Trends and Socioeconomic Disparities in Polypharmacy and Potentially Inappropriate Medication Use among 66-Year-Olds in Korea: A Nationwide Study, 2012–2021
Article information
Abstract
Background
With increasing life expectancy, the number of older adults with multiple chronic conditions requiring complex medication regimens is growing, raising concerns about polypharmacy and potentially inappropriate medication (PIM) use. This study investigated trends in polypharmacy and PIM use among 66-year-olds in South Korea from 2012 to 2021, considering participant characteristics, to inform interventions and policies.
Methods
A repeated cross-sectional study was conducted using the National Health Insurance Services database covering approximately 97% of Koreans. We included 3,397,044 individuals aged 66 who underwent the National Screening Program for Transitional Ages between 2012 and 2021. Polypharmacy was defined as the use of ≥5 medications for ≥90 days annually, hyper-polypharmacy as ≥10 medications, and PIM use as ≥1 PIM for ≥28 days. Trends were analyzed by sex, frailty, comorbidity, income, insurance type, and residence.
Results
Polypharmacy prevalence increased from 32.0% in 2012 to 35.4% in 2021, and hyper-polypharmacy also rose. PIM use slightly decreased from 55.7% to 53.7%. Higher rates of polypharmacy and PIM use were observed among rural residents, medical aid beneficiaries, and those with lower income. Despite improvements in comorbidity and frailty, socioeconomic disparities widened, particularly among medical aid beneficiaries. Frequently prescribed PIMs included NSAIDs, PPIs, muscle relaxants, and anxiolytics/hypnotics.
Conclusion
While PIM use slightly decreased over the study period, it remained above 50%, and polypharmacy prevalence increased among older adults in Korea. Socioeconomic disparities in medication use persist, highlighting the need for targeted interventions and policies to promote safe medication use among vulnerable groups.
INTRODUCTION
The global proportion of older adults is rising due to increased life expectancy and declining birth rates.1) This trend is accompanied by a rise in the number of frail individuals with multimorbidity.2) Socioeconomic development has improved access to healthcare systems, and advancements in medical technology have led to increased utilization of healthcare services. As societies continue to age, the frequency of healthcare contacts among older adults is expected to grow.3) This increase in healthcare utilization often results in more medication prescriptions, and older adults with multiple chronic conditions are particularly vulnerable to polypharmacy, where they are exposed to the simultaneous use of multiple medications.4)
All medications have potential side effects, and older adults are more susceptible to these adverse effects compared to younger individuals.5) Age-related physiological changes affect pharmacokinetic processes (absorption, distribution, metabolism, and excretion) and pharmacodynamic responses, often leading to different adverse outcomes than those observed in younger, healthier adults. Moreover, frail older adults are more likely to suffer from frequent and severe adverse drug reactions due to underlying pathologies.6) Multimorbidity increases the likelihood of drug-disease and drug-drug interactions, further complicating medication management.7)
Among the various medications, those that frequently cause adverse effects in older adults and whose risks outweigh their benefits are classified as potentially inappropriate medications (PIMs).8) Efforts to reduce polypharmacy and PIM prescriptions have led to the development of PIM lists and guidelines, such as the American Geriatrics Society's Beers Criteria and the European Geriatric Medicine Society's STOPP/START Criteria, which recommend safer alternatives.9,10) Older adults prescribed multiple medications require careful attention to avoid suboptimal treatment that includes PIMs. Despite these efforts, studies from various countries report high rates of polypharmacy (30.2%–61.7%) and PIM use (23.6%–47.0%) among older adults.11,12)
To mitigate polypharmacy and the prescription of PIMs in older adults, and to establish optimal medication guidelines and better policy alternatives, it is essential to understand the status and temporal trends of polypharmacy and PIM use.13) Analyzing these trends requires considering not only frailty and comorbidity, which significantly influence medication use, but also demographic characteristics of the population.14) Although there have been cross-sectional studies and small-scale longitudinal analyses examining frailty, polypharmacy, and PIM prescriptions, there is a lack of large-scale population-based studies. This study specifically focuses on the 66-year-old population, as this age represents a critical transition into older adulthood in South Korea. It is not only the age of eligibility for the National Screening Program for Transitional Ages (NSPTA) but also often coincides with major life events such as retirement and subsequent changes in social and economic status. Medically, it is a period where chronic conditions that may have developed in middle age begin to accumulate, often marking the onset of significant polypharmacy. Analyzing medication patterns at this pivotal age is therefore crucial for developing early, preventative strategies to mitigate medication-related harm throughout the later stages of old age. Therefore, we conducted a comprehensive analysis of polypharmacy and PIM prescription rates, as well as participant characteristics including frailty and comorbidity, and demographic characteristics among 66-year-old participants of the NSPTA in Korea from 2012 to 2021.
MATERIALS AND METHODS
Data Source, Study Design, and Study Population
This study utilized data from the National Health Insurance Services (NHIS) database, which encompasses approximately 97% of the South Korean population. The NHIS database provides comprehensive information, including prescription records, details of prescribing institutions, insurance qualifications, sex, residential information, income-based insurance premiums, medical diagnoses, health examination data, and long-term care insurance records.15) The NHIS operates the NSPTA for all insured individuals reaching the age of 66, a pivotal age marking the transition into older adulthood. The NSPTA includes comprehensive assessments such as blood tests, urine tests, radiological examinations, mental health screenings for depression and cognitive impairments, surveys on activities of daily living, and physical function tests including gait and balance assessments.16)
To compare and analyze the frequency of polypharmacy and PIM prescriptions, as well as participant characteristics such as frailty and comorbidity, from 2012 to 2021, the authors conducted a repeated cross-sectional study. This study included all individuals who underwent the NSPTA at 66 years old each year. Participant characteristics analyzed included sex, residence (categorized as metropolitan city, city, county), income level (classified into quintiles), type of health insurance (National Health Insurance or Medical Aid), Charlson Comorbidity Index (CCI), frailty, and long-term care grade. Healthcare utilization characteristics examined included hospital admissions and number of outpatient visits.
Polypharmacy and Potentially Inappropriate Medication
Polypharmacy was defined as the use of five or more distinct medications prescribed for at least 90 days within a year, based on claims data.17) Similarly, the use of 10 or more medications prescribed for at least 90 days within a year was defined as hyper-polypharmacy. The PIM list for this study was established by consolidating two institutional PIM lists from the researchers’ affiliated hospitals. Both of these lists were originally based on the Beers Criteria but had been independently modified to fit each hospital’s specific circumstances, such as by removing medications not available in their electronic medical record (EMR) systems. The research team unified these two lists into a single consensus list through a series of roundtable discussions. For the purpose of this study, injectable medications and topical agents were excluded. The final PIM list is presented in Supplementary Table S1. Within the study framework, the use of PIMs was defined as the prescription of one or more PIMs for a cumulative period of at least 28 days during the year of participation in the NSPTA.
Frailty and Charlson Comorbidity Index
The authors analyzed the longitudinal trends of polypharmacy and PIM use among cohorts of individuals who underwent the NSPTA annually from 2012 to 2021, employing a repeated cross-sectional methodology. Additionally, the study sought to examine the significant influence of comorbidity and frailty on medication use patterns. Comorbidity was assessed using the CCI, which was calculated by summing the weights of diagnostic codes present in the participants’ claim data from the year preceding the index date of NSPTA.18) In alignment with previous research, we constructed a frailty index using data from the NSPTA.19) The frailty index was computed from data on medical history in the past year from claim data, physical and psychological function, disability, and biomedical or laboratory measures collected during the NSPTA. Each deficit was coded as 0 (absent) or 1 (present), and the index was calculated as the proportion of deficits present. The frailty index ranged from 0 to 1, with higher values indicating increased frailty. Participants were stratified into four categories according to their frailty index scores: robust (<0.15), pre-frail (0.15 to <0.25), mild frail (0.25 to <0.35), and moderate to severe frail (≥0.35).
Statistical Analysis
This study aimed to evaluate the prevalence of polypharmacy and PIM prescriptions among subjects from 2012 to 2021. In addition, it investigated the prevalence of polypharmacy and PIM prescriptions according to subject characteristics and healthcare utilization patterns. The study also identified the top 10 most frequently prescribed PIMs. Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), with results considered statistically significant at a p-value of less than 0.05.
RESULTS
Participants
As the population of older adults increased, the number of individuals who reached the age of 66 and participated in the NSPTA gradually rose from 248,544 in 2012 to 461,824 in 2021. Over the 10-year period from 2012 to 2021, a total of 3,397,044 participants were included in the analysis (Table 1). Among all participants, 53.7% were female, 45.1% resided in metropolitan cities, 44.0% lived in cities, and 10.9% lived in counties. Additionally, 96.9% of the participants were covered by the National Health Insurance, while 3.1% were recipients of medical aid. The proportion of participants who had experienced hospitalization within one year prior to the index date showed a slight increase over the study period. However, the proportion of those who had more than 30 outpatient visits annually decreased.
Characteristics of participants by year 2012–2021 (unit: %)
In terms of comorbidities, 66.4% of participants in 2012 had a CCI score between 0 and 1, while 17.6% had a score of 2, and 16% had a score of 3 or higher. Over time, the proportion of individuals with a CCI score of 0–1 gradually increased, reaching a peak of 71.7% in 2021, whereas the proportion with a score of 3 or higher steadily declined to 12.5% in 2021. Due to the revisions in the NSPTA in 2018, certain variables used to calculate the frailty index, such as survey data and components of the physical and psychological tests, were modified. As a result, the frailty index was trackable over a 6-year period from 2012 to 2017. During this period, the proportion of robust individuals increased from 37.9% in 2012 to 42.3% in 2017, while the percentage of mild frail and moderate to severe frail individuals decreased from 6.3% and 0.5% in 2012 to 4.7% and 0.4% in 2017, respectively.
Trends in Polypharmacy Prevalence and Differences by Characteristics
The prevalence of polypharmacy in the study population varied between 31.6% and 35.4% between 2012 and 2021, while the prevalence of hyper-polypharmacy ranged from 7.4% to 8.8% (Table 2). Analyzing the trends by year, the prevalence of polypharmacy decreased from 32% in 2012 to 31.6% in 2015, before rising again to 35.4% in 2021 (Fig. 1). Similarly, the prevalence of hyper-polypharmacy declined from 7.7% in 2012 to 7.4% in 2015, but subsequently increased, reaching 8.8% in 2021. As the aging population expanded, and more individuals underwent the NSPTA at 66 years, the number of participants classified as experiencing polypharmacy increased from 79,638 in 2012 to 163,337 in 2021. Likewise, the number of individuals classified as experiencing hyper-polypharmacy rose from 19,148 in 2012 to 40,612 in 2021.
Polypharmacy prescription rate by participants characteristics (unit: %)
Annual trends in polypharmacy and hyper-polypharmacy prescription rates: (A) overall, (B) sex, (C) residence, (D) insurance type.
An analysis of polypharmacy rates by participant characteristics revealed that from 2012 to 2015, the rates were similar between males and females. However, from 2015 onward, the proportion of males prescribed polypharmacy (32.1%–37.1%) increased more substantially than that of females (31.3%–33.9%). Individuals residing in cities (32.0%–35.5%) and counties (35.9%–39.9%) demonstrated higher rates of polypharmacy compared to those in metropolitan areas (30.2%–34.2%). Regardless of residence, the proportion of individuals using polypharmacy showed a consistent upward trend. In terms of income levels, the lowest income quintile showed the highest rates of polypharmacy (35.5%–39.2%), while the second, third, and fourth quintiles displayed similar rates, and the highest income quintile had the lowest rates (29.3%–32.6%), with the smallest increase over time compared to other groups.
When comparing polypharmacy rates by type of health insurance coverage from 2012 to 2021, 58.6% to 69.7% of medical aid recipients were prescribed polypharmacy, significantly higher than those with health insurance (31.1%–34.4%), with the increase in rates being much more pronounced among medical aid recipients. Regarding comorbidity burden, individuals with a lower burden (CCI score 0–1) had the lowest rates of polypharmacy (21.1%–26.1%), whereas those with a higher comorbidity burden (CCI score ≥3) exhibited much higher rates (63.2%–68.4%). A similar pattern was observed with frailty from 2012 to 2017; robust individuals had the lowest rates of polypharmacy (25.7%–27.0%), while those classified as moderate-to-severe frail had the highest rates (52.0%–58.4%).
Trends in PIM Prevalence and Differences by Characteristics
Between 2012 and 2021, the proportion of participants prescribed PIMs ranged from 53.7% to 56.1% (Table 3, Fig. 2). By year, the rate was 55.7% in 2012 but gradually declined to 53.7% in 2021. In contrast, the absolute number of individuals receiving PIM prescriptions increased with the aging population, rising from 138,427 in 2012 to 248,225 in 2021, similar to trends observed with polypharmacy.
PIM prescription rate by participants characteristics (unit: %)
Annual trends in potentially inappropriate medication prescription rates: (A) overall, (B) sex, (C) residence, (D) insurance type.
An assessment of PIM prescription rates by participant characteristics revealed that females had slightly higher rates (54.2%–57.3%) compared to males (52.8%–54.7%), though the sex gap narrowed over time. Individuals residing in smaller cities (53.6%–56.5%) and counties (58.7%–62.1%) were prescribed PIMs more frequently than those in metropolitan areas (52.5%–54.1%). Regarding income levels, the highest PIM prescription rates were observed among those in the lowest income quintile (56.5%–58.8%), while the second, third, and fourth quintiles displayed similar rates. The lowest rates were recorded in the highest income quintile (50.6%–54.0%).
Analysis by insurance type showed that medical aid beneficiaries had significantly higher rates of PIM prescriptions (75.7%–79.4%) compared to National Health Insurance beneficiaries (53.0%–55.4%). Moreover, the rate among health insurance beneficiaries decreased over time, from 55.0% in 2012 to 53.0% in 2021, while medical aid beneficiaries experienced an increase from 75.7% to 79.4% during the same period. In terms of comorbidities, individuals with a CCI score of 0–1 demonstrated lower PIM prescription rates (46.3%–48.7%), whereas those with a score of 3 or higher had the highest rates (72.2%–77.8%). A similar trend was observed with frailty during 2012 to 2017, where robust individuals had lower PIM prescription rates (50.0%–51.2%), while moderate-to-severely frail individuals exhibited the highest rates (64.2%–71.1%).
Most Frequently Prescribed PIMs
The list of the most frequently prescribed PIMs remained largely consistent over the 10-year period from 2012 to 2021, with the top 10 medications presented in Table 4. Non-steroidal anti-inflammatory drugs (NSAIDs), proton pump inhibitors (PPIs), skeletal muscle relaxants, and anxiolytics or hypnotics, such as alprazolam and diazepam, were consistently ranked among the most frequently prescribed PIMs. Notably, the prescription frequency of PPIs increased over time compared to other medications.
Top 10 most frequently prescribed PIMs
Ethics Statements
The Institutional Review Board of the National Evidence-based healthcare Collaborating Agency (NECA) approved the study protocol (Approval No. NECAIRB23-001). The requirement for informed consent was waived by the Institutional Review Board because the dataset was de-identified to protect personal information.
DISCUSSION
The researchers conducted a repeated cross-sectional study on approximately 3.4 million individuals aged 66 who underwent NSPTA between 2012 and 2021, as they entered older adulthood. The study revealed that individuals residing in counties had higher rates of polypharmacy compared to those residing in large cities. Additionally, a higher rate of polypharmacy prescriptions was observed among individuals in the lowest income quintile, medical aid beneficiaries, those with multiple comorbidities, and frail individuals. The frequency of PIM prescriptions exhibited similar patterns to polypharmacy use, with higher rates among county residents, those in the lowest income quintile, medical aid beneficiaries, and individuals with multiple comorbidities or frailty.
An analysis of the temporal trends among 66-year-olds in Korea from 2012 to 2021 revealed a gradual decrease in comorbidity burden, an improvement in frailty, and a decline in the proportion of individuals making more than 30 outpatient visits per year. However, the proportion of individuals with polypharmacy increased slightly during this period, as did the frequency of hyper-polypharmacy. The increase in polypharmacy rates was more pronounced among county residents compared to those in large cities, and among individuals in the lowest income quintile compared to those in the highest quintile. Notably, medical aid beneficiaries experienced a substantial increase of over 10% in polypharmacy rates over the ten-year period. While the proportion of PIM users decreased slightly over time in most groups, it increased specifically among medical aid beneficiaries. The list of PIMs remained relatively consistent throughout the study period, with NSAIDs, PPIs, skeletal muscle relaxants, and anxiolytics/hypnotics ranking among the most frequently prescribed PIMs.
The findings of this study illustrate the complex dynamics of medication use among older adults in Korea. Over time, frailty and comorbidity burden among the 66-year-old population improved, and the proportion of individuals requiring frequent outpatient visits declined. However, the proportion of polypharmacy users increased, with a particularly marked rise in polypharmacy and hyper-polypharmacy among vulnerable populations. Moreover, while PIM use decreased in most groups, it increased among medical aid beneficiaries. This concerning trend may be attributable to several factors. Disparities in health literacy could mean that vulnerable populations have a lower understanding of the risks associated with polypharmacy and PIMs. Additionally, regional differences in healthcare, such as limited access to geriatric specialists in rural areas, may lead to less comprehensive medication management. Finally, the low co-payment structure for medical aid beneficiaries in Korea might inadvertently encourage more frequent healthcare visits and trials of various medications, increasing their exposure to polypharmacy and PIMs.
These trends highlight the need for targeted interventions that are specifically tailored to population characteristics. Previous studies have already demonstrated the impact of inequalities in residential area, education, income, and employment on polypharmacy and PIM use.20,21) Our study not only confirms these findings but also reveals a decade-long trend of widening disparities despite overall improvements in comorbidity and frailty, which reinforces the urgent need for policy adjustments to effectively address these inequalities.
The increasing prevalence of polypharmacy in Korea during the study period is consistent with findings from previous studies conducted in other countries.22,23) The increase in the proportion of individuals prescribed polypharmacy and hyper-polypharmacy may be due to the growing number of health conditions and related indicators requiring management, driven by advancements in medical technology, and increasingly stringent single disease-focused guidelines. Previous studies have also noted a significant rise in polypharmacy rates over time, particularly among older adults with cardiovascular disease or diabetes.23,24) Moreover, Korea’s fee-for-service healthcare system, which primarily incentivizes diagnostic testing, monitoring, and medication prescription rather than prevention, education, or lifestyle modification, may have contributed to the rising prevalence of polypharmacy. The limited reimbursement for preventive measures and lifestyle interventions further underscores the potential influence of this system on the increasing rates of polypharmacy.25-27)
The prescription rate of PIMs decreased slightly from 55.7% in 2012 to 53.7% in 2021. Despite this modest decline, the persistently high prevalence of PIM use above 50% represents a critical public health concern, particularly as this was observed in relatively younger older adults (age 66). This suggests that approximately half of the individuals entering older adulthood are exposed to medications with unfavorable risk-benefit profiles, potentially setting a trajectory for increased adverse drug events and healthcare costs as they age. Moreover, considering that PIM prescription rates in other countries range from 30% to 50%, the rate observed in this study remains alarmingly high.12,28,29) Although South Korea has implemented a nationwide real-time Drug Utilization Review (DUR) system since 2015, which has yielded some positive outcomes, our findings strongly suggest that this policy alone is insufficient to adequately address this issue.30) This underscores the urgent need for more proactive and multifaceted interventions. Specifically, the high frequency of prescriptions for NSAIDs, PPIs, skeletal muscle relaxants, and anxiolytics/hypnotics is often linked to patients’ preference for medication over non-pharmacological approaches for even minor symptoms.31) Therefore, a crucial part of the intervention should focus on shifting the perception that all temporary discomforts require pharmacological management. Furthermore, when these PIMs are necessary, a key strategy must be to limit their use to the shortest possible duration. Given the increasing absolute number of individuals taking PIMs due to the aging population, the continued attention of healthcare policymakers is also warranted.32)
While numerous studies have reported on the prevalence and user characteristics of polypharmacy and PIM use in older adults, most investigations to date have been conducted as cross-sectional studies examining specific populations at particular points in time. This has led to difficulties in conducting direct time-series analyses across multiple studies due to variations in the operational definitions of polypharmacy and PIM, specific study populations, and investigation periods. This study, a repeated cross-sectional study, investigated the trends in polypharmacy and PIM prescribing among 66-year-old individuals who underwent the NSPTA annually from 2012 to 2021, using consistent criteria for both. With approximately 80% of eligible individuals participating, this study facilitated time-series analysis as a birth cohort.33) Furthermore, it holds significance as the first nationwide study to investigate trends in the prevalence of polypharmacy and PIM use among older adults according to residence, income level, comorbidity burden, frailty, and insurance type.
However, several limitations should be considered. This study focused on individuals aged 66 who had just entered old age each year for 10 years, and thus may not be representative of the entire older adult population. Injectables and topical medications were excluded from the analysis, and over-the-counter drugs were not included because the analysis was based on health insurance claims data. Therefore, the prevalence of polypharmacy or PIM use may have been underestimated. Additionally, because the analysis was based on prescription records submitted to the NHIS, it was impossible to confirm patient adherence. We also could not validate the recorded diagnoses against clinical records or identify the specific grounds for each prescription. A medication classified as a PIM might have been clinically justified for an individual patient if, for example, no viable alternatives existed, or the benefit outweighed the risk. The study period included the COVID-19 pandemic, which may have affected polypharmacy and PIM prescribing. Furthermore, the findings of this study are specific to the South Korean healthcare system, with its unique insurance policies and prescribing culture, and thus should be generalized to other countries with caution. Finally, although the Beers Criteria and STOPP/START criteria were revised during the study period, resulting in changes to the PIM lists of the two research institutions, the final PIM list was used retrospectively.
In conclusion, an analysis of 66-year-olds from 2012 to 2021 revealed that polypharmacy and PIM use were more prevalent among individuals residing in rural areas, medical aid beneficiaries, and those with lower income levels. Although comorbidity burden and frailty among 66-year-olds improved over time, the proportion of polypharmacy users increased overall, with a notably larger increase among rural residents and socioeconomically disadvantaged groups. During the same period, the proportion of PIM users decreased slightly; however, an increasing trend was observed among medical aid beneficiaries. These findings suggest that health policymakers need to enhance and improve existing systems, such as the DUR program, and urgently develop quality improvement strategies to reduce polypharmacy and PIM use among specific population groups. These tailored interventions, by ensuring safe medication use, could play a crucial role in preventing disease, minimizing disability, and reducing the potential for iatrogenic harm from medications in an aging society.
Notes
CONFLICT OF INTEREST
Among the authors, Hee-Won Jung and Ji Yeon Baek serve as editors of the Annals of Geriatric Medicine and Research. However, they were not involved in any part of the review or decision-making process for this manuscript. Otherwise, no potential conflict of interest relevant to this article was reported.
FUNDING
This study was supported by grants from the National Evidence-based Healthcare Collaborating Agency (NECA-A-23-008) and Seoul National University Bundang Hospital Research Fund (SNUBH 14-2023-0010). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
AUTHOR CONTRIBUTIONS
Conceptualization, SwK, JYB, JEY, HWJ; Data curation, HJ, SJP, KP, JEY; Formal analysis, HWJ, HJ, KP, SJP, JYB, WYK, DK, ML, SgH, JEY, SwK; Writing-original draft, HWJ, HJ, KP, SJP, JYB, WYK, DK, ML, SgH, JEY, SwK; Writing-review & editing, HWJ, HJ, KP, SJP, JYB, WYK, DK, ML, SgH, JEY, SwK.
AVAILABILITY OF DATA AND MATERIALS
The data that support the findings of this study are available from the Korean National Health Insurance Service (NHIS) but restrictions apply to their availability. Access to the data is only possible through approval and oversight by the NHIS, and the data were analyzed onsite at NHIS data analysis centers.
SUPPLEMENTARY MATERIALS
Supplementary materials can be found via https://doi.org/10.4235/agmr.25.0134.
List of potentially inappropriate medications