Medication use patterns and polypharmacy among elderly in Iran: a cross-sectional study using national health insurance claims data

Abstract

Objectives

Polypharmacy, defined as the concurrent use of multiple medications, is a growing concern among the elderly, especially in low-income and middle-income countries such as Iran. This study aims to explore the prevalence and patterns of polypharmacy among the elderly in Iran, using health insurance claims data to identify common drug classes and coprescribed medications, with a focus on informing policy decisions and improving medication management.

Design

Retrospective population-based observational study.

Setting

Nationwide data from the Iran Health Insurance Organization (IHIO) across 24 provinces.

Participants

1 876 527 individuals aged 65 years and older, insured by the IHIO from 2014 to 2017. Individuals with incomplete demographic information or lacking medication records in the database were excluded from the analysis.

Primary and secondary outcome measures

Prevalence and patterns of polypharmacy, demographic factors associated with polypharmacy, and common drug classes used. Medications were classified using the Anatomical Therapeutic Chemical system. Polypharmacy was defined as the use of five or more medications, with cumulative polypharmacy considering total drug use over time, and consecutive polypharmacy focusing on the frequency of monthly drug use. Logistic regression and association rule mining were applied to explore demographic factors and medication patterns associated with polypharmacy.

Results

Of the study population, 74.9% experienced cumulative polypharmacy over 6 months and 64.6% over 1 month, with 7.6% experiencing consecutive polypharmacy. Females and those aged 75–79 were more prone to polypharmacy. Systemic glucocorticoids were the most commonly used medications (50.02%), followed by HMG-CoA reductase inhibitors (42.73%) and platelet aggregation inhibitors (41.92%). Polypharmacy was most strongly associated with medications related to the alimentary tract and metabolism, cardiovascular system, nervous system and blood and blood-forming organs.

Conclusions

Polypharmacy is highly prevalent among the elderly in Iran, with significant variations by gender, age, insurance fund and region. The findings highlight the need for targeted interventions to manage polypharmacy and improve medication safety in this population.

Strengths and limitations of this study.

• This study used a large, nationwide dataset covering 38% of the Iranian elderly population, providing robust statistical power.

• Cumulative and consecutive definitions of polypharmacy were employed, capturing different dimensions of medication use.

• Association rule mining provided novel insights into common drug combinations and their implications for polypharmacy.

• Limitations include the use of claims data lacking clinical details, such as adherence and reasons for prescriptions.

• Our analysis did not account for non-prescription medications, such as over-the-counter drugs and herbal supplements, which may lead to an underestimation of polypharmacy prevalence.

Introduction

Polypharmacy, defined by the WHO as the concurrent use of multiple medications, is a major public health concern worldwide.¹ This issue is particularly pronounced among the elderly and hospitalised patients.² Various studies have documented differing levels of polypharmacy prevalence, influenced by study design, the thresholds used, the time frame and the populations examined. Among the elderly, the prevalence of polypharmacy can vary dramatically, ranging from 4% in outpatient settings to over 96.5% in inpatient settings.³ Key factors contributing to polypharmacy include older age, the presence of multiple comorbidities, poor self-perceived health status, physical activity limitations, depression and pain. Cardiovascular drugs are among the most commonly prescribed medications in this demographic.⁴

The implications of polypharmacy are extensive and multifaceted. From a medicines management perspective, polypharmacy is associated with challenges such as non-adherence, increased risk of drug interactions, medication errors and inappropriate prescribing practices. These issues not only complicate patient care but also contribute to increased healthcare utilisation, reflected in higher hospitalisation rates and medical costs. Adverse health outcomes linked to polypharmacy include falls, malnutrition, frailty, disability, functional decline and adverse drug events.^{5 6}

In low-income and lower-middle-income countries such as Iran, the financial burden of pharmaceuticals is particularly pronounced, with a substantial portion of health expenditures allocated to medications.⁷ The WHO estimates that over half of all medicines are prescribed, dispensed or sold inappropriately, and that a similar proportion of patients fail to use medications correctly.⁸ This inappropriate use of medicines, which includes polypharmacy^{9 10} as well as the overuse of antibiotics, corticosteroids and injectable drugs,¹¹ leads to significant wastage of resources and poses extensive health risks.⁸

As populations age, the prevalence of multiple chronic conditions increases, making older adults more susceptible to drug-related problems such as polypharmacy, inappropriate medication use and adverse drug reactions.¹² Iran’s ageing population underscores the growing need for effective pharmaceutical management and equitable access to necessary medications. Healthcare systems must prioritise policies that promote the appropriate use of medications and monitor polypharmacy trends.⁷

Understanding the prevalence and patterns of polypharmacy is essential for developing strategies to manage this issue effectively. The Iran Health Insurance Organisation (IHIO), one of the principal insurers in Iran, covers about 50% of the population,¹³ offering a valuable resource for large-scale data analysis. Using health insurance claims data provides a cost-effective method to investigate medication use patterns and polypharmacy across different demographics and regions. This study aims to explore the prevalence of polypharmacy and identify the most frequently used drug classes and coprescribed medications among the elderly in Iran, using health insurance claims data from 2012 to 2015. These insights are crucial for informing policy decisions and improving medication management systems to enhance patient outcomes and achieve sustainable development.

Methods

Study design and population

This retrospective population-based observational study analysed prescription data extracted from health insurance claims of individuals aged 65 years and older who were insured by the IHIO across 24 provinces of Iran. The study encompassed a period from 21 March 2014 to 19 March 2017. In total, 19 726 266 prescription records were extracted, representing data from 1 876 527 elderly individuals with complete demographic information and medication records. According to the 2016 census, there were 4 871 518 people aged 65 and older in Iran.¹⁴ Therefore, this study covered approximately 38% of the Iranian elderly population.

The IHIO has five main insurance funds: the Civil Servants Fund, which covers all civil servants including employed, retired and pensioned individuals; the Rural Fund, which includes villagers, nomads and residents of cities with populations under 20 000; the Iranian and Universal Health Insurance Funds, available to all Iranians either through full premium payment or based on household income; the Foreign Citizens Fund, for non-Iranian nationals; and the Other Social Strata Fund, which covers veterans, warfare victims, students, disabled individuals, welfare recipients and prisoners and their families.¹³

Data collection and preparation

The extracted data incorporated individuals’ date of birth, sex, province of residence, insurance fund, prescription date and medication records. Each insured individual was identified by a unique anonymised code, facilitating identification of multiple prescriptions for the same individual. As part of the data preparation process, we assessed the completeness of demographic variables and medication records in the database, which led to the exclusion of 3341 individuals with missing birth dates. To ensure the accuracy of age classification, we calculated individuals’ ages based on their date of birth relative to the prescription date. Only prescriptions where the recipient was aged 65 or older at the time of each prescription were included for analysis.

Medication records included details such as dosage, route of administration, form and package number. Medications were classified according to the Anatomical Therapeutic Chemical (ATC) classification system. This internationally recognised system categorises active substances into groups based on their therapeutic, pharmacological and chemical properties across five hierarchical levels. The first level (ATC-1) consists of main anatomical or pharmacological groups: alimentary tract and metabolism (A), blood and blood-forming organs (B), cardiovascular system (C), dermatologicals (D), genitourinary system and sex hormones (G), systemic hormonal preparations excluding sex hormones and insulins (H), anti-infectives for systemic use (J), antineoplastic and immunomodulating agents (L), musculo-skeletal system (M), nervous system (N), antiparasitic products, insecticides and repellents (P), respiratory system (R), sensory organs (S) and various (V). Each of these groups is divided into second-level categories, which are further detailed into third and fourth levels representing more specific pharmacological or chemical subgroups. The fifth level (ATC-5) identifies the precise chemical substance.¹⁵

Our dataset initially included a broad spectrum of medications. However, after careful consideration and in consultation with an expert panel in gerontology, internal medicine and pharmacology, we excluded medications falling under specific ATC codes. These exclusions encompassed stomatological preparations (A01), blood substitutes and perfusion solutions (B05), topical anti-haemorrhoid (C05A) and anti-varicose drugs (C05B), hormonal contraceptives (G03A), immunoglobulins (J06), vaccines (J07), topical products for joint and muscular pain (M02), anaesthetics (N01), ectoparasiticides (P03), throat preparations (R02), respiratory stimulants and surfactants (R07), medications classified under ‘various’ (V) ATC group, topical dermatological agents, and local ophthalmological and otological agents (except for anti-infective, anti-inflammatory, anti-glaucoma/miotic and mydriatic/cycloplegic agents). These exclusions were based on their predominant use in hospital settings, diagnostic applications, infrequent administration, lack of relevance to the elderly population, minimal interaction risks with other medications and negligible clinical significance. Ultimately, our dataset comprised 2298 drugs categorised under 789 unique ATC-5 codes after applying these exclusion criteria.

Definitions of polypharmacy

Polypharmacy, broadly defined by the WHO as ‘the administration of an excessive number of drugs’, allows for various interpretations among researchers. Identifying polypharmacy depends on the number of medications taken within a specified time frame. While there is no universal consensus, the threshold of five medications is most frequently used in the literature. Polypharmacy can be further categorised based on the time frame into spontaneous, cumulative or continuous, with chronic polypharmacy defined by the frequency or duration of medication use.¹⁶

Based on the available data, measuring simultaneous polypharmacy was not feasible as it requires information on the duration of each drug’s administration. Additionally, using defined daily doses (DDD) was limited due to the lack of DDD information for many medications in the dataset and the discrepancy between DDD and national prescribing practices, which reduces its accuracy and generalisability. Therefore, we employed a cumulative definition of polypharmacy, considering the total number of different medications administered over a given period. For each individual, we calculated the number of different medications (based on ATC-5) received in 1-month and 6-month intervals. For individuals with multiple prescriptions, polypharmacy was assessed across all intervals generated from each prescription date. Individuals who received five or more different medications in at least one of these intervals during the 3-year study were identified as experiencing cumulative polypharmacy. To evaluate the severity of polypharmacy, we introduced a new definition: consecutive polypharmacy. This was defined as receiving five or more different medications every month or at least once every 2 months within a 6-month period. This definition focuses on prescription frequency rather than chronicity, as the drugs received consecutively could vary from month to month.

Data analysis

Quantitative variables were summarised using mean and ±SD, while categorical variables were presented as frequencies and percentages with 95% CIs. Prevalence rates of polypharmacy were calculated for various demographic and provincial groups, as well as for different drug classes based on ATC-1, stratified by polypharmacy status, age and sex. The associations between polypharmacy and various demographic factors, as well as consumption of different drug classes, were analysed using univariate and multiple logistic regression, with results reported as crude OR and adjusted ORs (AOR), all with 95% CI. The multiple logistic regression models were adjusted for age, sex, insurance fund and province.

Association rule mining was employed to identify commonly used drugs (based on ATC-4), drug combinations and their relation to polypharmacy. This data mining technique searches for frequent if-then patterns with significance assessed using Support, Confidence and Lift metrics.

Support measures the frequency of an item in the dataset:

$$Support\left(A\right)=\frac{NumberofrecordscontainingitemA}{Totalnumberofrecords}$$

Confidence evaluates the reliability of the association rule, estimating the conditional probability that the consequent is present given the antecedent:

$$Confidence\left(A\to B\right)=P\left(B|A\right)=\frac{Support(A\cup B)}{Support\left(A\right)}$$

Lift assesses the strength of the association relative to independent occurrences:

$$Lift\left(A\to B\right)=\frac{Support\left(A\cup B\right)}{Support\left(A\right)\times Support\left(B\right)}$$

We used different levels of the ATC classification system in this study, each chosen to fit the specific needs of our analyses. The ATC-5 level was used to define polypharmacy, as it provides the most detailed subclassification of medications, critical for identifying the use of multiple distinct drugs accurately. For logistic regression analysis, we applied the ATC-1 level, which organises medications into broad therapeutic groups, allowing us to focus on high-level patterns that influence polypharmacy without getting lost in detailed classifications. For association rule mining, we selected the ATC-4 level, which strikes a balance by grouping medications into clinically relevant therapeutic categories—detailed enough to uncover meaningful associations while remaining aggregated enough for the analysis to be manageable and interpretable.

The analysis was conducted using Python programming language V.3.11.4, with Pandas and NumPy libraries for data preparation, Matplotlib for visualisation, SciPy and statsmodels for statistical analysis, and mlxtend for association rule mining.

Patient and public involvement

Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this research. This study used anonymised secondary data from national health insurance claims, precluding direct engagement with patients or members of the public.

Results

Prevalence of polypharmacy

The study assessed polypharmacy prevalence among 1 876 527 elderly individuals who received at least one prescription. Polypharmacy was evaluated using different time windows: cumulative polypharmacy over 6-month and 1-month periods, and consecutive polypharmacy over a 6-month period. The results, along with baseline characteristics of the study population, are presented in table 1.

Table 1. Prevalence of polypharmacy and baseline characteristics.

Characteristics	Total population n=1 876 527	Cumulative polypharmacy (6-month window) n=1 405 654	Cumulative polypharmacy (1-month window) n=1 211 378	Consecutive polypharmacy (6-month window) n=470 873
	% (95% CI)	P% (95% CI)	P% (95% CI)	P% (95% CI)
Overall	100	74.907 (74.845 to 74.969)	64.554 (64.486 to 64.623)	7.615 (7.577 to 7.653)
Sex
Female	52.897 (52.825 to 52.968)	77.825 (77.744 to 77.907)	68.359 (68.267 to 68.45)	9.060 (9.003 to 9.116)
Male	47.103 (47.032 to 47.175)	71.630 (71.536 to 71.724)	60.282 (60.18 to 60.384)	5.993 (5.943 to 6.042)
Age group
65–69	37.600 (37.531 to 37.669)	74.713 (74.611 to 74.814)	63.550 (63.438 to 63.662)	6.923 (6.863 to 6.982)
70–74	20.553 (20.495 to 20.611)	76.583 (76.45 to 76.717)	66.415 (66.266 to 66.564)	8.546 (8.458 to 8.634)
75–79	17.414 (17.36 to 17.468)	77.010 (76.865 to 77.154)	67.403 (67.243 to 67.564)	8.921 (8.824 to 9.019)
80–84	13.626 (13.576 to 13.675)	75.129 (74.961 to 75.296)	65.546 (65.362 to 65.73)	8.048 (7.943 to 8.154)
≥85	10.808 (10.763 to 10.852)	68.729 (68.528 to 68.931)	58.667 (58.453 to 58.882)	5.602 (5.502 to 5.702)
Fund
Civil Servants	42.017 (41.946 to 42.087)	82.952 (82.869 to 83.035)	72.569 (72.47 to 72.667)	9.439 (9.374 to 9.504)
Rural	31.344 (31.277 to 31.41)	59.055 (58.929 to 59.18)	48.305 (48.178 to 48.433)	1.865 (1.831 to 1.9)
Iranian and Universal Health Insurance	16.747 (16.694 to 16.801)	77.356 (77.21 to 77.503)	65.598 (65.432 to 65.765)	5.466 (5.387 to 5.546)
Other Social Strata	9.739 (9.697 to 9.781)	87.317 (87.165 to 87.47)	80.789 (80.608 to 80.97)	22.066 (21.875 to 22.256)
Foreign Citizens	0.154 (0.148to 0.159)	55.312 (53.497 to 57.128)	44.757 (42.941 to 46.573)	0

P, prevalence.

Overall, 74.907% of the population experienced cumulative polypharmacy in a 6-month window, 64.554% in a 1-month window, and 7.615% experienced consecutive polypharmacy in a 6-month window. Females were more likely to experience polypharmacy across all measures compared with males. Age stratification revealed that polypharmacy prevalence increased with age, peaking in the 75–79 age group, while individuals aged 85 and older had the lowest rates of polypharmacy compared with all other age groups. Significant variations were also observed across different insurance funds, with the highest prevalence found in the ‘Other Social Strata’ group, while the ‘Foreign Citizens’ and ‘Rural’ groups had the lowest prevalences.

Drug usage patterns

Table 2 summarises drug usage and polypharmacy metrics by polypharmacy status. On average, the total population used 16.10 medications, compared with 3.29 medications for those not exposed to polypharmacy. In polypharmacy groups, usage increased significantly, with consecutive polypharmacy showing the highest average of 41.08 medications. Polypharmacy groups also had more prescriptions, with a mean of 13.44 for cumulative polypharmacy over 6 months and 36.29 for consecutive polypharmacy. The number of prescriptions with five or more drugs was notably higher in polypharmacy groups, especially in consecutive polypharmacy, which averaged 11.79. The mean number of drugs per prescription was similar across consecutive and cumulative polypharmacy groups.

Table 2. Summary statistics of drug usage and polypharmacy metrics according to polypharmacy status.

Variables	Total population	Not exposed to polypharmacy	Cumulative polypharmacy (6-month window)	Cumulative polypharmacy (1-month window)	Consecutive polypharmacy (6-month window)
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Number of drug classes (ATC-1)	5.65 (2.79)	2.34 (1.22)	6.76 (2.23)	7.04 (2.19)	8.95 (1.59)
Number of drugs (ATC-5)	16.10 (13.70)	3.29 (1.82)	20.35 (13.33)	22.08 (13.49)	41.08 (16.17)
Number of prescriptions	10.50 (12.20)	1.78 (1.38)	13.44 (12.80)	14.69 (13.30)	36.29 (18.33)
Number of prescriptions with ≥5 drugs	2.28 (4.00)	0	3.04 (4.36)	3.53 (4.21)	11.79 (6.90)
Number of drugs per prescription	3.20 (1.79)	2.11 (0.99)	3.25 (1.80)	3.31 (1.82)	3.71 (2.00)
Number of polypharmacy months	---	---	---	4.49 (4.68)	13.75 (5.60)
Polypharmacy duration by month (without gap)	---	---	---	1.24 (0.68)	1.97 (1.59)
Polypharmacy duration by month (1-month gap allowed)	---	---	---	1.90 (2.40)	5.82 (5.32)

ATC, Anatomical Therapeutic Chemical classification.

Figure 1 shows the distribution of drug consumption across main anatomical groups (ATC-1) by sex, age group and polypharmacy status. The most commonly used medications belonged to the alimentary tract and metabolism group (A). Females generally had higher consumption rates across most drug classes compared with males. However, males had higher consumption rates than females for anti-infectives (J), genitourinary drugs (G) and antineoplastic and immunomodulating agents (L). Consumption patterns remained fairly consistent across age groups, increasing with age until declining after the 80–84 age group. However, consumption of antiparasitic products (P) consistently decreased with age. Individuals with consecutive polypharmacy exhibited higher consumption rates across all anatomical groups, while those without polypharmacy had lower consumption rates.

Figure 1. Distribution of each drug class consumption by sex (a), age group (b), and polypharmacy status (c). A, alimentary tract and metabolism; ATC, Anatomical Therapeutic Chemical; B, blood and blood-forming organs; C, cardiovascular system; D, dermatologicals; G, genitourinary system and sex hormones; H, systemic hormonal preparations excluding sex hormones and insulins; J, anti-infectives for systemic use; L, antineoplastic and immunomodulating agents; M, musculoskeletal system; n, nervous system; p, antiparasitic products, insecticides and repellents; R, respiratory system; S, sensory organs.

Predictors of polypharmacy

Univariate and multivariate logistic regression analyses are summarised in table 3. Females had significantly higher odds of experiencing polypharmacy compared with males, with 1.484 times higher odds (95% CI: 1.467 to 1.501) for consecutive polypharmacy. Among age groups, individuals aged 75–79 had the highest odds of consecutive polypharmacy compared with the 65–69 group (AOR=1.317, 95% CI: 1.297 to 1.338). However, the trend reversed in individuals aged 85 and older, who had significantly lower odds of polypharmacy (AOR=0.744, 95% CI: 0.728 to 0.761).

Table 3. Univariate and multivariate logistic regression analysis of predictors of polypharmacy.

Drug class	Cumulative polypharmacy (6-month window)		Consecutive polypharmacy (6-month window)
	Crude OR (95% CI)	Adjusted OR (95% CI)	Crude OR (95% CI)	Adjusted OR (95% CI)
Sex
Male (reference)	1	1	1	1
Female	1.390 (1.381 to 1.399)	1.391 (1.382 to 1.401)	1.563 (1.545 to 1.580)	1.484 (1.467 to 1.501)
Age group
65–69 (reference)	1	1	1	1
70–74	1.107 (1.097 to 1.117)	1.185 (1.174 to 1.197)	1.256 (1.238 to 1.275)	1.290 (1.270 to 1.309)
75–79	1.134 (1.123 to 1.145)	1.232 (1.22 to 1.245)	1.317 (1.297 to 1.337)	1.317 (1.297 to 1.338)
80–84	1.022 (1.012 to 1.033)	1.118 (1.106 to 1.130)	1.177 (1.157 to 1.197)	1.158 (1.138 to 1.179)
≥85	0.744 (0.736 to 0.752)	0.749 (0.741 to 0.758)	0.798 (0.781 to 0.815)	0.744 (0.728 to 0.761)
Fund
Civil Servants (reference)	1	1	1	1
Rural	0.296 (0.294 to 0.299)	0.244 (0.242 to 0.246)	0.183 (0.179 to 0.187)	0.156 (0.153 to 0.160)
Iranian and Universal Health Insurance	0.702 (0.695 to 0.709)	0.731 (0.724 to 0.739)	0.557 (0.547 to 0.566)	0.601 (0.590 to 0.611)
Other Social Strata	1.415 (1.394 to 1.436)	1.277 (1.258 to 1.297)	2.727 (2.691 to 2.764)	2.448 (2.415 to 2.482)
Foreign Citizens	0.254 (0.236 to 0.274)	0.231 (0.214 to 0.249)	---	---
Main anatomical groups (ATC-1)
A (non-users as reference)	11.308 (11.221 to 11.396)	9.320 (9.246 to 9.395)	46.471 (43.778 to 49.329)	29.711 (27.985 to 31.544)
B (non-users as reference)	10.475 (10.382 to 10.569)	9.590 (9.502 to 9.679)	11.745 (11.528 to 11.965)	9.272 (9.098 to 9.448)
C (non-users as reference)	12.322 (12.227 to 12.418)	10.502 (10.419 to 10.587)	32.427 (31.118 to 33.792)	22.597 (21.681 to 23.553)
D (non-users as reference)	2.825 (2.576 to 3.097)	2.192 (1.995 to 2.409)	2.620 (2.432 to 2.823)	2.054 (1.9 to 2.221)
G (non-users as reference)	4.841 (4.778 to 4.905)	4.191 (4.135 to 4.248)	3.646 (3.605 to 3.687)	2.901 (2.867 to 2.936)
H (non-users as reference)	7.257 (7.2 to 7.314)	6.111 (6.061 to 6.161)	4.972 (4.9 to 5.044)	3.514 (3.462 to 3.566)
J (non-users as reference)	6.964 (6.914 to 7.015)	5.787 (5.743 to 5.831)	8.422 (8.239 to 8.608)	5.705 (5.579 to 5.834)
L (non-users as reference)	4.652 (4.523 to 4.785)	4.603 (4.473 to 4.737)	3.005 (2.948 to 3.063)	3.022 (2.961 to 3.084)
M (non-users as reference)	6.779 (6.727 to 6.831)	5.468 (5.424 to 5.512)	6.625 (6.519 to 6.732)	4.460 (4.387 to 4.534)
N (non-users as reference)	9.903 (9.829 to 9.978)	8.033 (7.971 to 8.096)	17.764 (17.215 to 18.33)	11.463 (11.106 to 11.832)
P (non-users as reference)	5.963 (5.88 to 6.048)	4.772 (4.705 to 4.841)	3.777 (3.735 to 3.819)	2.846 (2.813 to 2.88)
R (non-users as reference)	9.175 (9.098 to 9.252)	7.406 (7.343 to 7.470)	7.829 (7.7 to 7.961)	5.214 (5.125 to 5.304)
S (non-users as reference)	2.564 (2.544 to 2.585)	2.235 (2.216 to 2.253)	3.258 (3.222 to 3.295)	2.572 (2.542 to 2.601)

Adjustment variables: age group, sex, fund and province.

A, alimentary tract and metabolism; ATC, Anatomical Therapeutic Chemical classification; B, blood and blood-forming organs; C, cardiovascular system; D, dermatologicals; G, genitourinary system and sex hormones; H, systemic hormonal preparations excluding sex hormones and insulins; J, anti-infectives for systemic use; L, antineoplastic and immunomodulating agents; M, musculoskeletal system; N, nervous system; P, antiparasitic products, insecticides and repellents; R, respiratory system; S, sensory organs.

Insurance status also influenced polypharmacy risk. Compared with the Civil Servants fund, those in the Rural Fund and Iranian and Universal Health Insurance had lower odds of polypharmacy, while individuals in the Other Social Strata fund had a much higher risk. Foreign citizens had the lowest odds of cumulative polypharmacy (AOR=0.231, 95% CI: 0.214 to 0.249), but no cases of consecutive polypharmacy were observed in this group.

Logistic regression analysis revealed that having a prescription history of medications from certain main anatomical groups (ATC-1) was strongly associated with an increased likelihood of polypharmacy. Specifically, those with a prescription history from the alimentary tract and metabolism group (class A) were 29.711 times more likely (95% CI: 27.985 to 31.544) to experience consecutive polypharmacy compared with those who had never been prescribed medications from this class. Following class A, class C (cardiovascular system), class N (nervous system) and class B (blood and blood-forming organs) also showed strong associations, with AORs of 22.597, 11.463 and 9.272 for consecutive polypharmacy, respectively. For all drug classes, the AORs were lower than the crude ORs. In general, AORs for consecutive polypharmacy were similar to or lower than those for cumulative polypharmacy across most drug classes. However, exceptions were seen in classes A, C and N, where the ORs for consecutive polypharmacy were higher than those for cumulative polypharmacy.

Regional variations

The analysis revealed significant regional variation in polypharmacy across provinces (figure 2). Prevalence rates ranged from 60.69% (Hormozgan) to 80.85% (Isfahan) for cumulative polypharmacy and 2.76% (Hormozgan) to 13.60% (Ilam) for consecutive polypharmacy. Multivariate logistic regression analysis showed that Fars had the highest odds for cumulative polypharmacy (AOR=2.36), while Chaharmahal and Bakhtiari had the highest odds for consecutive polypharmacy (AOR=4.21), both in comparison to Tehran.

Figure 2. Polypharmacy prevalence and ORs by province.

Commonly used medications and their association with polypharmacy

Table 4 presents the most commonly used medications by pharmacological subgroup (ATC-4) overall and by polypharmacy status. Systemic glucocorticoids were the most commonly used overall (50.02%), followed by HMG-CoA reductase inhibitors (42.73%) and platelet aggregation inhibitors excluding heparin (41.92%). HMG-CoA reductase inhibitors were the most prevalent in consecutive polypharmacy (83.54%), followed by platelet aggregation inhibitors (82.84%), while glucocorticoids were the most common in non-polypharmacy (18.26%) and cumulative polypharmacy (60.66%) groups. As shown in table 5, platelet aggregation inhibitors had the strongest association with organic nitrates (lift=2.01) and were coprescribed in 20.78% of the population. Additionally, platelet aggregation inhibitors and HMG-CoA reductase inhibitors were strongly associated with each other, as well as with beta-blockers and angiotensin-II receptor blockers (ARBs). Glucocorticoids were also associated with acetic acid derivatives, third-generation cephalosporins and anilides.

Table 4. The most commonly used medications in total and in different polypharmacy status.

Total		Non-polypharmacy		Cumulative polypharmacy		Consecutive polypharmacy
Pharmacological subgroup (ATC-4)	Proportion (%)	Pharmacological subgroup (ATC-4)	Proportion (%)	Pharmacological subgroup (ATC-4)	Proportion (%)	Pharmacological subgroup (ATC-4)	Proportion (%)
Glucocorticoids (H02AB)	50.02	Glucocorticoids (H02AB)	18.26	Glucocorticoids (H02AB)	60.66	HMG-CoA reductase inhibitors (C10AA)	83.54
HMG-CoA reductase inhibitors (C10AA)	42.73	Acetic acid derivatives and related substances (M01AB)	11.06	HMG-CoA reductase Inhibitors (C10AA)	53.94	Platelet aggregation inhibitors excl. heparin (B01AC)	82.84
Platelet aggregation inhibitors excl. heparin (B01AC)	41.92	Vitamin D and analogues (A11CC)	9.97	Platelet aggregation inhibitors excl. heparin (B01AC)	52.90	Glucocorticoids (H02AB)	79.62
Acetic acid derivatives and related substances (M01AB)	36.83	Fluoroquinolones (J01MA)	9.75	Angiotensin-II receptor blockers (ARBs), plain (C09CA)	45.97	Angiotensin-II receptor blockers (ARBs), plain (C09CA)	76.22
Angiotensin-II receptor blockers (ARBs), plain (C09CA)	36.31	Antibiotics, ophthalmological (S01AA)	9.29	Acetic acid derivatives and related substances (M01AB)	45.46	Anilides (N02BE)	74.68
Vitamin D And analogues (A11CC)	35.80	HMG-CoA reductase inhibitors (C10AA)	9.29	Anilides (N02BE)	44.98	H2-receptor antagonists (A02BA)	72.53
Anilides (N02BE)	35.67	Platelet aggregation inhibitors excl. heparin (B01AC)	9.14	Vitamin D and analogues (A11CC)	44.45	Acetic acid derivatives and related substances (M01AB)	71.50
H2-receptor antagonists (A02BA)	32.78	Corticosteroids, ophthalmological (S03BA)	8.79	H2-Receptor antagonists (A02BA)	41.43	Vitamin D And analogues (A11CC)	70.49
Third-generation cephalosporins (J01DD)	32.10	Third-generation cephalosporins (J01DD)	7.96	Third-generation cephalosporins (J01DD)	40.18	Third-generation cephalosporins (J01DD)	67.61
Beta blocking agents, selective (C07AB)	30.17	Anilides (N02BE)	7.89	Beta blocking agents, selective (C07AB)	38.38	Beta-blocking agents, selective (C07AB)	65.31
Fluoroquinolones (J01MA)	27.48	Angiotensin-II receptor blockers (ARBs), plain (C09CA)	7.46	Fluoroquinolones (J01MA)	33.42	Proton pump inhibitors (A02BC)	62.57
Proton pump inhibitors (A02BC)	26.24	H2-receptor antagonists (A02BA)	6.95	Proton pump inhibitors (A02BC)	33.30	Benzodiazepine derivatives (N05BA)	61.34
Benzodiazepine derivatives (N05BA)	26.05	Corticosteroids, plain, ophthalmological (S01BA)	6.38	Benzodiazepine derivatives (N05BA)	33.07	Organic nitrates (C01DA)	61.19
Macrolides (J01FA)	25.15	First-generation cephalosporins (J01DB)	5.80	Macrolides (J01FA)	31.90	Macrolides (J01FA)	55.84
Organic nitrates (C01DA)	24.70	Beta-blocking agents, selective (C07AB)	5.66	Organic nitrates (C01DA)	31.85	Fluoroquinolones (J01MA)	55.43

ATC, Anatomical Therapeutic Chemical classification.

Table 5. The most common medications taken together.

Combination (ATC-4)	Support (%)	Lift
Platelet aggregation inhibitors excl. heparin (B01AC) + Organic nitrates (C01DA)	20.78	2.01
Platelet aggregation inhibitors excl. heparin (B01AC) + Beta-blocking agents, selective (C07AB)	22.21	1.76
Platelet aggregation inhibitors excl. heparin (B01AC) + HMG-CoA reductase inhibitors (C10AA)	31.39	1.75
HMG-CoA reductase inhibitors (C10AA) + Beta-blocking agents, selective (C07AB)	21.68	1.68
Platelet aggregation inhibitors excl. heparin (B01AC) + Angiotensin-II receptor blockers (ARBs), plain (C09CA)	25.09	1.65
HMG-CoA reductase inhibitors (C10AA) + Angiotensin-II receptor blockers (ARBs), plain (C09CA)	25.02	1.61
Vitamin D and analogues (A11CC) + Acetic acid derivatives and related substances (M01AB)	20.77	1.58
Acetic acid derivatives and related substances (M01AB) + Glucocorticoids (H02AB)	27.33	1.48
Third-generation cephalosporins (J01DD) + Glucocorticoids (H02AB)	22.77	1.42
Anilides (N02BE) + Glucocorticoids (H02AB)	25.13	1.41

ATC, Anatomical Therapeutic Chemical classification.

The simultaneous presence of glucocorticoids and HMG-CoA reductase inhibitors in an individual’s prescription history showed the most prominent relationship with consecutive polypharmacy (confidence=20.6%, lift=2.71). Additionally, simultaneous consumption of ARBs with HMG-CoA reductase inhibitors (confidence=20.4%, lift=2.68) and platelet aggregation inhibitors (confidence=20.2%, lift=2.65) significantly increased the possibility of consecutive polypharmacy. Other medications, such as H2-receptor antagonists, third-generation cephalosporins, anilides, vitamin D and acetic acid derivatives, showed notable but less frequent associations with consecutive polypharmacy (table 6).

Table 6. Association rules of medications or their combinations and consecutive polypharmacy.

Antecedent (ATC-4)	Support (%)	Confidence (%)	Lift
Glucocorticoids (H02AB) + HMG-CoA reductase inhibitors (C10AA)	5.02	20.60	2.71
Angiotensin-II receptor blockers (ARBs), plain (C09CA) + HMG-CoA reductase inhibitors	5.10	20.40	2.68
Angiotensin-II receptor blockers (ARBs), plain (C09CA) + Platelet aggregation inhibitors excl. heparin (B01AC)	5.07	20.20	2.65
HMG-CoA reductase inhibitors (C10AA) + Platelet aggregation inhibitors excl. heparin (B01AC)	5.61	17.88	2.35
H2-receptor antagonists (A02BA)	5.52	16.85	2.21
Third-generation cephalosporins (J01DD)	5.15	16.04	2.11
Angiotensin-II receptor blockers (ARBs), plain (C09CA)	5.80	15.99	2.10
Anilides (N02BE)	5.69	15.94	2.09
Platelet aggregation inhibitors excl. heparin (B01AC)	6.31	15.05	1.98
Vitamin D and analogues (A11CC)	5.37	15.00	1.97
HMG-CoA reductase inhibitors (C10AA)	6.36	14.89	1.95
Acetic acid derivatives and related substances (M01AB)	5.44	14.78	1.94
Glucocorticoids (H02AB)	6.06	12.12	1.59

ATC, Anatomical Therapeutic Chemical classification.

Discussion

In this study, we investigated the prevalence and characteristics of polypharmacy among the elderly population in Iran over a 3-year period using IHIO claims data. Our findings revealed a high prevalence of polypharmacy among older adults, with approximately 75% of the study population being prescribed more than five different medications within a 6-month period. Additionally, one-third of these individuals (7.6% of the total population) experienced sustained exposure to polypharmacy, a measure we defined as consecutive polypharmacy. The analysis revealed significant demographic patterns, with females more likely to experience polypharmacy across all measures, and the highest prevalence observed in the 75–79 age group. There were also substantial variations in polypharmacy prevalence across different insurance funds and provinces, indicating the influence of socioeconomic and regional factors.

When comparing our findings to similar studies conducted in other countries, the prevalence of polypharmacy among the elderly in Iran is notably high but consistent with global patterns. In the same region, a retrospective analysis in Qatar in 2017 found a polypharmacy prevalence of 75.5% among older adults using Electronic Medical Records from primary healthcare centres over a 6-month period, closely matching our results.¹⁷ Similarly, a 2016 study in Saudi Arabia found that 66.3% of those aged 60 and above were on five or more medications based on a 6-month study of outpatient electronic health records.¹⁸ In East Asia, the highest prevalence was observed in Korea, where 86.4% of the elderly were affected, according to data from the Korea Health Insurance Review and Assessment Service from 2010 and 2011.¹⁹ In Tokyo, Japan, 63.5% of outpatients aged 75 and older were prescribed five or more drug types during a 6-month period in 2014.²⁰ In contrast, a study in Switzerland using 2016 health insurance claims data found that 50.4% of the elderly experienced polypharmacy during a 3-month period.²¹ These comparisons highlight that the high prevalence of polypharmacy in Iran is part of a broader global trend, with variations likely due to differences in healthcare systems, prescribing practices and population health profiles.

Consistent with other studies, polypharmacy prevalence was significantly higher among females in our study,^{17 18 21 22} though some studies have reported higher rates in males.^{19 20} These inconsistencies may stem from differences in physicians’ prescribing attitude towards males and females,^{23 24} as well as variations in health-seeking behaviours between sexes.²⁵ Polypharmacy is also strongly linked to non-communicable diseases and the number of chronic conditions.^{10 17} An Iranian study using latent class analysis found females more likely to fall into cardiovascular-metabolic and cognitive-metabolic multimorbidity classes, both associated with polypharmacy.²⁶ Thus, the higher polypharmacy prevalence in females may be partly due to their greater burden of certain multimorbidities.

Our study found that individuals aged 75–79 had the highest odds of polypharmacy compared with the 65–69 age group, while those aged 85 and older showed significantly lower odds. This pattern aligns with other research showing an initial increase in polypharmacy with age, that can be due to the accumulation of chronic conditions and multimorbidity, followed by a decrease in the very elderly.^{19 21} An Italian study on prescription patterns found that medication use significantly declined after age 90, with reductions most notable for chronic disease treatments (eg, antihypertensives) and less so for acute condition medications (eg, antibiotics and glucocorticoids).²⁷ Our findings similarly indicated a decrease in overall medication use among the oldest age group, with a more pronounced reduction in classes like cardiovascular (class C) and nervous system (class N) drugs compared with anti-infectives (class J). These findings may reflect a more cautious medical approach in the oldest age group due to changes in the risk–benefit profile of treatments. Chronic disease management may also become less prioritised given limited life expectancy, with medications for acute conditions or symptoms taking precedence over preventive therapies. Additionally, a healthy survivor effect may explain why only the fittest individuals, requiring fewer medications, live past 85. Lastly, reduced access to care and potential age-related discrimination may also contribute to lower polypharmacy in the very old.²⁸

When examining the impact of health insurance status on polypharmacy, our study found that individuals in the Rural and Foreign Citizens Funds had significantly lower odds of polypharmacy compared with those in the Civil Servants Fund. This difference may stem from reduced access to healthcare services, fewer medical consultations and more conservative healthcare-seeking behaviours in rural areas.²⁹ For foreign citizens, factors such as cultural preferences, legal barriers and cost concerns are likely to contribute to their lower polypharmacy rates. More targeted studies are needed to explore the underlying causes of these disparities and to develop strategies to address potential gaps in healthcare delivery.

The most commonly prescribed drug classes associated with consecutive polypharmacy in our study included those related to the alimentary tract and metabolism, cardiovascular system, nervous system and blood and blood-forming organs. Specifically, vitamin D and analogues, H2-receptor antagonists and proton pump inhibitors were frequently prescribed for alimentary tract and metabolism issues, while HMG-CoA reductase inhibitors, ARBs, beta-blockers and organic nitrates dominated cardiovascular prescriptions. In the nervous system category, anilides and benzodiazepines were prominent, and platelet aggregation inhibitors were common among drugs acting on blood and blood-forming organs. This pattern of medication use aligns with the chronic conditions prevalent in the elderly population.³⁰ Supporting our findings, a study in Tehran found that cardiovascular drugs accounted for 20.8% of all prescriptions among elderly patients, with a significant portion also receiving drugs for diabetes and nutritional supplements.³¹ Additionally, the Pars Cohort Study in Iran highlighted that cardiovascular drugs, drugs acting on blood and blood-forming organs, and alimentary tract and metabolism drugs were the most frequently used among patients with cardiovascular diseases.³² Consistently, association rule mining analysis in our study showed that HMG-CoA reductase inhibitors, ARBs and platelet aggregation inhibitors were most commonly used together in patients with polypharmacy.

Systemic glucocorticoids were the most prevalent pharmacological subgroup (ATC-4) in our study, with half of the elderly population having used them during the study period. This high prevalence is consistent with other studies that highlight the widespread use of glucocorticoids, particularly dexamethasone, among general physicians.¹¹ The popularity of glucocorticoids can be attributed to their rapid alleviation of disease symptoms, making them a favoured choice for both doctors and patients despite their well-documented side effects, including hypertension, weight gain, increased blood sugar, osteoporosis, mood swings, cardiovascular diseases, increased risk of infections and delay in diagnosis and treatment of diseases.³³ However, the frequent use of glucocorticoids is concerning, as overprescription exacerbates health risks for patients and increases healthcare costs by prolonging treatment and leading to additional complications. The high rate of glucocorticoid prescriptions may reflect a combination of factors, including gaps in physician knowledge, strong patient demand, limited access to alternative treatments and weak regulatory oversight.³⁴ Addressing these issues by enhancing physician education and improving regulatory frameworks is crucial to curbing the overuse of glucocorticoids and reducing their associated risks.

While prescribing multiple medications can be clinically justified and isn't inherently inappropriate, it does increase the risk of adverse drug events and healthcare costs. Polypharmacy also raises the concern of overtreatment, where medications offer little clinical benefit for the patient’s remaining lifespan or where the risks outweigh the potential benefits of additional treatments.³⁵ Our study sheds light on medication use patterns and polypharmacy, providing valuable insights for healthcare providers, emphasising the need for targeted interventions to manage and optimise medication use among the elderly. Implementing regular medication reviews, deprescribing strategies and, when appropriate, represcribing—the addition of beneficial medications that are often omitted—could help minimise the risks associated with both overtreatment and undertreatment.36,39 In this context, structured tools such as the Screening Tool to Alert to Right Treatment/Screening Tool of Older Persons’ Prescriptions (START/STOPP) criteria and the Fit fOR The Aged (FORTA) list, which have demonstrated clinical benefits in randomised trials, offer promising patient-centred approaches. Unlike purely drug-oriented listing tools like the Beers criteria, these instruments consider both potentially inappropriate medications and potentially omitted medications, supporting a more holistic medication optimisation strategy known as represcribing.³⁹ Further research should focus on exploring the long-term impacts of polypharmacy on patient outcomes and evaluating the effectiveness of interventions designed to reduce inappropriate prescribing.

This study has several strengths, including the use of a large, nationwide dataset that allows for analysis with sufficient statistical power and provides a comprehensive overview of polypharmacy among the elderly in Iran. Furthermore, we employed association rule mining to investigate patterns of polypharmacy and identify drugs commonly used together, which is a novel method in this context and adds depth to our analysis. The focus on consecutive polypharmacy offers valuable insights into the persistence of polypharmacy, a critical factor in understanding its potential risks.

However, there are several limitations to consider. The use of claims data, while extensive, lacks detailed clinical information such as the rationale for prescriptions and patient adherence, which limits our ability to fully assess the appropriateness of polypharmacy. Additionally, our analysis did not account for non-prescription medications, such as over-the-counter drugs and herbal supplements, which may lead to an underestimation of polypharmacy prevalence. Another significant limitation is the lack of a standardised definition for polypharmacy across different studies. This variation in definitions makes it challenging to directly compare our findings with those from other research, as the criteria for what constitutes polypharmacy can differ widely. Finally, while our study is representative of the elderly population in Iran, the findings may not be fully generalisable to the entire population of interest, as our data were limited to elderly individuals covered by the IHIO who received at least one medication during the study period.

Conclusions

This study provides a comprehensive analysis of polypharmacy prevalence and patterns among the elderly population in Iran using health insurance claims data. The findings highlight the widespread occurrence of polypharmacy, with nearly three-quarters of the elderly experiencing cumulative polypharmacy and a smaller yet significant portion affected by consecutive polypharmacy. Polypharmacy was more common among females and those aged 75–79, with significant variations across insurance funds and provinces. The study underscores the significant role of specific drug classes, such as those related to the alimentary tract, cardiovascular system and nervous system, in driving polypharmacy. Overall, this study contributes to a better understanding of polypharmacy in the Iranian elderly population, providing a foundation for future research and policy initiatives aimed at improving medication safety and health outcomes in this vulnerable group.

Data availability statement

Data are available on reasonable request.