Family Practice Advance Access published online on December 11, 2006
Family Practice, doi:10.1093/fampra/cml067
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Prevalence and predictors of polypharmacy among older primary care patients in Germany
Department of General Practice, Hannover Medical School, D-30623 Hannover, Germany. Correspondence to U Junius-Walker; Email: junius-walker.ulrike{at}mh-hannover.de
Received 2 June 2006; Revised 7 November 2006; Accepted 10 November 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Junius-Walker U, Theile G and Hummers-Pradier E. Prevalence and predictors of polypharmacy among older primary care patients in Germany.
Background. Older people consume an increasing amount of medication. Polypharmacy is associated with an elevated risk of adverse health outcomes resulting in hospitalizations and sometimes death.
Objectives. To describe the prevalence of prescribed and over-the-counter (OTC) medications among older general practice patients living in the community. To determine predictors of polypharmacy (five or more prescribed drugs) from a variety of patient- and doctor-related factors.
Methods. Sixty-seven randomly selected practices in two areas of Germany and 466 of their older patients (70+ years) were recruited for a geriatric assessment study. A cross-sectional analysis of health problems, GPs' awareness and their interventions was conducted. In this post hoc analysis, we assessed the medication use as reported by older patients and compared it with doctors' perceived medication regimens for their respective patients. The detailed assessment of patients' health and well-being enabled us to explore a variety of predictors of polypharmacy using logistic regression analysis with forward selection.
Results. Study participants consumed an average of 3.7 prescribed medicines and an additional 1.4 OTC drugs. In all, 26.7% of patients used five and more chronically prescribed drugs. A set of five determinants predicted polypharmacy best: breathlessness, hypertension, dependency on instrumental activities of daily living, low subjective health and medication disagreement between doctors and patients.
Conclusion. This older general practice population in Germany is among the top pharmaceutical user group of European study samples. Apart from disease-specific determinants, GPs should be aware that low subjective health and medication disagreement are independent predictors of polypharmacy.
Keywords. Ageing, preventive medicine, prescribing, patient safety, family medicine.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Escalating pharmaceutical costs, new budgetary demands and a growing awareness of health risks for patients with polypharmacy exert pressure on GPs to reduce medication. This necessitates a good understanding of how multiple drug use comes about. There is to date no common definition of polypharmacy available. It is determined either as the simultaneous use of a certain number of medications (two to six and more)1–3 or as the unnecessary overuse of drugs.4 It can refer to perceptions of prescribers or consumers and may or may not include over-the-counter (OTC) remedies. If defined as use of five or more drugs, between 4%1 and 34%5 of people aged 65 years and above are affected by polypharmacy. A number of studies investigated determinants of prescribed polypharmacy and reported relevant socio-demographic factors (age, gender, education, employment and socio-economic status),6,7 influence of disease (multimorbidity, multiple complaints, well-being and chronic illness)8,9 and health system factors (prescriber related, perceived patient pressure and free access to medications).10–13 These studies employed either limited numbers of health determinants or looked at overall health as an abstract concept when predicting polypharmacy.
Our analysis is based on a comprehensive health check for older people. Hence, we aimed to look at a large variety of patient- and doctor-related factors that predict the simultaneous and regular use of five and more chronically prescribed medications. We also wanted to determine the amount of self-reported medications taken by older primary care patients in Germany as there are very little data available on this issue.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
This post hoc analysis is derived from a multicentre cross-sectional study on health complaints of older primary care patients and their GPs' awareness of them. Part of the study was to assess the actual medication use of older patients in comparison to the medication regimen as perceived by their GPs.
All GPs and primary care internists in the area of Leipzig and Hannover were included in a computer-generated randomization process. Practices, which were willing to take part, then provided study participants in a systematic but non-random fashion. This meant that the first two patients of at least 70 years entering the practice 1 hour after opening in the morning and afternoon were chosen every day in the course of 1 week. Patients who required regular home visits or lived in institutions were excluded.
All participants received a precursor version of the STEP-Assessment14 that had been tested in a pilot feasibility study. The patients were given a set of partly validated questionnaires to rate the presence or absence of 31 somatic diseases, functional limitations and psychological and social problems. The instruments and their sources have been published elsewhere.15,16 Additionally, patients were asked to number the different chronically prescribed drugs and OTC remedies in daily use. Likewise, the doctors were independently invited to specify their patients' health problems and the quantity of prescribed medications and OTC drugs. As part of the assessment, the doctors also performed simple standard examinations which included repeated blood pressure measurement, pulse examination, dementia test, get-up-and-go test as well as a blood sample for thyroid stimulating hormone, cholesterol and glucose.
All 31 health observations, the two medication variables (number of prescribed and OTC medications) and socio-demographic data (gender, age and formal education) were used in the analysis as well as a medication disagreement item. This parameter was defined as the percentage of patients for whom doctors either over- or underestimated the prescribed medications in reference to the patients' statements. For baseline characteristics, we classified health variables into four groups that constitute the sum of 17 somatic, 3 functional, 4 psychological and 6 social variables. Univariate comparisons for all health variables were performed to describe the polypharmacy group (defined in this study as five and more prescribed medications) as opposed to the reference group. Probability statements were given using two-tailed t-tests and chi-square tests. Logistic regression with forward selection was chosen to explore predictors of prescribed multiple medication use with polypharmacy as the binary outcome and significant univariate coefficients as predictors. The model was internally validated using the Hosmer–Lemeshow test.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Description of the sample
In all, 67 primary care practices agreed to participate. They represented 30% and 65% of the invited practices in the area of Hannover and Leipzig, respectively. In all, 713 patients were systematically recruited in the practices and 466 (65.4%) participated in the survey. Table 1 presents an overview of health and socio-demographic characteristics by gender. Twice as many women as men took part in this study. Significant gender differences existed for formal education and social health problems.
|
Medication status
Altogether, 419 observations were available for the prescribed and 221 for the OTC medication status. Figure 1 provides a flow chart of available data for each analysis. A total of 418 (99.8%) participants reported to take at least one prescribed drug on a regular basis. In all, 112 patients were affected by polypharmacy, which constitutes approximately a quarter of the study population. A total of 184 (83.3%) consumed at least one non-prescribed remedy daily. For the combined prescribed and OTC medications (209 observations), the proportion of patients with at least five drugs was 53.7% (Table 1). The association between the number of prescribed medication and age was fairly weak and non-significant (r = 0.36). Likewise, there was only a small but significant correlation between prescribed medication and the number of health problems (r = 0.32, P < 0.01).
|
Agreement between doctors and patients
There were 344 observations available for cross-tabulation to analyse the patients' and doctors' agreement on the regular intake of prescribed drugs. Overall, doctors and patients agreed on the same number of prescribed drugs in 56.4%. The agreement was greater in the subgroup of patients with zero to four drugs (60.9%). In this low-user group (N = 256), 22.3% took more medications than their doctor expected and 16.8% took less. In the polypharmacy group (N = 88), the agreement of prescribed medication use was significantly less, only 43.2%. For 54.5% of the cases, doctors underrated the number of consumed medications and they overestimated their use in 2.3%.
Health characteristics of the low-user versus high-user medication groups
We compared the prevalence of health problems for patients on five or more prescribed drugs (N = 112) with the low-user group (N = 307). For somatic problems, quite a number of conditions were significantly more common in the polypharmacy group (Table 2). These were syncope, dizziness, shortness of breath, pain, difficulty in chewing, chronic bronchitis, hypertension and arrhythmia. In the functional health area, a problem in either independent or basic activities of daily living was significantly associated with multiple drug use. Only depression as one of four psychological problems was significantly more prevalent in the high-user group. None of the social problems demonstrated a significant relation to polypharmacy.
|
Low subjective health was defined as having rated the overall health as less good or bad as opposed to good and excellent. In all, 15.6% of the participants who considered themselves to be in good health consumed five and more drugs as opposed to 38.9% with a low rating. Subjective health was moderately correlated with depression (r = 0.52, P < 0.01) and the number of health problems (r = 0.55, P < 0.01).
Predictors of polypharmacy
Logistic regression analysis with forward selection was applied to explore into risk factors of polypharmacy. Health problems that had shown a significant difference between high- and low-user groups entered the model as well as self-perceived health and medication disagreement between doctors and patients. Age and gender were added to adjust the model. In all, 325 patients, who presented a complete data set of self-rated problems as well as self- and doctor-rated medication use, were included. Table 3 presents the model derived from logistic regression with stepwise forward selection. Five variables adjusted by age and gender predict the use of polypharmacy best. The odds of being breathless, being hypertensive, being dependent on instrumental activities of daily living, having a low perceived health or having a medication disagreement were between two and three times higher in the polypharmacy group. Syncope, dizziness, pain, problems with chewing, chronic bronchitis, arrhythmia, dependency in basic activities of daily living and depression did not improve the model and were excluded in the process of forward selection. The Hosmer–Lemeshow test demonstrated that the model provides a good fit (P = 0.93).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Our study population of independent older general practice patients consumed a mean of 3.7 prescribed medicines and an additional 1.4 OTC drugs. In all, 26.7% were on prescribed polypharmacy. This increased to 53.6% when OTC drugs were considered as well. Surprisingly, just one patient out of our sample of 466 did not take any medicines at all. This puts our German practice sample into the top bracket of pharmaceutical consumers in comparison with results of other pharmaco-epidemiological studies in Europe. An Italian practice survey reported that 95% of women and 91% of men aged 75 years and more were on regular medication.17 Population-based surveys record user rates from 65% to 96% among older people.2,8,5 This substantial variation can also be observed in the average number of prescribed pharmaceuticals within European study samples ranging from 2.718 to 4.25 in older community-dwelling people. However, comparability is limited due to differing inclusion criteria between studies, such as age group and independent living. Medication data have been collected in a variety of ways (computer-based practice data, interview, presentation of drugs in surgery and inspection of drugs at home), and they are presented with diverse or in some cases indistinct definitions and specifications on medication use. Our study population was also somewhat selected, as only patients who visited the practice were invited to participate and patients requiring regular home visits or living in nursing homes were excluded.
The disagreement between drugs simultaneously taken by patients and described by their doctors was lower (45%) than in other studies (53%–58%).19,20 However, in our study we based the definition of agreement on the quantity of drug use alone and not on the type of medication. Doctors over- and underrated medication use of the low-user group to an equal extent. Their estimation became unbalanced for the high users. Over half of these patients were assumed to take fewer drugs.
This was confirmed in our prediction model for the use of five and more prescribed medications. Besides medical problems, most notably hypertension and breathlessness, the medication disagreement was an equally powerful and innovative predictor of polypharmacy. One reason for the disagreement is likely to be ‘different prescribers’, a risk factor previously identified for polypharmacy.11 The German health system with its community specialists, who work relatively independent without referrals from GPs, promotes professional ignorance about patients' medication status. Another cause may be an insufficient patient–doctor communication on medications. This may be determined by consultation lengths, which are among the shortest of European countries.21 A somewhat neglected determinant is the generally low subjective health rating among multiple drug users. Subjective health and morbidity were only moderately correlated and both had an independent influence on the quantity of prescribing. Whether patients with low perceived health exert the well-known perceived pressure on doctors to prescribe12 remains unclear. Several other known determinants could not be verified in our sample, such as being female and advancing age within the old age group.18 Secondary prescription data collected by a German health insurance company support our latter finding that the number of prescribed medications decreases in the highest age group.22
It is well known that multiple medication use entails health risks. According to several European primary care studies, inappropriate drug prescribing occurs in approximately 20% of older general practice patients.23 Older people exposed to potential drug interactions rise from 25% in the age group of between 60 and 80 years to 36% in the 80+ age category.24 The average prevalence of drug-related hospitalizations lies within 2.5%–6% among all hospitalized patients with increasing age and polypharmacy as significant predictors for admission, length of stay, re-admission and indeed death.25 Therefore, we can reasonably assume that a continuous medication assessment by primary care prescribers will reduce adverse outcomes. However, a recent meta-analysis on this subject showed no benefit when primary health care professionals conducted a medication review. Pharmacist-led interventions fared better, but there was a significant heterogeneity between studies.26
Our findings have to be interpreted with caution in view of it being a post hoc analysis. It was not the primary intention to assess the medication status of older primary care patients in detail but to merely identify patients with polypharmacy. The overall study aim was rather to describe self-reported health and the effects of geriatric assessment in German general practice. For reasons of practicality, we therefore only assessed the number of simultaneously and chronically used medications and did not attempt to record type and dosage. However, the similarity of pharmaco-epidemiological findings previously reported in Germany is reassuring.27 The intention to analyse OTC data had to be discarded because of high missing values for these items in the patient—as well as the doctor—questionnaires. Generally, there is a lack of understanding on self-administered remedies for different reasons such as complexity and the grey area of the definition of OTC drugs.28 Consequently, OTC drugs are hardly ever taken into account when drug-related problems or drug interventions are assessed.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Our study indicates that many older people in German general practice are high consumers of prescribed and OTC drugs. GPs underrate the number of prescribed drugs especially for multiple medication users. When issuing prescriptions, doctors should consider the possibility of polypharmacy and its predictors. In addition to disease, specific predictor knowledge of non-specific disease determinants such as poor subjective health and medication disagreement may facilitate good prescribing. Solutions to improve the quality of prescribing are not established in Germany yet, such as regular medication assessments, an electronic medication card or computer-aided prescribing. Future research could focus on medication assessment methodology as well as targeting high-risk groups for adverse drug effects for intervention.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
Funding: It was funded by the Federal Government of Education and Research.
Conflicts of interest: None.
|
Acknowledgments |
|---|
The study has been approved by the ethics committee of Hannover Medical School.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionDeclarationReferences |
|---|
1 Veehof L, Stewart R, Haaijer-Ruskamp F, Meyboom-de-Jong B. (2000) The development of polypharmacy. A longitudinal study. Fam Pract 17:261–267.
2 Kennerfalk A, Ruigomex A, Wallander M, Wilhelmsen L, Johannson S. (2002) Geriatric drug therapy and health care utilization in the United Kingdom. Ann Pharmacother 36:797–803.[Abstract]
3 Fialova D, Topinkova E, Gambassi G, et al. (2005) Potentially inappropriate medication use among elderly home care patients in Europe. J Am Med Assoc 293:1348–1358.
4 Avorn J. (2004) Polypharmacy. A new paradigm for quality drug therapy in the elderly? Arch Intern Med 164:1957–1958.
5 Barat I, Andreasen F, Damsgaard E. (2000) The consumption of drugs by 75-year-old individuals living in their own homes. Eur J Clin Pharmacol 56:501–509.[CrossRef][Web of Science][Medline]
6 Odubanjo E, Bennett K, Feely J. (2004) Influence of socioeconomic status on the quality of prescribing in the elderly—a population based study. Br J Clin Pharmacol 58:496–502.[CrossRef][Web of Science][Medline]
7 Perry B and Turner L. (2001) A prediction model for polypharmacy: are older, educated women more susceptible to an adverse drug event? J Women Aging 13:39–51.[Web of Science][Medline]
8 Linjakumpu T, Hartikainen S, Klaukka T, Veijola J, Kivelä S, Isoaho R. (2002) Use of medications and polypharmacy are increasing among the elderly. J Clin Epidemiol 55:809–817.[CrossRef][Web of Science][Medline]
9 Al-Windi A. (2005) Determinants of medicine use in a Swedish primary health care practice population. Pharmacoepidemiol Drug Saf 14:47–51.[CrossRef][Web of Science][Medline]
10 Bjerrum L, Sogaard J, Hallas J, Kragstrup J. (1999) Polypharmacy in general practice: differences between practitioners. Br Med J 49:195–198.
11 Hessel A, Gunzelmann T, Geyer M, Brähler E. (2000) Utilization of medical services and medication intake of patients over 60 in Germany. Z Gerontol Geriatr 33:289–299.[CrossRef][Medline]
12 Little P, Dorward M, Warner G, Stephens K, Senior J, Moore M. (2004) Importance of patient pressure and perceived pressure and perceived medical need for investigations, referral, and prescribing in primary care: nested observational study. Br Med J 328:444–447.
13 Perkins A, Kroenke K, Unützer J, et al. (2004) Common co-morbidity scales were similar in their ability to predict health care costs and mortality. J Clin Epidemiol 57:1040–1048.[CrossRef][Web of Science][Medline]
14 Williams I, Fischer G, Junius U, et al. (2002) An Evidence Based Approach to Assessing Older People. Occasional Papers 82(Royal College of General Practitioners, London).
15 Junius U, Fischer G, Kemmnitz W. (1995) Neue hausärztliche Versorgungsformen für ältere Patienten—Teil I: vom Screening zur Intervention. Das Hannoversche ambulante geriatrische Screening. Geriat Forsch 5:71–78.
16 Junius U, Kania H, Fischer G. (1996) Ein Präventionsprogramm für Gesundheitsprobleme im Alter. Das ambulante geriatrische Screening für den Einsatz in der Hausarztpraxis. Fortschr Med 114:33–41.[Medline]
17 Nobili A, Tettamanti M, Frattura L, et al. (1997) Drug use by the elderly in Italy. Ann Pharmacother 31:416–422.[Abstract]
18 Stuck A, Gloor B, Pfluger D, Minder C, Beck J. (1995) [Gender differences in medication use by community-dwelling older persons: an epidemiologic study in Bern, Switzerland]. Z Gerontol Geriatr 28:394–400.[Web of Science][Medline]
19 Atkin P, Stringer R, Duffy J, et al. (1998) The influence of information provided by patients on the accuracy of medication records. Med J Aust 169:85–88.[Web of Science][Medline]
20 Barat I, Andreasen F, Damsgaard E. (2001) Drug therapy in the elderly: what doctors believe and patients actually do. Br J Clin Pharmacol 51:615–622.[CrossRef][Web of Science][Medline]
21 Deveugele M, Derese A, van den Brink-Muinen A, Bensing J, de Maeseneer J. (2002) Consultation length in general practice: cross sectional study in six European countries. Br Med J 325:472–478.
22 Kuhlmey A, Winter M, Maaz A, Hofmann W, Nordheim J, Borchert C. (2003) [High utilization of health care services by older adults]. Z Gerontol Geriatr 36:233–240.[Web of Science][Medline]
23 Hooft C, Jong G, Dieleman J, et al. (2005) Inappropriate drug prescribing in older adults: the updated 2002 Beers criteria—a population-based cohort study. Br J Clin Pharmacol 60:137–144.[CrossRef][Web of Science][Medline]
24 Bjerrum L, Andersen M, Petersen G, Kragstrup J. (2003) Exposure to potential drug interactions in primary health care. Scand J Prim Health Care 21:153–158.[CrossRef][Web of Science][Medline]
25 Campbell S, Seymour D, Primrose W. (2004) A systematic literature review of factors affecting outcome in older medical patients admitted to hospital. Age Ageing 33:110–115 ACMEPLUS project.
26 Royal S, Smeaton L, Avery A, Hurwitz B, Sheikh A. (2006) Interventions in primary care to reduce medication related adverse events and hospital admission: systematic review and meta-analysis. Qual Saf Health Care 15:23–31.
27 Anders J, V Renteln-Kruse W, Dapp U, Gillman G, Stuck A. (2005) [Drug use and self-perceived health as reported by community-dwelling elderly persons in Hamburg]. Z Gerontol Geriatr 38:173–181.[CrossRef][Web of Science][Medline]
28 Francis S, Barnett N, Denham M. (2005) Switching of prescription drugs to over-the-counter status: is it a good thing for the elderly? Drugs Aging 22:361–370.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||