Abstract
What is already known about this subject
Medication-related problems are more common among the elderly and are associated with poor outcomes.
Identification of elderly patients at high risk of medication misadventure and timely interventions could avoid unnecessary hospitalizations.
What this study adds
Three-quarters of elderly patients living in sheltered housing complexes had at least one risk factor for medication-related problems.
Sheltered housing residents using five or more medications or using any medication with a narrow therapeutic index were more likely to have unplanned hospitalizations.
Use of medications with high potential for adverse drug reactions in the elderly or self-reported non-adherence to prescribed medications were not independent predictors of unplanned hospitalizations.
Aim
To identify risk factors for unplanned hospitalizations among residents of sheltered housing complexes (SHCs).
Methods
Medication-related risk factors for health outcomes among residents of SHCs in Aberdeen (n = 1137) were assessed using a postal questionnaire. Predictors of unplanned hospitalization/emergency department (ED) visit were identified using logistic regression.
Results
Of the 695 (61.1%) responses received, 645 were from residents (mean age 78.2 years) using prescribed medications. One or more risk factors for medication-related problems was seen in 467 (72.4%) respondents; 488 (75.7%) were using medications with high potential for adverse drug reactions (ADRs) in the elderly. Unplanned hospitalizations/ED visits (n = 230) were found to be associated with use of drugs of narrow therapeutic index [P < 0.001; odds ratio (OR) 2.98, 95% confidence interval (CI) 1.69, 5.28]; use of five or more different medications (P = 0.001; OR 2.10, 95% CI 1.34, 3.31); and greater disability (Townsend score) (P = 0.005; OR 1.06, 95% CI 1.02, 1.11).
Conclusion
Residents of SHCs using drugs of narrow therapeutic index, using five or more different medications, and with greater disability warrant periodic monitoring.
Keywords: adverse drug reactions, elderly, hospitalization, medication, risk, sheltered housing
Introduction
Medication-related problems are more common among the elderly, accounting for considerable morbidity, mortality and additional healthcare costs [1]. Lack of medication administration routine, therapeutic duplication, hoarding, confusion between generic and trade names, multiple prescribers, discontinued medication repeats retained and multiple storage locations have been reported among elderly patients living in their own homes [2].
Data suggest that more than half of medication-related hospital admissions are potentially avoidable if appropriate measures are taken [3]. Identification of elderly patients at high risk of health outcomes is therefore important, particularly in those with minimal social support [4]. Several screening tools and prognostic indicators have been used by health professionals to identify high-risk patients [5, 6]. Studies on medication-related risk factors among the elderly have focused on those living in their own homes and nursing homes [2, 7]; limited data are available on those receiving intermediate care, i.e. those living in sheltered housing complexes (SHCs). This study aimed to identify risk factors associated with unplanned hospitalizations among residents of SHCs.
Methods
A questionnaire specifically enquiring about use of medications with a high potential for adverse drug reactions (ADRs) in the elderly – antihypertensives; digoxin; diuretics; hypnotics; nonsteroidal anti-inflammatory drugs (NSAIDs); Parkinson's drugs; and warfarin [8]– validated scales for patient self-assessment of medication risk [6], adherence [9] and disability [10, 11], and general items on demographics, health and medication use was developed. For better patient understanding, layman terms were used for medication classes, where possible (e.g. blood pressure tablets for antihypertensives, water tablets for diuretics and sleeping tablets for hypnotics). The use of psychotropics was not explored due to difficulty in translating it into layman terms and concerns about respondent bias. The medication risk questionnaire (MRQ), originally developed for self-administration in elderly ambulatory patients to identify those at risk of a medication-related problem [6], has 10 items regarding medication use, each with a ‘yes’ or ‘no’ response option. The Morisky scale, a patient self-reported adherence tool, has four closed questions on medication use with binary response options (yes/no) [9]. It has demonstrated satisfactory sensitivity and specificity in the British population [12]. The disability scale was the modified version of the Townsend scale [10], which has been validated in elderly people in Britain [11]. The scale consists of nine items on which the subjects have to report their level of difficulty on a scale of 0–2 (0, yes, with no difficulty; 1, yes, with some difficulty; and 2, no, need help). The total score is the sum of the scores for the nine items, all of which are weighted equally.
The questionnaire was pilot tested in residents from one of the SHCs (n = 58) randomly chosen from the list of all SHCs (n = 27) in Aberdeen. Based on the responses (n = 35; 60.3%), minimal changes were made to the questionnaire. The final questionnaire was mailed in November 2005 to all residents (n = 1137) in 24 SHCs within Aberdeen (excluding the pilot group and two SHCs that were not interested due to recent participation in other surveys). Nonrespondents were sent one reminder at 2 weeks.
Responses were entered into an SPSS (version 13.0) database (SPSS Inc., Chicago, IL, USA). Internal consistency of the disability scale was assessed using Cronbach's α. The predictors of the outcome as a binary variable (yes/no) – any unplanned hospitalization/emergency department (ED) visit in the last 12 months – were identified using Student's t-test/χ2/Mann–Whitney U-test. Significant (P< 0.05) variables were further analysed in a logistic regression model (forward selection method and verified using backward selection method) to identify the independent predictors of the outcome. Results are reported as odds ratio (OR) with 95% confidence intervals (95% CI). Correlations between number of hospital admissions/ED visits, self-reported general health and disability score were assessed using Spearman's ρ.
This study was approved by the Ethical Review Panel of the School of Pharmacy at The Robert Gordon University. Grampian Research Ethics Committee advised that this study did not require formal review by a National Health Service ethics committee.
Results
Of the 1137 questionnaires sent out, 695 (61.1%) responses were received. Responses from 50 residents not using any prescribed medications were excluded from analysis. Those using prescribed medications (n = 645; 92.8%) were aged 78.2 ± 7.8 (mean ± SD) years. Their characteristics are summarized in Table 1. Only 161 (23.2%) reported that their community pharmacist talked to them about their medications and 111 (16.0%) got help from someone to use their medications correctly. The use of reminders to help medication use was reported by 184 (28.5%) respondents.
Table 1.
Description of the respondents using prescribed medications (n = 645*)
| Characteristic | N(%) |
|---|---|
| Sex | |
| Male | 214 (33.2) |
| Female | 430 (66.7) |
| Number of times seen by a doctor/nurse in the last 12 months | |
| 0 | 18 (2.8) |
| 1–3 | 160 (24.8) |
| 4–6 | 187 (29.0) |
| 7–9 | 66 (10.2) |
| 10–12 | 78 (12.1) |
| >12 | 60 (9.3) |
| Number of unplanned hospital admissions/emergency department visits in the last 12months | |
| 0 | 380 (58.9) |
| 1 | 136 (21.1) |
| 2 | 43 (6.7) |
| 3 | 25 (3.9) |
| 4 | 12 (1.9) |
| ≥5 | 14 (2.3) |
| Self-reported general health on a scale of 1 (as bad as it can be) to 5 (as good as it can be) | |
| 1 | 29 (4.5) |
| 2 | 79 (12.2) |
| 3 | 207 (32.1) |
| 4 | 136 (21.1) |
| 5 | 171 (26.5) |
| Disability based on Townsend score | |
| 0 (no disability) | 39 (6.0) |
| 1–2 (slight disability) | 42 (6.5) |
| 3–6 (some disability) | 117 (18.1) |
| 7–10 (appreciable disability) | 128 (19.8) |
| 11–14 (severe disability) | 127 (19.7) |
| 15–18 (very severe disability) | 92 (14.3) |
| Help in completing the questionnaire | |
| Yes | 162 (25.1) |
| No | 455 (70.5) |
Responses do not add up to 645 due to missing responses.
According to the MRQ, 467 (72.4%) respondents had at least one risk factor for medication-related problems; the median number of risk factors in the cohort was three (range 0–9). Usage of at least one medication with a high potential for ADRs in the elderly was reported by 488 (75.7%); the utilization of individual medications/medication classes were: antihypertensives (317; 49.1%); digoxin (41; 6.4%); diuretics (253; 39.2%); hypnotics (89; 13.8%); NSAIDs (153; 23.7%); Parkinson's drugs (12; 1.9%); and warfarin (46; 7.1%). The internal consistency of the disability scale was high (α = 0.91). During the 12 months prior to the study, 230 residents self-reported one or more unplanned hospitalizations/ED visits. The number of unplanned hospitalizations/ED visits had moderate, but significant correlations with self-reported general health (ρ = −0.19; P < 0.001) and disability score (ρ = 0.25; P < 0.001).
In univariate analysis, several items regarding medication use including those in the MRQ (Table 2) and the following factors were associated with unplanned hospitalizations/ED visits: poor self-reported general health (P< 0.001); number of medications with risk of ADRs in the elderly (P = 0.008); and greater disability score (P< 0.001). In logistic regression (n = 428), use of drugs of narrow therapeutic index (P< 0.001; OR 2.98, 95% CI 1.69, 5.28); use of five or more different medications (P = 0.001; OR 2.10, 95% CI 1.34, 3.31); and greater self-reported disability (Townsend score) (P = 0.005; OR 1.06, 95% CI 1.02, 1.11) were the only independent predictors of one or more unplanned hospitalizations/ED visits (n = 162).
Table 2.
Univariate analysis of responses to medication risk questionnaire and other items (n = 645†)
| Yes | No | |||
|---|---|---|---|---|
| Item | HospitalizedN (%) | Not hospitalizedN (%) | HospitalizedN (%) | Not hospitalizedN (%) |
| Do you currently take five or more different medicines?‡ | 156 (24.2)*** | 168 (26.0) | 63 (9.8) | 199 (30.9) |
| Do you currently take 12 or more tablets or capsules per day?‡ | 62 (9.6)*** | 56 (8.7) | 149 (23.1) | 300 (46.5) |
| Are you taking medicines for three or more medical problems?‡ | 163 (25.3)*** | 208 (32.2) | 57 (8.8) | 155 (24.0) |
| Does more than one doctor prescribe medicines for you on a regular basis?‡ | 103 (16.0)* | 134 (20.8) | 119 (18.4) | 235 (36.4) |
| Have your medicines or the instructions on how to take them been changed four times or more in the last 12 months?‡ | 39 (6.0)*** | 27 (4.2) | 179 (27.8) | 341 (52.9) |
| Is it difficult for you to take your medicines as prescribed?‡ | 15 (2.3) | 14 (2.2) | 216 (33.5) | 367 (56.9) |
| Do you take any of the following medicines: carbamazepine, lithium, phenytoin, warfarin, digoxin, phenobarbital, procainamide, theophylline?‡ | 58 (9.0)*** | 37 (5.7) | 170 (26.4) | 346 (53.6) |
| Do you collect your own medicines from the pharmacy?‡ | 113 (17.5) | 248 (38.4) | 115 (17.8)*** | 135 (20.9) |
| Are your prescriptions always dispensed at the same pharmacy?‡ | 192 (29.8) | 319 (49.5) | 38 (5.9) | 66 (10.2) |
| Do you know why you are taking all of your medicines?‡ | 213 (33.0) | 360 (55.8) | 15 (2.3) | 21 (3.3) |
| Do you take any of the following medicines: antihypertensives, digoxin, diuretics, hypnotics, nonsteroidal anti-inflammatory drugs, Parkinson's drugs, warfarin? | 180 (27.9) | 278 (43.1) | 48 (7.4) | 104 (16.1) |
| Does your pharmacist ever speak to you about your medicines? | 60 (9.3) | 91 (14.1) | 166 (25.7) | 288 (44.7) |
| Do you ever forget to take your medicine?§ | 38 (5.9) | 190 (29.5) | 57 (8.8) | 320 (49.6) |
| Are you always careful about taking your medicine?§ | 15 (2.3) | 213 (33.0) | 18 (2.8) | 358 (55.5) |
| When you feel better, do you sometimes stop taking your medicine?§ | 17 (2.6) | 210 (32.6) | 23 (3.6) | 352 (54.6) |
| Sometimes, if you feel worse when you take your medicine, do you stop taking it?§ | 29 (4.5) | 196 (30.4) | 47 (7.3) | 328 (50.9) |
| Do you use anything to help you remember to use or take your medicines? | 85 (13.2)*** | 93 (14.4) | 142 (22.0) | 287 (44.5) |
| Does any one help you to use or take your medicines correctly? | 51 (7.9)* | 56 (8.7) | 180 (27.9) | 322 (49.9) |
| Do you think you need any more help to use your medicines? | 5 (0.8) | 3 (0.5) | 222 (34.4) | 376 (58.3) |
Responses do not add up to 645 due to missing responses.
Items from the medication risk questionnaire.
Items from the Morisky scale.
P < 0.05
P < 0.001.
Discussion
Three-quarters of respondents in the study were found to be using medications with a high risk of ADRs in the elderly and had at least one risk factor for medication-related problems. Use of medications of narrow therapeutic index, use of five or more different medications and greater self-reported disability were found to be associated with health outcomes. Presence of drugs that need therapeutic monitoring was found to be significantly associated with an increased risk of adverse events in both children and adults who visited ambulatory clinics [5]. The concurrent use of five or more medications (major polypharmacy) was known to be a risk factor for poor clinical outcomes as early as the 1970s [13]. The association between disability and hospitalization observed in our study is not surprising considering the number of participants reporting their disability as either appreciable or worse. Self-assessment of medication risk is known to be effective in identifying medication-related risk factors among ambulatory patients attending primary care clinics in the USA [14]. We have tested the suitability of a self-assessment postal questionnaire (MRQ) for identifying medication-related risk factors in older patients living alone. Our findings would assist in identifying sheltered housing residents who are at risk of hospitalization and are in greatest need of interventions. A multidisciplinary service model, involving general practitioners and pharmacists, delivering medication review to Australian patients in the community at risk of medication misadventure has been shown to have positive trends in clinical outcomes and costs [15]. The role of such interventions in health outcomes of patients living in SHCs, especially those at greater risk, needs to be further studied.
Evidence suggests that one in five elderly patients is at risk of receiving at least one inappropriate drug prescription [16, 17]. The majority of the respondents in our study were using medications with a high potential for ADRs in the elderly and the number of medications with a risk of ADRs in the elderly was a significant predictor of hospitalization in univariate analysis; however, this was not an independent predictor of the outcome. The overlap between some of the medications listed as having risk of ADRs and medications of narrow therapeutic index might explain this lack of association. Patients using medications with a risk of ADRs might have been using five or more medications and/or might have self-reported greater disability. Similar to the findings of a recent Canadian study [18], one-third of sheltered housing residents self-reported medication non-adherence, but it was not associated with hospitalization. In a study of community-dwelling elderly at high risk of adverse drug events, failure to prescribe indicated medications, monitor medication use appropriately, document necessary information, educate patients and maintain continuity were more common prescribing problems than use of inappropriate drugs [19]. These findings suggest that efforts to improve patient health outcomes should focus not only on exclusion of inappropriate medications but also on ensuring prescription of the right medications and their monitoring, data recording and patient education.
Our study has a few limitations. We have responses from only two-thirds of the SHC residents. Although a quarter of the respondents reported receiving help in completing the questionnaire, we cannot eliminate the possibility of nonresponse due to disability or hospitalization at the time of the study. However, we included all the SHCs in Aberdeen and the characteristics of our respondents are similar to those who participated in a study carried out in SHCs in England [20]. The majority of our respondents were using prescribed medications and it is possible that the nonrespondents were either uninterested in the study or were not using prescription medications. We relied on self-reported unplanned hospitalization/ED visit as the marker for health outcomes. The cross-sectional study design does not allow us to draw conclusions regarding the cause–effect relationships between the outcome and risk factors. However, unscheduled admissions are known to be important risk factors for future health outcomes [21, 22]. Correlations between the number of hospitalizations, general health and disability were in the predicted direction and confirm the validity of our data. Since the study did not collect any personal details that could reveal the respondents' identity, we could not follow up the patients to evaluate their health outcomes more objectively. In an attempt to improve accuracy in self-report, we sought hospitalization information in the previous 12 months only. Moreover, hospitalization was treated as a binary variable rather than a continuous variable in data analysis. Psychotropics also have high a risk of ADRs in the elderly [8]; however, their use was not explored due to the sensitivity of the topic and concerns about patient nonresponse. The validity and reliability of items in the MRQ have not been established in British patients. It is possible that some of the respondents using dosage forms other than tablets and capsules might have misunderstood the term ‘medicines’ in the MRQ items and responded incorrectly.
This study has shown that sheltered housing residents using medications of narrow therapeutic index, using five or more medications and with greater disability, were at greater risk of hospital admissions and should be periodically monitored. Interventions for improving health outcomes of residents needs to be studied, targeting those at greater risk.
Acknowledgments
Competing interests: None declared.
This study was funded by the School of Pharmacy, The Robert Gordon University, Aberdeen, Scotland. The authors acknowledge the input of E. Fallon, L. McGuire and V. Thomson in data collection. Abstracts based on this work have been accepted for presentation at the 2006 British Pharmaceutical Conference in Manchester.
References
- 1.Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, Farrar K, Park BK, Breckenridge AM. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ. 2004;329:15–9. doi: 10.1136/bmj.329.7456.15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sorensen L, Stokes JA, Purdie DM, Woodward M, Roberts MS. Medication management at home: medication-related risk factors associated with poor health outcomes. Age Ageing. 2005;34:626–32. doi: 10.1093/ageing/afi202. [DOI] [PubMed] [Google Scholar]
- 3.Hallas J, Harvald B, Gram LF, Grodum E, Brosen K, Haghfelt T, Damsbo N. Drug related hospital admissions: the role of definitions and intensity of data collection, and the possibility of prevention. J Intern Med. 1990;228:83–90. doi: 10.1111/j.1365-2796.1990.tb00199.x. [DOI] [PubMed] [Google Scholar]
- 4.Boult C, Pualwan TF, Fox PD, Pacala JT. Identification and assessment of high-risk seniors. HMO Workgroup on Care Management. Am J Manag Care. 1998;4:1137–46. [PubMed] [Google Scholar]
- 5.Koecheler JA, Abramowitz PW, Swim SE, Daniels CE. Indicators for the selection of ambulatory patients who warrant pharmacist monitoring. Am J Hosp Pharm. 1989;46:729–32. [PubMed] [Google Scholar]
- 6.Barenholtz Levy H. Self-administered medication-risk questionnaire in an elderly population. Ann Pharmacother. 2003;37:982–7. doi: 10.1345/aph.1C305. [DOI] [PubMed] [Google Scholar]
- 7.Lapane KL, Hughes CM. Identifying nursing home residents at high risk for preventable adverse drug events: modifying a tool for use in the Fleetwood Phase III Study. Consult Pharm. 2004;19:533–7. doi: 10.4140/tcp.n.2004.533. [DOI] [PubMed] [Google Scholar]
- 8.Mehta DK, editor. British National Formulary. London: Pharmaceutical Press; 2005. Prescribing for the elderly; pp. 18–9. [Google Scholar]
- 9.Morisky DE, Green LW, Levine DM. Concurrent and predictive validity of a self-reported measure of medication adherence. Med Care. 1986;24:67–74. doi: 10.1097/00005650-198601000-00007. [DOI] [PubMed] [Google Scholar]
- 10.Townsend P. Poverty in the United Kingdom. Pelican: Harmondsworth; 1979. [Google Scholar]
- 11.McGee MA, Johnson AL, Kay DW. The description of activities of daily living in five centres in England and Wales. Medical Research Council Cognitive Function and Ageing Study. Age Ageing. 1998;27:605–13. doi: 10.1093/ageing/27.5.605. [DOI] [PubMed] [Google Scholar]
- 12.George CF, Peveler RC, Heiliger S, Thompson C. Compliance with tricyclic antidepressants: the value of four different methods of assessment. Br J Clin Pharmacol. 2000;50:166–71. doi: 10.1046/j.1365-2125.2000.00244.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stewart RB, Cluff LE. A review of medication errors and compliance in ambulant patients. Clin Pharmacol Ther. 1972;13:463–8. doi: 10.1002/cpt1972134463. [DOI] [PubMed] [Google Scholar]
- 14.Langford BJ, Jorgenson D, Kwan D, Papoushek C. Implementation of a self-administered questionnaire to identify patients at risk for medication-related problems in a family health center. Pharmacotherapy. 2006;26:260–8. doi: 10.1592/phco.26.2.260. [DOI] [PubMed] [Google Scholar]
- 15.Sorensen L, Stokes JA, Purdie DM, Woodward M, Elliott R, Roberts MS. Medication reviews in the community: results of a randomized, controlled effectiveness trial. Br J Clin Pharmacol. 2004;58:648–64. doi: 10.1111/j.1365-2125.2004.02220.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.van der Hooft CS, Jong GW, Dieleman JP, Verhamme KM, van der Cammen TJ, Stricker BH, Sturkenboom MC. Inappropriate drug prescribing in older adults: the updated 2002 Beers criteria – a population-based cohort study. Br J Clin Pharmacol. 2005;60:137–44. doi: 10.1111/j.1365-2125.2005.02391.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhan C, Sangl J, Bierman AS, Miller MR, Friedman B, Wickizer SW, Meyer GS. Potentially inappropriate medication use in the community-dwelling elderly: findings from the 1996 Medical Expenditure Panel Survey. JAMA. 2001;286:2823–9. doi: 10.1001/jama.286.22.2823. [DOI] [PubMed] [Google Scholar]
- 18.Vik SA, Hogan DB, Patten SB, Johnson JA, Romonko-Slack L, Maxwell CJ. Medication nonadherence and subsequent risk of hospitalisation and mortality among older adults. Drugs Aging. 2006;23:345–56. doi: 10.2165/00002512-200623040-00007. [DOI] [PubMed] [Google Scholar]
- 19.Higashi T, Shekelle PG, Solomon DH, Knight EL, Roth C, Chang JT, Kamberg CJ, MacLean CH, Young RT, Adams J, Reuben DB, Avorn J, Wenger NS. The quality of pharmacologic care for vulnerable older patients. Ann Intern Med. 2004;140:714–20. doi: 10.7326/0003-4819-140-9-200405040-00011. [DOI] [PubMed] [Google Scholar]
- 20.Ryan-Woolley BM, Rees JA. Initializing concordance in frail elderly patients via a medicines organizer. Ann Pharmacother. 2005;39:834–9. doi: 10.1345/aph.1E148. [DOI] [PubMed] [Google Scholar]
- 21.Pacala JT, Boult C, Boult L. Predictive validity of a questionnaire that identifies older persons at risk for hospital admission. J Am Geriatr Soc. 1995;43:374–7. doi: 10.1111/j.1532-5415.1995.tb05810.x. [DOI] [PubMed] [Google Scholar]
- 22.Brody KK, Johnson RE, Ried LD, Carder PC, Perrin N. A comparison of two methods for identifying frail Medicare-aged persons. J Am Geriatr Soc. 2002;50:562–9. doi: 10.1046/j.1532-5415.2002.50127.x. [DOI] [PubMed] [Google Scholar]