Abstract
Background:
Polypharmacy, given its link with drug interactions, potentially inappropriate medications and medication non-adherence, may pose a significant health risk, especially among the elderly. A comprehensive understanding of the epidemiology of polypharmacy is essential for countries facing population ageing and growing chronic disease burden, like Singapore.
Objectives:
We assessed the prevalence and risk factors of polypharmacy (⩾5 prescription medications) among community-dwelling elderly in Singapore and established the association of polypharmacy with medication non-adherence.
Methods:
We used data from a national sample of 1499 community-dwelling elderly aged ⩾66 years. Using logistic regression, we assessed the association of socio-demographic, health and healthcare use variables with polypharmacy, and polypharmacy with medication non-adherence.
Results:
The weighted prevalence of polypharmacy was 14.5%. In multivariable analysis, elderly who were men (versus women), with ⩾2 (versus 0–1) chronic diseases, aged ⩾85 (versus 66–69) years, and of Malay and Indian (versus Chinese) ethnicity were significantly more likely to have polypharmacy; healthcare use variables were not associated. Polypharmacy was significantly correlated with medication non-adherence.
Conclusions:
The prevalence of polypharmacy among community-dwelling elderly in Singapore is lower than that reported in other countries; however, polypharmacy is associated with medication non-adherence. Elderly subgroups, defined by gender, health status, age and ethnicity, who are at a higher risk of polypharmacy will benefit from medication review and de-prescribing services.
Introduction
On average, more than 50% of the elderly in developed countries have multiple chronic conditions. 1 Given that multiple morbidities often require more prescription medications, polypharmacy is expected to be an increasingly pertinent issue in countries with ageing populations.2,3 Polypharmacy, most commonly defined as taking ⩾5 medications, 4 is considered suitable when it supports optimal patient outcomes; 5 however, it is also known to entail health risks.6,7 The elderly, among whom chronic diseases and medication use are more common than those younger, are at greater risk of polypharmacy and its possible negative implications. Polypharmacy affects several dimensions of medication safety. First, it is linked with a higher risk of adverse drug reactions, especially so among the elderly, due to physiological changes occurring with age that potentially alter the pharmacokinetics and pharmacodynamics of certain classes of medications. 8 Second, it is related to a higher likelihood of potentially inappropriate medications (PIMs). 9 Third, it has been reported to be associated with medication non-adherence in a disease-specific older population. 10 Additionally, higher medication intake can contribute to rising healthcare costs. 2 Therefore, it is necessary to gain a holistic understanding of the epidemiology of polypharmacy among the elderly, especially in countries experiencing population ageing and growing chronic disease burden.
Singapore, with a rapidly ageing population, 11 is a relevant setting to explore polypharmacy-related issues. In 2010, 64% of years of life lost were attributed to chronic diseases, compared with 60% in 1990. 12 In addition, between the years 2004 and 2010, the total disease burden rose by 10%, while cardiovascular diseases and cancers mainly accounted for disability-adjusted life years in 2010. 13 These demographic and epidemiological trends portend an increase in medication intake for chronic diseases, which in turn necessitates research on medication use patterns among the elderly in Singapore.
Despite its relevance as a public health and policy issue in Singapore, polypharmacy has not been adequately studied. Previous studies focused on the outcomes of polypharmacy, such as its relationship with PIMs 7 and with hospital admissions due to drug-related problems. 14 Less is known about the prevalence and risk factors of polypharmacy and the link of polypharmacy with medication non-adherence. Moreover, published research primarily focused on institutional settings, such as nursing homes 7 and hospitals, 14 while community studies are lacking. Equal emphasis should be placed on community studies as ageing countries move from hospital-centred medical care towards community-oriented care, which is driven by the predicted rise in healthcare demands 11 and a push towards enabling older people to ‘age-in-place’. 15
Thus, we assessed the prevalence and risk factors of polypharmacy among community-dwelling elderly Singaporeans and the association of polypharmacy with medication non-adherence. In doing so, we aim to fill existing research gaps and inform community-based initiatives and policies aimed at addressing polypharmacy in Singapore.
Methods
Data source and analytical sample
We used data from Wave 3 of the Panel on Health and Ageing of Singaporean Elderly (PHASE), a nationally representative longitudinal survey of community-dwelling elderly in Singapore, conducted in 2015. PHASE collected data on various aspects of ageing, covering social, mental and physical health of older Singaporeans. PHASE was approved by the institutional review board at the National University of Singapore and all participants provided written informed consent. All interviews were conducted in the interviewee’s preferred language (English/Mandarin/Malay/Tamil/other Chinese dialects) by trained interviewers. The sampling approach for participant selection in PHASE has been detailed elsewhere. 16 Briefly, PHASE was initiated in 2009 (Wave 1), interviewing a stratified random sample of 4990 Singapore citizens or permanent residents aged ⩾60 years. Of this group, 3103 participated in 2011–2012 (Wave 2). From this group of Wave 2 participants, Wave 3 entailed interviewing 1572 elderly, who were aged ⩾66 years in 2015; six years older since the initial wave. The reduced sample sizes across different waves were due to mortality or refusal to participate. Only Wave 3 included data on prescription medication use of the participants. After excluding 73 respondents who did not report their exact number of prescription medications taken regularly, the analytical sample comprised 1499 elderly (or their proxies, if the elderly could not respond for health reasons).
Outcome variables: polypharmacy and medication non-adherence
We defined polypharmacy as taking ⩾5 prescription medication (yes; no), using the most common threshold. 4 We assessed it using a self-reported question: ‘How many different prescription medications do you (subject) take on a regular basis?’ Those not taking any medications were included under ‘no’. We excluded health supplements that were not prescribed by a doctor and/or traditional medications. In sensitivity analyses, we operationalized polypharmacy in two other ways, that is, using the second most common threshold (⩾6 medications) 4 and including only those taking ⩾1 medication.
Medication non-adherence was assessed for 1104 respondents, after sequentially excluding elderly who were not taking any medications (n=270; they were not at risk of medication non-adherence) and had a proxy respondent (n=125; non-adherence could not be determined from proxies). We used two questions to ascertain non-adherence: ‘In the last one month, how often did you take your medications as prescribed by the doctor?’ (all the time; nearly all the time; most of the time; about half the time; less than half of the time); 17 and ‘At times, do you forget to take your prescription medications?’ (yes; no). 18 Respondents were classified as adherent only if they answered ‘all the time’ to the first question and ‘no’ to the second question.
Potential risk factors of polypharmacy
We assessed the association of a range of socio-demographic (age, gender, ethnicity, highest educational status, and housing type), health (number of chronic diseases and activities of daily living (ADL) limitations) and past healthcare use (emergency room visit (six-month), doctor visit (three-month), number of hospitalizations (one-year) and use of the Pioneer Generation card (three-month; this card provides eligible elderly with benefits and subsidies at public and selected private clinics [https://www.pioneers.sg/en-sg/Pages/Overview.aspx])) variables with polypharmacy. The potential risk factors, all self-reported, were chosen based on relevant international literature, their importance in the context of ageing and health in Singapore and their availability in the PHASE dataset.
Polypharmacy and medication non-adherence: covariates
To investigate the relationship between polypharmacy and medication non-adherence, we controlled for age, gender, ethnicity, educational status and cognitive status, assessed using the Short Portable Mental Status Questionnaire, which has been validated. 19
Statistical analysis
We assessed differences in the distribution of socio-demographic, healthcare and health variables by polypharmacy status using chi-square test or t-test, as applicable. We used multivariable logistic regression to determine the risk factors of polypharmacy and establish the association between polypharmacy and medication non-adherence. We applied survey sampling weights specific to PHASE Wave 3. All analyses were done in STATA 14.
Results
The number of prescription medications being taken on a regular basis by community-dwelling elderly Singaporeans ranged between 0 and 18 (Supplementary Figure 1 in Supplementary Material online). The weighted prevalence of polypharmacy was 14.5%. In unadjusted analysis (Table 1), old age, ethnic minority groups, lower educational status, past hospitalization, past emergency room visit, past doctor visit, use of Pioneer Generation card and a higher number of chronic conditions and ADL limitations were significantly associated with a higher likelihood of polypharmacy.
Distribution of socio-demographic, healthcare use and health variables, and of polypharmacy by these variables among community-dwelling elderly Singaporeans.
ADL: activities of daily living
In multivariable analysis (Table 2), men and those aged ⩾85 years were twice more likely to have polypharmacy as women and those aged 66–69 years, respectively. Malays and Indians were more susceptible to polypharmacy than the Chinese. While healthcare use variables were not associated with polypharmacy, the odds of polypharmacy were higher among elderly with multiple morbidities (versus 0–1 chronic illness) and those with more ADL limitations.
Association of potential risk factors with polypharmacy among community-dwelling elderly Singaporeans: multivariable logistic regression (N = 1499).
ref.: reference; ADL: activities of daily living
Nearly 30% of the elderly reported being non-adherent to their prescription medications. Polypharmacy was associated with a higher likelihood of medication non-adherence. Furthermore, compared with those with secondary education or better, those without formal education were more likely to be non-adherent (Table 3).
Association of polypharmacy with medication non-adherence among community-dwelling elderly Singaporeans: multivariable logistic regression (N = 1104).
ref.: reference
In sensitivity analyses, the weighted prevalence of polypharmacy was, as expected, lower, at 7.5%, for the threshold of ⩾6 prescription medications, and 18% when restricting the sample to those taking ⩾1 medications. The risk factors of polypharmacy in multivariable models were largely similar to those identified in the main analysis – gender, ethnicity and the number of chronic diseases were consistent risk factors of polypharmacy (Supplementary Table 1 in the Supplementary Material). The significant association of polypharmacy with medication non-adherence was maintained when defining polypharmacy using the threshold of ⩾6 prescription medications (Supplementary Table 2 in the Supplementary Material). When the sample was restricted to those taking ⩾1 medications, the association between polypharmacy and medication non-adherence was identical to the main analysis. This was because elderly not taking any prescription medications were already excluded in the main analysis (as they could not report medication non-adherence).
Discussion
In our study, the weighted prevalence of polypharmacy among community-dwelling elderly Singaporeans was 14.5%. This is relatively lower than the prevalence reported for other Asian countries such as Malaysia (45.9%; polypharmacy threshold: ⩾5 medications, though the study also included non-prescription medications) 20 and South Korea (86.4%; ⩾6 medications), 21 and Western countries such as Germany (26.7%; ⩾5 medications) 6 and the United Kingdom (UK) (30.5%; ⩾5 medications). 22 Our estimate is lower, possibly because we categorized elderly who reported not taking any medications as ‘no polypharmacy’ to arrive at a national prevalence of polypharmacy, whereas the prevalence estimates in all the aforementioned studies, except the study conducted in the UK, included mostly (>99%) individuals taking ⩾1 medications.6,20,21 Although we did the same in our sensitivity analysis, the prevalence estimate increased only marginally, to 18%, which is still lower than that reported elsewhere.
We observed age ⩾85 years to be a risk factor for polypharmacy. Given their age (⩾85 years), the oldest-old may be susceptible to more chronic diseases and greater severity of their health conditions than those younger (age 66–69 years), leading to greater consumption of prescription medications. While we adjusted for the number of chronic diseases and proxy measures of disease severity (number of ADL limitations and past healthcare use) in our multivariable model, we do acknowledge residual confounding by disease severity as an alternative explanation. Previous findings on gender and medication use have been inconsistent. A study from Brazil reported that women aged 60–79 years used more medications than men of the same age; 23 whereas, akin to our findings, studies from Malaysia and South Korea reported polypharmacy to be more prevalent among older men.20,21 Such disparities could be due to country-specific differences in the distribution of chronic diseases by gender. Another reason could be gender differences, favouring women, in preventive health behaviours that guard against the escalation of illnesses. A recent national population health survey showed that Singaporean women exhibited better health screening behaviours. 24
Ethnicity, an important part of one’s cultural identity, can influence general patterns of medication use. An American study hypothesized that acculturation, or incompatibility between ethnic-specific health beliefs and the healthcare system, might explain some variances in polypharmacy, although an association was not found eventually. 25 In Malaysia, a multi-ethnic country like Singapore, it was reported that Indian elderly had a higher risk of polypharmacy compared with the Chinese and Malay elderly. 20 In our study, ethnic minority groups in Singapore, viz. Malays and Indians, were more susceptible to polypharmacy than the Chinese, controlling for other socio-demographic, health and healthcare use variables. According to the Singapore National Health Survey 2010, 26 Malays consistently had the lowest rates of screening for cancers, diabetes, hypertension and high cholesterol. Undetected or unaddressed health problems can potentially lead to greater disease severity and polypharmacy risks in old age among ethnic minority groups in Singapore. However, not much else is known about the specific reasons behind the higher rates of polypharmacy among ethnic minority groups in Singapore – this should be explored in future studies. Nonetheless, efforts to minimize polypharmacy should focus on ethnic minority groups.
In a previous study, lower socio-economic status (SES) was associated with polypharmacy, with more chronic diseases being the explanatory mechanism. 22 In our study, SES markers, such as educational status and housing type, were not associated with polypharmacy. We opine that the health and healthcare use variables likely mediated the relationship of these SES markers with polypharmacy, rendering the associations non-significant in our multivariable analysis.
International studies have reported multiple morbidities to be associated with polypharmacy.20,21 Parallel to such studies, we also observed the risk of polypharmacy to be higher among those with ⩾2 chronic conditions and to rise with an increase in the number of ADL limitations. These results allude to the importance of pharmacotherapy in chronic disease management, and the centrality of medication management among chronically-ill elderly. Measures of healthcare use, such as the frequency of ambulatory care visits and hospital admissions, have been previously linked with polypharmacy,21,27 a potential explanation being that hospitalization brings new diagnoses and medications on top of existing ones. 27 However, our study did not find an association between past healthcare use variables and polypharmacy in the adjusted analyses.
Implications
Our findings indicate that among community-dwelling elderly in Singapore, those older (⩾85 years), men, those of poorer health status and those belonging to ethnic minority groups (Malay or Indian) have a higher risk of polypharmacy. These findings have an immediate translational impact. Local healthcare institutions can design relevant interventions meant to address polypharmacy, targeting patients who fit these criteria. Healthcare policymakers and practitioners can plan and conduct home-based medication reviews for the oldest-old. Policymakers and practitioners can also seek out ethnic minority groups by promoting medication management programmes at selective locations, such as religious establishments. Ongoing public health initiatives, such as health screenings and exercise regimes, have been implemented at mosques. 28 It has been suggested that visiting patients’ homes and religious establishments could motivate patients and improve their health and medication-taking habits. 28 Such targeted interventions can complement broad community-level initiatives, such that younger elderly, women and the Chinese are not entirely excluded.
The significant association between polypharmacy and medication non-adherence also provides some evidence to support the development of a comprehensive de-prescribing framework in Singapore. De-prescribing is the process of discontinuing drugs, with the aim of reducing inappropriate polypharmacy and improving patient outcomes. 29 In Singapore, a pharmacist-initiated de-prescribing initiative received a 77.9% acceptance rate from physicians. 3 In a separate study, 93.4% of the patients reported a willingness to stop a medication if approved by their doctor. 30 However, some barriers have prevented a large-scale implementation of de-prescribing. These barriers include a lack of a specific approach to de-prescribing and insufficient knowledge; one-third of junior doctors expressed reluctance to de-prescribe. 31 Thus, efforts to develop de-prescribing guidelines for elderly with multiple morbidities and co-morbidities would facilitate and allow prescribers to withdraw treatment confidently whenever appropriate.
De-prescribing efforts may also benefit from integrating a non-dispensing pharmacist (NDP), also known as a clinical pharmacist, in primary care practice. NDPs generally provide pharmacy services beyond medication dispensing, such as medication education and medication review. 32 These services target patients at risk of drug-related problems, such as polypharmacy. A recent systematic review demonstrated a positive association between the degree of integration of NDPs in primary care and health outcomes, which include resolved medication-related problems and improved quality of care for chronic disease management. 32 Although similar pharmacy initiatives have been progressively introduced in Singapore’s public primary care institutions, such services are lacking in the private primary care sector. Integrating NDPs into Singapore’s primary care settings may see the decline of polypharmacy among community-dwelling elderly.
Strengths and limitations
Our study has its strengths. It is the first community-based study in Singapore to ascertain the prevalence and risk factors of polypharmacy, that too using a national sample. Its findings, pertaining to community-dwelling, non-institutionalized elderly, are directly relevant to future plans centred around bringing healthcare beyond the hospital to the community, and empowering patients to handle medications in a community setting. 15 In 2017, Singapore’s public healthcare system was re-organized into three integrated clusters to anchor care in the community, through allowing partnerships to be built among acute hospitals, community hospitals and primary care providers. 33 Presently, there are plans for pharmaceutical care services, including medication management and de-prescribing, to be extended to the community. 15 When such plans are implemented with consideration of our findings, they could benefit a growing number of elderly who are currently seeking chronic care from primary healthcare settings. In Singapore, between 2010 and 2014, the percentage of elderly visiting General Practitioner clinics almost doubled, from 6% to 11%, and patients with chronic disease increased from 12% to 20%. 34 In 2014, more than 50% of polyclinic visits were chronic disease patients, and these figures are expected to continue growing. 34 Hence, our community-based study is important in guiding this shift towards more inclusive community care.
We must consider these findings with a few limitations in the measurement of some study variables. First, the number of medications, chronic diseases and functional limitations were based on self-report. Interpretations of such questions were largely left to the participants. We were unable to tally the reported number of medications with patients’ medical records, and thus cannot perform further checks on the accuracy of this variable. Second, the PHASE dataset did not contain an explicit measure for disease severity; we considered the number of ADLs and past healthcare use as proxy measures. Third, some associations found in this study might be slightly weak, suggesting a need to conduct further research on polypharmacy among Singaporean elderly. Finally, a validated scale was not utilized in ascertaining medication non-adherence. Nevertheless, we made our measure more comprehensive by using questions that capture both intentional and unintentional non-adherence. The first question – about how often participants were adherent to their prescription medications 17 – covered a range of possible factors for intentional non-adherence, while the second question – on whether they sometimes forgot to take their prescription medications 18 – reflected greater involuntariness.
Conclusion
The prevalence of polypharmacy among community-dwelling elderly in Singapore is lower than that reported in other countries; however, polypharmacy is associated with medication non-adherence. Subgroups of the elderly who are at a higher risk of polypharmacy, defined by gender, health status, age and ethnicity, will benefit from medication review and de-prescribing services.
Supplemental Material
Supplementary_Material_Submission_PoSH – Supplemental material for Polypharmacy among community-dwelling elderly in Singapore: Prevalence, risk factors and association with medication non-adherence
Supplemental material, Supplementary_Material_Submission_PoSH for Polypharmacy among community-dwelling elderly in Singapore: Prevalence, risk factors and association with medication non-adherence by Yi Wen Tan, Sumithra Suppiah, Mary Ann C Bautista and Rahul Malhotra in Proceedings of Singapore Healthcare
Footnotes
Author contribution
TYW conducted the analysis and wrote the first draft of the manuscript, SS edited the manuscript, BMAC conceptualized the research and edited the manuscript, MR conceptualized the research, guided the analysis and edited the manuscript. All authors approved the final version of the manuscript.
Availability of data and materials
The datasets generated and/or analyzed during the current study are available from the corresponding author.
Declaration of conflicting interests
The authors declare that there is no conflict of interest.
Ethical approval
Ethical approval for this study was obtained from the National University of Singapore’s institutional review board (NUS-IRB Ref: B-14-235).
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study uses data from Wave 3 of the Panel on Health and Ageing of Singaporean Elderly (PHASE), which was funded or supported by the Singapore Ministry of Health’s National Medical Research Council under its Clinician Scientist – Individual Research Grant – New Investigator Grant (NMRC-CNIG-1124-2014) and the Duke-NUS Geriatric Research Fund.
Informed consent
Written informed consent was obtained from all subjects before the study.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
