubl SR, Tan WA, Foody JM et al. Incidence and clinical course of thrombotic thrombocytopenic purpura due to ticlopidine following coronary 4. Schneider JK, Mion LC, Frengley JD. Adverse drug reactions in an elderly stenting. EPISTENT Investigators. Evaluation of Platelet IIb/IIIa Inhibitor for outpatient population. Am J Hosp Pharm 1992;49:90–96.
5. Veehof LJG, Stewart RE, Meyboom-de Jong B et al. Adverse drug reactions 35. Tisdale JE. Antiplatelet therapy in coronary artery disease: Review and up- and polypharmacy in the elderly in general practice. Eur J Clin Pharmacol date of efficacy studies. Am J Health Syst Pharm 1998;55:S8–S16.
36. Green D, Podrazik P. Key references: Antiplatelet drugs in the elderly. J 6. Mannesse CK, Derkx FHM, de Ridder MAJ et al. Contribution of adverse Thromb Thrombolysis 2000;9:115–116.
drug reactions to hospital admission of older patients. Age Ageing 2000;29: 37. Hogan DB. Revisiting the O complex: Urinary incontinence, delirium and polypharmacy in elderly patients. Can Med Assoc J 1997;157:1071–1077.
7. von Renteln-Kruse W, Thiesemann N, Thissemann R et al. Does frailty predis- 38. Katz WA. Pharmacology and clinical experience with tramadol in osteoar- pose to adverse drug reactions in older patients? Age Ageing 2000;29:461–462.
SLOANE ET AL.
39. Forman WB. Opioid analgesic drugs in the elderly. Clin Geriatr Med 1996; 58. Kadowaki T, Hagura R, Kajinuma H et al. Chlorpropamide-induced hy- ponatremia: incidence and risk factors. Diabetes Care 1983;6:468–471.
40. Miller RR. Evaluation of the analgesic efficacy of ibuprofen. Pharmacother- 59. Tanay A, Firemann Z, Yust I et al. Chlorpropamide-induced syndrome of inap- propriate antidiuretic hormone secretion. J Am Geriatr Soc 1981;29:334–336.
41. Pearce V, Robson PJ. Double-blind crossover trial or oral meptazinol, penta- 60. Bassotti G, Whitehead WE. Biofeedback, relaxation training, and cognitive zocine and placebo in the treatment of pain in the elderly. Postgrad Med J behaviour modification as treatments for lower functional gastrointestinal disorders. Q J Med 1997;90:545–550.
42. Mallet L, Kuyumjian J. Indomethacin-induced behavioral changes in an el- 61. McDowell BJ, Engberg S, Sereika S et al. Effectiveness of behavioral therapy derly patient with dementia. Ann Pharmacother 1998;32:201–203.
to treat incontinence in homebound older adults. J Am Geriatr Soc 1999;47: 43. Johnson AG. NSAIDs and blood pressure. Clinical importance for older pa- tients. Drugs Aging 1998;12:17–27.
62. Appell RA. Clinical efficacy and safety of tolterodine in the treatment of 44. Johnson AG, Seidemann P, Day RO. NSAID-related adverse drug interac- overactive bladder: A pooled analysis. Urology 1997;50:90–96.
tions with clinical relevance: An update. Int J Clin Pharmacol Ther 1994;32: 63. Zorzitto ML, Holliday PJ, Jewett MA. Oxybutynin chloride for geriatric uri- nary dysfunction: A double-blind, placebo-controlled study. Age Ageing 45. Hoppmann RA, Peden JG, Ober SK. Central nervous system side effects of nonsteroidal anti-inflammatory drugs. Aseptic meningitis, psychosis, and 64. Burgio KL, Locher JL, Goode PS et al. Behavioral vs drug treatment for urge cognitive dysfunction. Arch Intern Med 1991;151:1309–1313.
urinary incontinence in older women: A randomized controlled trial. JAMA 46. Brooks PM, Kean WF, Kassam Y et al. Problems of antiarthritic therapy in the elderly. J Am Geriatr Soc 1984;32:229–234.
65. Pritchard JF, Bryson JC, Kernodle AE et al. Age and gender effects on on- 47. Ray WA. Psychotropic drugs and injuries among the elderly: A review. J Clin dansetron pharmacokinetics: Evaluation of healthy aged volunteers. Clin 48. Darcourt G, Pringuey D, Salliere D et al. The safety and tolerability of zolpi- 66. Isah AO, Rawlins MD, Bateman DN. The pharmacokinetics and effects of dem–an update. J Psychopharmacol 1999;13:81–93.
prochlorperazine in elderly female volunteers. Age Ageing 1992;21:27–31.
49. Dashevsky BA, Kramer M. Behavioral treatment of chronic insomnia in psy- 67. Hulme A, MacLennan WJ, Ritchie RT et al. Baclofen in the elderly stroke pa- chiatrically ill patients. J Clin Psychiatry 1998;59:693–699.
tient: Its side-effects and pharmacokinetics. Eur J Clin Pharmacol 1985;29: 50. Freeman C. Drug treatment of insomnia in the elderly. Conn Med 1992;56: 68. Urba SG. Nonpharmacologic pain management in terminal care. Clin Geriatr 51. Dommisse CS, DeVane CL. Buspirone: A new type of anxiolytic. Drug Intell 69. Fishback DB. An approach to the treatment of hypertension in the aged. An- 52. Hart RP, Colenda CC, Hamer RM. Effects of buspirone and alprazolam on the cognitive performance of normal elderly subjects. Am J Psychiatry 1991; 70. Ganzini L, Walsh JR, Millar SB. Drug-induced depression in the aged. What can be done? Drugs Aging 1993;3:147–158.
53. Smith M, Buckwalter KC. Medication management, antidepressant drugs, 71. Blom MW, Sommers DK. Placebo substitution for methyldopa in geriatric and the elderly: An overview. J Psychosoc Nurs Ment Health Serv 1992;30: hypertensive patients. S Afr Med J 1993;83:335–336.
72. Thompson TL 2d, Filley CM, Mitchell WD et al. Lack of efficacy of hydergine in 54. Schweizer E, Rickels K, Hassman H et al. Buspirone and imipramine for the patients with Alzheimer’s disease. N Engl J Med 1991;324:197–198.
treatment of major depression in the elderly. J Clin Psychiatry 1998;59:175–183.
73. Rogers SL. Perspectives in the management of Alzheimer’s disease: Clinical 55. Roberto V, Vitaliano B, Donnatella P et al. Disopyramide pharmacokinetics profile of donepezil. Dement Geriatr Cogn Disord 1998;9:29–42.
in the elderly after single oral administration. Pharmacol Res Commun 1988; 74. Shintani EY, Uchida KM. Donepezil: An acetylcholinesterase inhibitor for Alzheimer’s disease. Am J Health Promot 1997;54:2805–2810.
56. Connolly SJ. Evidence-based analysis of amiodarone efficacy and safety. Cir- 75. Katz IR, Jeste DV, Mintzer JE et al. Comparison of risperidone and placebo for psychosis and behavioral disturbances associated with dementia: A ran- 57. Krentz AJ, Ferner RE, Bailey CJ. Comparative tolerability profiles of oral an- domized, double-blind trial. Risperidone Study Group. J Clin Psychiatry tidiabetic agents. Drug Safety 1994;11:223–241.
INAPPROPRIATE MEDICATIONS IN RESIDENTIAL CARE
aspirin allergy or those who have failed effects (dizziness, confusion), which may Efficacy is questionable; low-dose aspirin More anticholinergic activity than other Toxic metabolite may accumulate in older Mixed agonist/antagonist agents should be may lead to frontal headaches; may cause antihypertensive effect of beta-blockers; risk of psychosis in dementia patients.
Appendix 1. Potentially Inappropriately Prescribed Medications Studied and Possible Alternative Treatments
SLOANE ET AL.
delirium, sedation, postural hypotension Use of agents with long half-lives leads to Potent negative inotrope; may induce heart Long half-life drug, therefore more likely to Appendix 1 (Continued)
INAPPROPRIATE MEDICATIONS IN RESIDENTIAL CARE
potential for benefit; avoid long-term use; otential for toxic reactions greater than potential for benefit; highly anticholinergic; people; can aggravate peptic ulcers.
Potential for toxic reactions greater than Not effective for incontinence and detrusor May cause extrapyramidal adverse events; Therapeutically ineffective; extrapyramidal nonsteroidal antiinflammatory drug.
ϭ Appendix 1 (Continued)

Source: http://www.heleo.org/sitebuildercontent/sitebuilderfiles/inappropriatemedicationprescribinginresidentialcareassistedlivingfacilities.pdf