FormalPara Key Points

The prevalence of polypharmacy in people aged 65 years and older remains high when compared with those younger than 65 years, putting them at even greater risk for adverse drug events.

Women are at increased risk of having potentially inappropriate medications (PIMs), and living in rural areas is associated with increased risks of both polypharmacy and PIMs.

Effective medication management and prescribing practices should be a top priority for primary care physicians when it comes to aging populations.

1 Introduction

Polypharmacy is the simultaneous use of five or more medications, whereas excessive polypharmacy is often defined as 10 or more medications [1,2,3,4]. Polypharmacy is more common in elderly patients with multimorbidity. While one can benefit from a proper combination of medications to manage multiple conditions, there is a risk of being prescribed potentially inappropriate medications (PIMs) such as some opioids and benzodiazepines, that could result in adverse pharmacological interactions, poor medication adherence, increased hospitalizations, and higher mortality [5]. Older adults have a lower metabolic reserve and are more vulnerable to adverse drug reactions, which are worsened with high-risk medications.

In the US, the number of people 65 years and older is projected to nearly double from 56 million in 2020 to 95 million in 2060 [6], highlighting the importance of identifying risk factors associated with polypharmacy and PIMs. Previous studies have shown that the prevalence of polypharmacy is common in older populations and ranges between 45 and 65% in the US [7, 8] and up to 90% in some populations worldwide [1]. While factors associated with polypharmacy and PIM usage have not been consistently described, they are often seen as linked with older age, living in a rural area, sex, race and ethnicity, and lifestyle factors [1, 7, 9,10,11]. More studies on this subject are necessary to produce a more precise and reliable estimate of the prevalence of polypharmacy to understand the burden and factors associated with it. This study describes the factors associated with prevalent polypharmacy and PIMs usage among people aged 65 years and older in the US.

2 Methods

2.1 Data Source and Study Population

We used annual cross-sectional study data from the CDC's National Ambulatory Medical Care Survey (NAMCS) collected by the National Center for Health Statistics [12]. The NAMCS is a nationally representative survey that collects a sample of visits to non-federally employed office-based physicians who provide direct patient care but are not in the specialties of anesthesiology, pathology, and radiology. Data are released to the public without any personally identifiable information. The NAMCS database provides a description of the characteristics of patients who have visited physician offices, along with their medical conditions, treatments, medications, and services offered. From 2010 to 2011, the survey collected a maximum of eight medications per patient, which increased to ten medications in 2012 and 2013. From 2014 onwards, the number of medications expanded to 30 medications. In order to keep the number of medications comparable over the years, analyses were limited to the first eight recorded medications.

We included adults aged 65 years and older as reported during their visits between 2010 and 2016. Out of 280,511 visits during the study period, only 81,295 visits with patients aged 65 years and older were included (Fig. 1). Patients were later categorized as having polypharmacy or not by the number of medications that were prescribed at their visit. Missing data were removed from analyses since overall missing values were <5%. Our final analytical sample was 81,295.

Fig. 1
figure 1

Flowchart of study sample

Full size image

2.2 Variables

We used a general definition of polypharmacy as having five or more medications prescribed or renewed at an office visit for this study, hence patients were classified into two categories, no polypharmacy (fewer than five medications) and polypharmacy (five or more medications). PIMs were defined as high-risk medications in the elderly (chronic use of opioids, benzodiazepines, tricyclic antidepressants, muscle relaxants, antipsychotics, and medications with significant anticholinergic effects) as outlined in the Beers criteria [13,14,15]. The Beers criteria includes a list of inappropriate medications whose risks outweigh the benefits in the geriatric population. The BEERS list is extensive, but we limited our PIMs to classes of medications that are at the highest risk and associated with an increased risk of falls and cognitive impairment in the elderly. Characteristics of the sample such as age, sex, race, and ethnicity, smoking status, geographic settings, source of payment, provider specialty, overweight status, and the number of chronic conditions were collected in the survey and were self-reported. The total number of chronic conditions is a pre-calculated variable that is derived from a list of conditions asked about in the survey. Chronic conditions were stratified arbitrarily into fewer than five and five or more.

2.3 Statistical Analysis

Descriptive analyses were stratified by polypharmacy use. Multivariable logistic regression was used to examine the effects of each selected risk factor individually on polypharmacy and prescribing of PIMs and then adjusting for the same covariates (i.e., age, race/ethnicity, and geographic setting) in the adjusted model. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were reported. Analyses were conducted using SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA) with a significance level of 0.05. The study followed STROBE guidelines for reporting of cross-sectional studies [16].

3 Results

3.1 Participant Characteristics

Out of 280,511 patients identified, 199,216 patients were excluded as they were younger than 65 years of age at the time of the visit. As a result, a total of 81,295 patients aged ≥ 65 years and older were included in the analysis. Of this analytic sample, 45.1% were men and 54.9% were women. The mean (SD) age of the sample is 75 ± 7.3 years. Prevalence of polypharmacy and prescription of PIMs was 42.0% and 18.8%, respectively, between 2010 and 2016. See Table 1 for a list of patient characteristics.

Table 1 Characteristics of US ambulatory care visits among patients 65 years old or older by polypharmacy status and occurrence of potentially inappropriate medications, unweighted, 2010–2016
Full size table

3.2 Associations Between Risk Factors and Polypharmacy, PIMs

After adjusting for the other variables included in the model, participants in rural areas had higher odds of both polypharmacy (aOR: 1.15, 95% CI 1.07–1.23) and PIMs (aOR: 1.19, 95% CI 1.09–1.29) than participants in urban areas. Compared with men, women were at greater odds of being prescribed PIMs (aOR: 1.31, 95% CI 1.23–1.40). In the study period 2010–2016, for every 5-year increment in age, the odds of being on polypharmacy increased by 8% (95% CI 1.06–1.10) but the odds of being on PIMs decreased by 3% (95% CI: 0.95–0.99). See Table 2 for the factors associated with polypharmacy and PIM usage.

Table 2 Factors associated with polypharmacy and PIM usage in the elderly population in the US (2010–2016)
Full size table

4 Discussion

Polypharmacy is linked to adverse outcomes due to drug side effects, drug–drug interactions, falls, frailty, and mortality. Older adults are particularly vulnerable to polypharmacy risks. Our study using a nationally represented primary care database demonstrated a high prevalence of polypharmacy (42%) and PIM usage (18%) among adults aged 65 years and older. Recent estimates on the prevalence of polypharmacy in the general population has been estimated to be 37% (95% CI 31–43) [8].

Our study shows significant variations in polypharmacy and PIMs by gender and urbanization level. Rural residence was an independent risk factor for polypharmacy and PIMs. This could be due to prevalence of chronic disease, low educational attainment, and barriers in access to healthcare which could mean polypharmacy goes unchecked without regular visits and monitoring in rural areas [17, 18]. Prior studies examining the relationship between urbanization and polypharmacy have been inconsistent and thus more studies are needed to investigate the quality of care in rural regions.

Greater PIMs usage in women is consistent with prior studies, thought to be related to women having a longer life expectancy and differences in patterns of chronic conditions, healthcare utilization, and health behaviors [1, 19]. The odds of being on polypharmacy increasing with advancing age is likely related to an increase in chronic health conditions. Decreased odds of PIMs could result from the many clinical tools available to detect PIMs (e.g., Beers and STOPP criteria) and more regular doctor visits allowing for comprehensive care and planning. The Beers list has had some success in reducing the number of PIMs and unplanned hospitalization in older people [20].

Regular access to primary care physicians for medication reconciliation and deprescribing would be helpful to reduce polypharmacy [21]. Patients perceive polypharmacy as a considerable burden but are reluctant to discuss it with their providers due to fear of worsening symptoms with deprescribing. Physicians should prioritize discussion and consider lowering doses and deprescribing where possible, especially among elderly patients. Aside from primary care physicians, clinical pharmacists have also been shown to play a role in creating a positive impact on the quality of prescribing medications and on polypharmacy reduction through a medical review form intervention [22,23,24]. For example, the Pharmacist Consultant program in Slovenia was beneficial for elderly people with multimorbidity in being more cost effective, with fewer drug-related problems, and better treatment guideline adherence [24].

The main strength of our study is a large, diverse, primary care sample using the NAMCS database which has been previously validated as being representative of outpatient visits at a national level over time. However, this study has some limitations to consider. Medications provided at visits might only be related to the visit and not necessarily include all patients’ medications. The results of PIMs could have thus been underestimated because medications were limited to the first eight recorded. Including lifestyle factors could have been informative and meaningful in describing their associations with polypharmacy and PIMs, but due to missing data, they were excluded as covariates. Lastly, our study is cross-sectional in design thus further prospective studies are needed to establish temporality.

5 Conclusions

Effective medication management should be a top priority for primary care physicians with aging patient populations. Our study highlights a high prevalence of polypharmacy and prescribing of PIMs among elderly patients. This warrants a focus on the elderly, particularly women and people living in rural areas, aimed at reducing these rates. Ongoing surveillance of polypharmacy trends and predictors is important to develop strategic interventions that can improve the quality of life for elderly patients in all settings.