Abstract
Background
The prevalence of multimorbidity is increasing and is associated with polypharmacy (PP) and a reduced quality of life (QoL). Polypharmacy consistently correlates with poorer outcomes; however, its relationship with QoL in a broad general practice population is underexplored. For the general practitioner, QoL assessment is resource-intensive; therefore, this study examines whether the number of redeemed unique prescription medications can serve as an indicator for QoL in patients with multimorbidity managed in general practice.
Methods
This nationwide cross-sectional study was conducted with data from questionnaires sent to 160,584 adults from 250 general practices who consulted their general practitioner for an annual chronic disease consultation in 2022. Questionnaire data were linked with socioeconomic and medication data from Danish Registries. We examined the association between the number of redeemed unique prescription medicines and six domains of QoL with linear models and tested for effect modification with multivariable linear models. The multivariable models were adjusted for the covariates sex, age, education, and cohabitation. The number of unique prescription medications that constitute a minimal clinically important difference (MCID) was calculated. Further, we presented associations between redeemed unique prescription medications and QoL stratified by the covariates.
Results
All 35,977 patients who participated in the survey were included in this study, with 18,665 (51.9%) being female. A linear association was found between redeemed unique prescription medications and more burdened QoL in all six domains. The largest association (2.95; 95% CI 2.87–3.03) and lowest MCID (4.4 redemptions) were found in the domain of physical ability. Stratified analyses appeared visually parallel with no clinically meaningful effect modification by any covariate in any QoL domain.
Conclusions
Findings indicate a potential use of redeemed prescriptions as an indicator for the physical ability domains of QoL in clinical practice. The study found no clinically relevant effect modification.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12916-026-04861-5.
Keywords: Quality of life, Primary care, General practice, Multimorbidity, Polypharmacy
Background
The prevalence of multimorbidity (MM) is increasing due to an aging population, improved treatments, and a greater emphasis on early disease detection [1]. Higher degrees of MM (more illness) are naturally associated with higher health care use and reduced quality of life (QoL), and thereby often accompanied by polypharmacy (PP) [2–5].
Polypharmacy is the simultaneous use of several medications, and its prevalence is increasing worldwide [6]. Polypharmacy has no clear definition but is often discussed as either minor PP (two to four medications), major PP (five to nine medications), or excessive polypharmacy (ten or more medications), and classified as either cumulative, continuous, or simultaneous [7, 8]. Multiple definitions and classifications of PP entail study heterogeneity; however, the association of PP with reduced health outcomes, hereunder QoL, persists regardless of the method, likely due to its strong collinearity with MM [9].
The assessment of QoL is a cumbersome process: in research, this entails the dissemination of many questionnaires, and in clinical practice, systematic inquiry into QoL aspects. The former requires many resources; the latter requires too much time in the consultation. In research, QoL is often measured through generic measures such as the EuroQol-5D or the 36-Item Short Form Health Survey [10]. However, a generic measure may inquire into things not relevant to a population with different diseases or may miss relevant issues. A systematic review by Møller et al. evaluated patient-reported outcome measures (PROMs) used in MM research and found no PROM with the specific aim of measuring QoL in patients with MM [11]. Since then, a novel Multimorbidity Questionnaire 1 (MMQ1) instrument has been developed specifically for patients with MM, providing a focused understanding of the association between MM and QoL [12].
General practitioners (GPs) manage nearly all patients with MM and are responsible for most prescriptions. The identification of patients with chronic diseases managed exclusively in primary care in research is, however, limited, as they are not reported to central registries [13]. Data on diagnoses from primary care are not accessible, as the Danish Registers only hold diagnosis data coded at contact with secondary care. In contrast, data on prescription medicines are recorded systematically at redemption. Prescription data, therefore, provide a complete and patient-centered source of information [14]. Positively, medicine consumption often reflects the level of morbidity and may even add additional information about severity [3].
In clinical practice, access to the patients’ medication lists is more accessible than PROMs and systematic inquiry, and the list is often looked at by the GP during the consultation. Since previous studies have found a linear association between the number of conditions and reduced QoL, we wondered whether a similar association exists between the number of medications and QoL [15]. If so, knowledge of medication use can provide clinicians with insight into patients’ QoL in routine clinical practice, warranting clinician attention.
Objective
This study aimed to explore the correlation between the number of redeemed unique prescription medications and QoL in a cohort of adult patients who visited their GP for a chronic disease consultation in 2022. Additionally, we evaluated whether the number of redeemed unique prescription medications may serve as an indicator of different domains of QoL.
Methods
Study design, setting, and participants
In Denmark, all citizens have a unique civil registration number, which makes it possible to link data from questionnaires with Danish registry data. Virtually all Danish citizens are listed with a specific GP clinic. Long-term conditions are managed annually through chronic care consultations, which are systematically recorded. Data on diagnoses are not registered at the primary care level; however, data on medicine use are registered at redemption at the pharmacy.
This national cross-sectional study is based on the baseline data of the Cohort of Health-Related Outcomes in chroNic Illness Care in General Practice (CHRONIC-GP). In 2022, all Danish GP clinics were invited to participate in the MM600 trial that was to commence in 2023 [16]. In conjunction with this trial, all adult patients who had attended an annual chronic disease consultation or were listed as chronic-care patients in the 250 participating practices in 2022 were sent a questionnaire survey via secure online mail in early 2023, before any MM600 trial intervention activity. The patients who initiated the questionnaire constitute the baseline population of the CHRONIC-GP that has been described elsewhere [17].
Variables and data sources/measurement
The outcome was needs-based QoL measured using the six domains of MMQ1 disseminated in the 2023 questionnaire survey [17]. The measure was developed and validated in patients with chronic condition(s). It assesses six domains of needs-based QoL: physical ability (six items), worries (six items), limitations in everyday life (ten items), social life (six items), self-image (six items), and economy (three items). The items had four response options, valued 0 to 3, which were summed and scaled to obtain a score for each domain ranging from 0 (not burdened) to 100 (highly burdened). MMQ1 has shown excellent psychometric properties and notably did not have differential item functioning with disease groups [12, 18].
Exposure was the number of redeemed unique prescription medicines at the fourth Anatomical Therapeutic Chemical (ATC) level. The fourth level was used, as two subscriptions from the same fifth-level ATC group are only observed in practice when a patient switches between medications within the same chemical, pharmacological, or therapeutic subgroup, e.g., due to side effects [19]. A cumulative approach was applied, as the Danish National Prescription Registry does not contain information on dosage. A simultaneous method would require several assumptions and could introduce additional bias. Only medications redeemed between 0 and 120 days prior to questionnaire initiation were included. Medication data include medications purchased in any pharmacy in Denmark and do not include medicines dispensed at hospitals. As this study concerns long-term conditions, systemic medications for infectious diseases (ATC J group) were omitted. This methodological decision further minimized the risk of overestimating PP. The maximum number of redemptions was 26, but to make the analysis more robust to outliers, participants with more than 16 redemptions (the 99.7% percentile) were added to the 16 + group.
Sex (female, male) and age (categorized into six groups: 18–39, 40–49, 50–59, 60–69, 70–79, and 80 + years) at the time of questionnaire distribution were determined based on the participants’ civil registration numbers. Data on cohabitation status (living alone, not living alone) were self-reported and originated from the questionnaire. Data from Statistics Denmark provided information on the highest level of education completed. Education was categorized according to the International Standard Classification of Education (ISCED): up to 10 years of education—primary school (ISCED levels 0–2), 11–13 years—secondary school (ISCED level 3), and 14 or more years—higher educations (ISCED levels 4–8); individuals with no information on education were analyzed as a separate group.
This study presents the number of redeemed unique prescription medicines that determine a minimal clinically important difference (MCID) for each MMQ1 domain. The MCID was defined as the mean difference of MMQ1 scores between the global health scale item responses “poor” and “acceptable.” The method has previously been described by Jørgensen et al., and the MCID found in that study was scaled to this study’s 0–100 range (multiplied by 33.3) [15]. The number of redeemed unique prescriptions that corresponded to an MCID for each MMQ1 domain was calculated by dividing the scaled MCID by the coefficients of the linear models of Fig. 1.
Fig. 1.
Unadjusted association between unique prescription medicines redeemed (at the 4th ATC level) 120 days preceding their initiation of the questionnaire and QoL assessed using the six-dimensional MMQ1 tool. The left y-axis shows the number of participants, the right y-axis shows MMQ1 score (a higher score indicates a more burdened QoL), and the x-axis shows the number of unique prescription medicines redeemed. The equations are linear models with the exposure variable set to numeric, and the brackets contain the 95% CI of the coefficient. The colored lines are the effects that are modelled in a categorical model, where each instance of the exposure variable is its own category. The reported MCID values of each domain reflect the change of MMQ1 score necessary for a clinically important difference. The dashed line represents the sum of the MCID and the MMQ1 score at the intercept (0 redemptions). The number of redeemed unique prescription medications that constitute an MCID is found on the x-axis at the intersection between the dashed line and the linear fit. R2 values of the respective domains: physical ability, 0.12; worries, 0.08; limitations in everyday life, 0.10; social life, 0.05; self-image, 0.04; economy, 0.01. All NAs were due to missing outcome variables. ATC, anatomical therapeutic chemical; QoL, quality of life; MMQ1, MultiMorbidity Questionnaire 1; N, number of participants; NA, not applicable; MCID, minimal clinically important difference; CI, confidence interval
Study size
A total of 160,584 people were invited to the baseline survey of the CHRONIC-GP [17]. All 35,977 participants who initiated the questionnaire were included in this study.
Statistical methods
Continuous variables are described by mean and standard deviation (SD), categorical variables by number (n) and percentage (%). The association between redeemed unique prescription medications and the MMQ1 domains was analyzed in linear regression models with redeemed unique prescription medications as a continuous variable. Effect modification analyses were performed for all covariates using both adjusted and unadjusted multivariable linear models. Slope estimates relative to the unstratified slopes were presented. We excluded observations with missing data from the linear regressions due to a few missing data. Averages based on fewer than three observations were not shown, and the total number of observations that were not shown was listed as censored. R Core (Version 2025.05.0) was used for all computations [20].
Results
The population consisted of equal proportions of females (51.9%) and males, with a mean age of 65.5 years (SD 12.9) (Table 1). Most participants (72.5%) were aged 60 years or older. Participants with higher educational levels were more represented in the groups with fewer redeemed unique prescription medications. In contrast, participants with lower educational levels were more prevalent in groups with more unique medications. Living alone was reported by 23.8% of participants with zero or one redemption, increasing to 35.8% among those with ten or more redemptions. On average, participants redeemed 4.7 (SD 3.2) unique prescription medications in the 120 days preceding their initiation of the questionnaire.
Table 1.
Baseline characteristics
| Variable | Overall (n = 35,977) | 0–1 medications* (n = 5090) | 2–4 medications* (n = 14,355) | 5–9 medications* (n = 13,605) | 10 + medications* (n = 2927) | p-value1 |
|---|---|---|---|---|---|---|
| Female, n (%) | 18,665 (51.9%) | 2845 (55.9%) | 7691 (53.6%) | 6702 (49.3%) | 1427 (48.8%) | < 0.0001 |
| Age groups, n (%) | < 0.0001 | |||||
| 18–39 years | 1595 (4.4%) | 557 (10.9%) | 785 (5.5%) | 233 (1.7%) | 20 (0.7%) | |
| 40–49 years | 2263 (6.3%) | 534 (10.5%) | 1032 (7.2%) | 583 (4.3%) | 114 (3.9%) | |
| 50–59 years | 6045 (16.8%) | 1163 (22.8%) | 2690 (18.7%) | 1878 (13.8%) | 314 (10.7%) | |
| 60–69 years | 10,499 (29.2%) | 1434 (28.2%) | 4349 (30.3%) | 3900 (28.7%) | 816 (27.9%) | |
| 70–79 years | 11,766 (32.7%) | 1125 (22.1%) | 4299 (29.9%) | 5175 (38.0%) | 1,167 (39.9%) | |
| 80 + years | 3809 (10.6%) | 277 (5.4%) | 1200 (8.4%) | 1836 (13.5%) | 496 (16.9%) | |
| Education, n (%) | < 0.0001 | |||||
| No information | 598 (1.7%) | 134 (2.6%) | 222 (1.5%) | 188 (1.4%) | 54 (1.8%) | |
| Primary school | 6940 (19.3%) | 705 (13.9%) | 2392 (16.7%) | 3043 (22.4%) | 800 (27.3%) | |
| Secondary school | 15,888 (44.2%) | 2084 (40.9%) | 6203 (43.2%) | 6259 (46.0%) | 1342 (45.8%) | |
| Higher educations | 12,551 (34.9%) | 2167 (42.6%) | 5538 (38.6%) | 4115 (30.2%) | 731 (25.0%) | |
| Living alone, n (%) | 9051 (27.3%) | 1049 (23.8%) | 3341 (25.3%) | 3695 (28.9%) | 966 (35.8%) | < 0.0001 |
| Missing | 2882 (8.0%) | 678 (13.3%) | 1152 (8.0%) | 827 (6.1%) | 225 (7.7%) | |
| Unique redemptions (4th ATC level), mean (SD) | 4.7 (3.2) | 0.7 (0.5) | 3.0 (0.8) | 6.5 (1.3) | 11.7 (2.1) |
Baseline characteristics of the participants stratified into commonly used groups of PP: 0–1 medications = No PP, 2–4 medications = minor PP, 5–9 medications = major PP, 10 + medications = excessive polypharmacy. All exposure variables and possible confounders are listed. PP polypharmacy, ATC anatomical therapeutic chemical. *Redeemed unique prescription medicines at the 4th ATC level; 1Pearson’s Chi-squared test or one-way analysis of means
Figure 1 illustrates the unadjusted association between redeemed unique prescriptions and impaired needs-based QoL. Higher MMQ1 scores indicate a more burdened QoL. Their association follows a positive, linear, and significant dose–response pattern in all domains of the MMQ1. The number of redeemed unique prescriptions that correspond to an MCID for each MMQ1 domain was as follows: 4.4 (physical ability), 7.9 (worries), 8.8 (limitations in everyday life), 13.1 (social life), 13.3 (self-image), and 33.0 (economy).
Figure 2 and Additional file 1: Figs. S1–S5 visualize the association between redeemed unique prescription medications and QoL in the different domains of MMQ1 stratified by the covariates sex, age, education, and cohabitation status. Analysis of the stratified analyses in the different domains reveals only minor effect modification (Table 2). Despite the gaping confidence intervals, it was noticeable how visually parallel the slopes in the charts appeared. The statistical significance was likely attributable to the large sample size.
Fig. 2.
Visualizing the association between unique prescription medicines redeemed (at the 4th ATC level) 120 days preceding questionnaire initiation and the MMQ1 domain Physical ability strata by sex, age, education, and cohabitation status. Higher scores indicate a more burdened QoL. ATC, anatomical therapeutic chemical; MMQ1, MultiMorbidity Questionnaire 1; QoL, quality of Life
Table 2.
Unadjusted and adjusted linear models on the effect modifiers of the association between redeemed unique 4th-level ATC prescription medications in 120 days and the six domains of MMQ1. Slopes were estimated using a linear regression model including the main effect of the potential effect modifier and its interaction with the number of redeemed unique prescription medications. The main effect of the number of redeemed unique prescription medications was included only through the interaction term. The adjusted models were adjusted for all covariates
| Unadjusted model | Adjusted model | Unadjusted model | Adjusted model | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Slope of regression (relative to unstratified slope) | Lower CI | Upper CI | Slope of regression (relative to unstratified slope) | Lower CI | Upper CI | Slope of regression (relative to unstratified slope) | Lower CI | Upper CI | Slope of regression (relative to unstratified slope) | Lower CI | Upper CI | |
| Physical ability | Worries | |||||||||||
| 18–39 years | 3.35 (0.17) | 2.79 | 3.92 | 2.97 (− 0.06) | 2.38 | 3.56 | 3.13 (0.18) | 2.53 | 3.73 | 2.80 (− 0.07) | 2.17 | 3.43 |
| 40–49 years | 3.53 (0.35) | 3.19 | 3.88 | 3.37 (0.34) | 3.02 | 3.72 | 3.22 (0.27) | 2.86 | 3.59 | 3.15 (0.28) | 2.77 | 3.52 |
| 50–59 years | 3.27 (0.09) | 3.05 | 3.48 | 3.07 (0.04) | 2.85 | 3.29 | 3.12 (0.17) | 2.89 | 3.36 | 3.03 (0.16) | 2.80 | 3.27 |
| 60–69 years | 3.07 (− 0.11) | 2.91 | 3.22 | 2.90 (− 0.13) | 2.74 | 3.05 | 2.84 (− 0.11) | 2.68 | 3.01 | 2.76 (− 0.11) | 2.59 | 2.92 |
| 70–79 years | 3.15 (− 0.03) | 3.01 | 3.29 | 3.04 (0.01) | 2.90 | 3.18 | 2.94 (− 0.01) | 2.79 | 3.09 | 2.89 (0.02) | 2.74 | 3.04 |
| 80 + years | 3.29 (0.11) | 3.04 | 3.54 | 3.15 (0.12) | 2.90 | 3.40 | 2.89 (− 0.06) | 2.62 | 3.15 | 2.74 (− 0.13) | 2.47 | 3.01 |
| Female | 3.12 (0.13) | 3.01 | 3.24 | 3.13 (0.10) | 3.01 | 3.24 | 2.60 (0.07) | 2.48 | 2.72 | 2.91 (0.04) | 2.79 | 2.94 |
| Male | 2.85 (− 0.14) | 2.73 | 2.96 | 2.93 (− 0.10) | 2.81 | 3.05 | 2.46 (− 0.07) | 2.33 | 2.59 | 2.81 (− 0.06) | 2.68 | 3.04 |
| Living alone | 2.88 (− 0.01) | 2.72 | 3.03 | 2.93 (− 0.10) | 2.78 | 3.08 | 2.26 (− 0.20) | 2.09 | 2.42 | 2.54 (− 0.33) | 2.38 | 2.70 |
| Not living alone | 2.89 (0.00) | 2.79 | 2.99 | 3.08 (0.05) | 2.97 | 3.18 | 2.54 (0.08) | 2.43 | 2.65 | 3.01 (0.24) | 2.90 | 3.12 |
| No educational information | 3.06 (0.22) | 2.46 | 3.66 | 3.49 (0.46) | 2.88 | 4.10 | 2.40 (− 0.03) | 1.74 | 3.06 | 3.27 (0.40) | 2.62 | 3.92 |
| Primary school | 2.67 (− 0.17) | 2.50 | 2.85 | 2.80 (− 0.23) | 2.62 | 2.98 | 2.14 (− 0.29) | 1.95 | 2.34 | 2.51 (− 0.36) | 2.32 | 2.71 |
| Secondary school | 2.96 (0.10) | 2.84 | 3.08 | 3.14 (0.11) | 3.02 | 3.27 | 2.62 (0.19) | 2.48 | 2.75 | 3.03 (0.16) | 2.90 | 3.17 |
| Higher educations | 2.76 (− 0.08) | 2.62 | 2.91 | 3.00 (− 0.03) | 2.85 | 3.15 | 2.36 (− 0.07) | 2.20 | 2.52 | 2.84 (− 0.03) | 2.69 | 3.00 |
| Limitations in everyday life | Social life | |||||||||||
| 18–39 years | 3.42 (0.53) | 2.87 | 3.97 | 3.04 (0.29) | 2.47 | 3.61 | 2.45 (0.65) | 2.00 | 2.89 | 2.17 (0.49) | 1.72 | 2.62 |
| 40–49 years | 3.49 (0.60) | 3.16 | 3.83 | 3.33 (0.58) | 2.99 | 3.66 | 2.63 (0.83) | 2.36 | 2.90 | 2.48 (0.80) | 2.21 | 2.75 |
| 50–59 years | 3.02 (0.13) | 2.80 | 3.23 | 2.85 (0.10) | 2.64 | 3.06 | 2.20 (0.40) | 2.03 | 2.37 | 2.06 (0.38) | 1.88 | 2.23 |
| 60–69 years | 2.65 (− 0.24) | 2.51 | 2.80 | 2.51 (− 0.24) | 2.36 | 2.65 | 1.69 (− 0.11) | 1.57 | 1.81 | 1.57 (− 0.11) | 1.46 | 1.69 |
| 70–79 years | 2.83 (− 0.06) | 2.70 | 2.97 | 2.74 (− 0.01) | 2.61 | 2.88 | 1.57 (− 0.23) | 1.46 | 1.68 | 1.49 (− 0.19) | 1.38 | 1.59 |
| 80 + years | 3.09 (0.20) | 2.86 | 3.33 | 2.93 (0.18) | 2.69 | 3.17 | 1.75 (− 0.05) | 1.56 | 1.94 | 1.62 (− 0.06) | 1.43 | 1.82 |
| Female | 2.77 (0.16) | 2.66 | 2.88 | 2.88 (0.13) | 2.77 | 2.99 | 1.58 (0.10) | 1.49 | 1.68 | 1.77 (0.09) | 1.68 | 1.86 |
| Male | 2.45 (− 0.16) | 2.33 | 2.56 | 2.61 (− 0.14) | 2.49 | 2.72 | 1.36 (− 0.12) | 1.27 | 1.46 | 1.59 (− 0.09) | 1.50 | 1.68 |
| Living alone | 2.58 (0.08) | 2.43 | 2.73 | 2.73 (− 0.02) | 2.58 | 2.88 | 1.65 (0.28) | 1.53 | 1.77 | 1.86 (0.18) | 1.75 | 1.98 |
| Not living alone | 2.46 (− 0.04) | 2.36 | 2.56 | 2.75 (0.00) | 2.66 | 2.85 | 1.25 (− 0.12) | 1.17 | 1.33 | 1.60 (− 0.08) | 1.52 | 1.68 |
| No educational information | 3.15 (0.67) | 2.56 | 3.75 | 3.80 (1.05) | 3.21 | 4.38 | 2.11 (0.72) | 1.62 | 2.59 | 2.80 (1.12) | 2.34 | 3.27 |
| Primary school | 2.28 (− 0.20) | 2.10 | 2.45 | 2.50 (− 0.25) | 2.32 | 2.67 | 1.27 (− 0.12) | 1.13 | 1.42 | 1.52 (− 0.16) | 1.38 | 1.66 |
| Secondary school | 2.61 (0.13) | 2.49 | 2.73 | 2.86 (0.11) | 2.74 | 2.98 | 1.49 (0.10) | 1.39 | 1.58 | 1.75 (0.07) | 1.66 | 1.85 |
| Higher educations | 2.41 (− 0.07) | 2.26 | 2.55 | 2.70 (− 0.05) | 2.56 | 2.84 | 1.28 (− 0.11) | 1.16 | 1.40 | 1.62 (− 0.06) | 1.51 | 1.74 |
| Self-image | Economy | |||||||||||
| 18–39 years | 2.49 (0.46) | 1.97 | 3.00 | 2.24 (0.32) | 1.72 | 2.77 | 3.57 (1.91) | 3.01 | 4.12 | 3.45 (1.93) | 2.88 | 4.01 |
| 40–49 years | 2.89 (0.86) | 2.58 | 3.20 | 2.76 (0.84) | 2.45 | 3.07 | 3.37 (1.71) | 3.03 | 3.70 | 3.21 (1.69) | 2.88 | 3.55 |
| 50–59 years | 2.59 (0.56) | 2.39 | 2.79 | 2.46 (0.54) | 2.26 | 2.65 | 2.96 (1.30) | 2.74 | 3.17 | 2.77 (1.25) | 2.56 | 2.98 |
| 60–69 years | 1.95 (− 0.08) | 1.81 | 2.08 | 1.83 (− 0.09) | 1.69 | 1.97 | 1.78 (0.12) | 1.64 | 1.93 | 1.64 (0.12) | 1.49 | 1.78 |
| 70–79 years | 1.79 (− 0.24) | 1.66 | 1.91 | 1.72 (− 0.20) | 1.60 | 1.85 | 0.94 (− 0.72) | 0.80 | 1.07 | 0.83 (− 0.69) | 0.70 | 0.97 |
| 80 + years | 1.75 (− 0.28) | 1.53 | 1.97 | 1.65 (− 0.27) | 1.43 | 1.88 | 0.70 (− 0.96) | 0.46 | 0.94 | 0.62 (− 0.90) | 0.38 | 0.86 |
| Female | 1.68 (0.14) | 1.57 | 1.78 | 2.05 (0.13) | 1.94 | 2.15 | 0.98 (0.01) | 0.86 | 1.10 | 1.50 (− 0.02) | 1.39 | 1.61 |
| Male | 1.39 (− 0.15) | 1.28 | 1.50 | 1.79 (− 0.13) | 1.68 | 1.90 | 0.97 (0.00) | 0.85 | 1.10 | 1.55 (0.03) | 1.43 | 1.66 |
| Living alone | 1.45 (0.00) | 1.30 | 1.59 | 1.78 (− 0.14) | 1.65 | 1.92 | 0.98 (0.07) | 0.82 | 1.13 | 1.42 (− 0.10) | 1.28 | 1.57 |
| Not living alone | 1.45 (0.00) | 1.35 | 1.54 | 1.98 (0.06) | 1.89 | 2.07 | 0.88 (− 0.03) | 0.77 | 0.98 | 1.57 (0.05) | 1.47 | 1.66 |
| No educational information | 1.57 (0.14) | 1.00 | 2.14 | 2.69 (0.77) | 2.15 | 3.23 | 1.03 (0.18) | 0.40 | 1.67 | 2.33 (0.81) | 1.75 | 2.92 |
| Primary school | 1.40 (− 0.03) | 1.23 | 1.57 | 1.83 (− 0.09) | 1.67 | 1.99 | 0.64 (− 0.21) | 0.45 | 0.83 | 1.23 (− 0.29) | 1.05 | 1.40 |
| Secondary school | 1.56 (0.13) | 1.45 | 1.68 | 2.02 (0.10) | 1.90 | 2.13 | 1.05 (0.20) | 0.92 | 1.18 | 1.67 (0.15) | 1.55 | 1.79 |
| Higher educations | 1.24 (− 0.19) | 1.10 | 1.38 | 1.81 (− 0.11) | 1.68 | 1.94 | 0.69 (− 0.16) | 0.54 | 0.85 | 1.46 (− 0.06) | 1.32 | 1.60 |
The adjusted models are adjusted for all covariates. CI confidence interval, MMQ1 MultiMorbidity Questionnaire 1, ATC anatomical therapeutic chemical
Discussion
This national observational study described a significant, linear, and positive association between redeemed unique prescription medicines and impaired QoL in adults. The finding supports the ability to use the number of redeemed unique prescription medications as an indicator for the physical ability domain of QoL in research and perhaps for the clinician. The number of redeemed unique prescription medications that constitute an MCID on QoL was between five and 33 redeemed unique prescription medications, with the domain of physical ability being the most sensitive. Further, the study found no clinically meaningful effect modification by any of the covariates.
The overall positive association between redeemed unique prescription medicines and burdened QoL aligned with our hypothesis that medicine redemption is a sign of underlying disease and, therefore, associated with more burdened QoL. This trend resembles findings from previous studies investigating the association between MM and QoL [5]. A study using the MMQ1 revealed that the domain of physical ability was the most sensitive, and that a difference in self-reported conditions of two disease groups exceeded the MCID in this domain [15]. Correspondingly, the present study found that a difference of 4.4 (five) or more redeemed unique prescription medications, i.e., the difference of no medicine use and major polypharmacy, exceeded the MCID in the domain of physical ability. Counting medications, however, is a pragmatic approach, and we are aware that some medications perhaps correlate with a larger burden.
To our knowledge, the association between PP and QoL has not previously been examined to this extent. Existing minor studies have reported significant associations between PP and burdened QoL, but the exposures were grouped, and the populations were either age-restricted or limited to specific diseases [21–24]. Similarly, we found significant associations between increasing numbers of redeemed unique prescription medications and burdened QoL. It is, however, challenging to compare the studies due to the different measuring tools. While the other studies examined health-related QoL, the MMQ1 scale assessed needs-based QoL.
Another study disentangled the direct effect of PP on QoL, finding that only about half of the observed impact was attributable to PP itself [25]. A study by Olsen et al. highlighted the contribution of underlying diseases to the overall effect/burden, and another study found that medication burden mediated 10% of the effect of the number of medications on health-related QoL [26, 27]. These findings support the use of medication data as a cumulative measure of disease and treatment, thus reflecting not only the direct pharmacological effect but also the presence of underlying conditions, disease severity, and treatment burden. This is a clinically relevant finding, as a simple assessment of the medication list may provide the clinician with an overall impression of the patient’s QoL extending beyond the effects of medication use alone, irrespective of the underlying disease(s).
This study found numerically small effect modification in all domains of QoL. Generally, less educated patients reported a lower sensitivity to each additional redemption. Since both MM and PP were more prevalent in less educated participants, perhaps this normalization had a protective effect on education’s interaction with redeemed unique prescription medications. Another observation was made in the 40–49-year-old participants. This age group reported a high sensitivity (steeper slope) to additional redeemed unique prescription medications across all QoL domains. However, it was relatively uncommon for patients aged 40–49 years to redeem larger amounts of unique prescription medication. This deviation from their peers—and possibly from their own expectations regarding health at that age—may partly explain the observed sensitivity [28]. Older participants consistently reported a low economic burden, despite an increasing number of redemptions. On the other hand, participants of working age reported significantly greater economic concerns with more redeemed unique prescription medications. Questions in the domain of economy addressed whether illness limits their ability to achieve a good economy, and whether participants worry about their financial situation and social status due to their illness. A previous study described financial concerns in patients with MM [29]. These combined findings emphasize the need for particular attention from both general practitioners and municipalities on younger patients with extensive medication lists, as it can be an indication of financial concerns. However, the observed differences were numerically small, underscoring that findings should be interpreted with caution, as they do not necessarily imply clinical relevance.
Generally, age was visually associated with large absolute differences in MMQ1 scores between the youngest and oldest participants. This potential finding, however, is beyond the scope of this paper.
The study had several strengths. It was a large and diverse cohort of adult patients in general practice with unselected chronic conditions. The cohort was linked to Danish registries with high validity, and the questionnaire was developed for patients with MM. Redeemed unique prescriptions reflected participant behavior since they were registered at redemption. The cumulative approach to the predictor variable allowed us to include all redeemed prescription medications and was not limited or biased by a drug’s assigned defined daily dose (DDD) [7]. The risk of selection bias was noticeably small because eligibility happened passively based on the participating practices’ routine care for annual chronic disease consultations.
However, it also has several limitations. There is a risk of participation bias, since respondents may differ systematically from non-respondents. A previous study on the CHRONIC-GP showed that participants were generally a bit older than the total invited population and that males had a marginally higher response rate [30]. The cohort profile discussed several reasons for not participating, e.g., the risk of false eligibility due to the inclusion method and lack of time due to a job or family situation [17]. Lastly, participants with more than 16 unique redemptions were added to the group of 16 + to avoid leverage points; this made it impossible to assess QoL for participants with more than 16 redemptions.
Even though using a questionnaire developed and validated for measuring needs-based and non-disease-specific QoL in patients with MM was a strength, this may also induce a limitation because participants were not preselected based on MM status. Consequently, some may not fulfill MM criteria and would have been better off with a more generic questionnaire. Further, we estimate that a substantial part of the invited population without MM did not identify with the intended target group and consequently dropped out. This may explain the higher dropout rates among those with 0–1 redemptions, as suggested by the missing responses on cohabitation seen in Table 1. These factors likely shifted the sample toward participants with higher medicine use and greater burdened QoL.
A notable decision was not to adjust for diagnoses or self-reported conditions, despite data being available for the cohort. We did not adjust for them because of collinearity. Further, adjustment would have introduced bias and reduced interpretability, as diagnoses are inconsistently reported. Furthermore, the clinician cannot adjust for underlying disease(s) during the brief medication list assessment. However, using medications as a predictor perhaps underestimates the disease prevalence of conditions that are not medically treated, which could partly explain why the group without redeemed medications lies slightly above the linear fit. This pattern may also reflect healthy user bias. Moreover, the lack of distinction between incident and prevalent users may further contribute to such bias. Three further limitations are the cross-sectional design, which precludes causal inference and does not protect against reverse causation, the Danish health care setting, which may limit generalizability, and the potential deprescribing in end-of-life care.
Future research should explore whether the use of medicine data as an indicator for QoL is a better measure than diagnosis data sampled in secondary care only. Additionally, it would be valuable to identify which main groups of the ATC classification system are most strongly associated with impaired QoL. Lastly, it would be of great interest to study the absolute differences between subgroups of the covariates, as they seem to differ significantly.
Conclusions
Conclusively, an increasing number of redeemed unique prescriptions was associated with both a statistically and clinically significant burdened QoL in the physical ability domain. The consistent linear relation shows the ability to use the number of redeemed unique prescription medications as an indicator for the physical ability domain of QoL in research and to some extent for the clinician. The study’s findings can be applied easily in clinical practice, as GPs have access to both medication lists and diagnoses. Diagnoses, however, may be outdated or of limited relevance to the patient, whereas redeemed prescription medications reflect current treatment and disease status. Often, GPs review the medication list during consultations—an informal practice that has until now lacked empirical evidence. Lastly, we found no clinically relevant effect modification.
Supplementary Information
Additional file 1: The association between unique prescription medicine redemptions at the 4th ATC level and the MMQ1 domains worries, limitations in everyday life, social life, self-image, and economy strata by sex, age, education, and cohabitation status; description of data: Figure S1. Stratified visualization. Figure S2. Stratified visualization. Figure S3. Stratified visualization. Figure S4. Stratified visualization. Figure S5. Stratified visualization. Higher scores indicate a more burdened QoL. ATC = Anatomical Therapeutic Chemical; MMQ1 = MultiMorbidity Questionnaire 1; QoL = Quality of Life.
Acknowledgements
The PhD of HHPL is funded by the Danish Foundation for General Practice (Fonden for Almen Praksis, Denmark). The authors have used the following AI technologies in the writing process to improve readability and language of the study: ChatGPT (GPT-5.1), Grammarly (Free version v1.2.213.1791), and the Spelling and Grammar function of Word (Microsoft 365 Apps for enterprise).
Abbreviations
- MM
Multimorbidity
- QoL
Quality of life
- PP
Polypharmacy
- GP
General practitioner
- PROM
Patient-reported outcome measure
- MMQ1
MultiMorbidity Questionnaire 1
- CHRONIC-GP
Cohort of Health-Related Outcomes in chroNic Illness Care in General Practice
- ATC
Anatomical therapeutic chemical
- ISCED
International Standard Classification of Education
- MCID
Minimal clinically important difference
- SD
Standard deviation
- DDD
Defined daily dose
Authors’ contributions
HHPL, VS, TGW, ABL, AP, FBW, and AH took part in the conceptualization of the study. HHPL, AH, and VS did the data curation, methodology, and formal analysis. HHPL did the visualization. HHPL, VS, TGW, ABL, AP, FBW, and AH took part in the interpretation of data. HHPL, TGW, and AH wrote the original draft. HHPL, VS, TGW, ABL, AP, FBW, and AH critically revised and edited the manuscript. AH acquired funding. HHPL, VS, TGW, ABL, AP, FBW, and AH had full access to all the data in the study and had final responsibility for the decision to submit for publication. All authors read and approved the final manuscript.
Funding
Open access funding provided by Copenhagen University The funding source did not influence the study design, data collection, data analysis, data interpretation, manuscript writing, or the decision to submit the paper for publication.
Data availability
The CHRONIC-GP datasets analyzed are not publicly available due to privacy regulations, as they include personal health data. Access is thus restricted to authorized researchers. Anonymized data are available from the corresponding author (HHPL) to authorized researchers only upon reasonable request and with permission from the relevant Danish Authorities.
Declarations
Ethics approval and consent to participate
The Capital Region’s Ethical Committee has assessed the CHRONIC-GP as a quality improvement project that does not need any further ethical oversight (F-24026474 and H-22041229). Participants were informed about their rights and confirmed to have received information on the scope and purpose of their participation in the research project. The consent was sampled at the initiation of the questionnaire.
Consent for publication
Not applicable.
Competing interests
HHPL holds shares in Novo Nordisk and Genmab, and the Danish Foundation for General Practice funds the PhD. AH declares having received funding for research from the Novo Nordic Foundation. TGW has received royalties from the Association of Danish Medical Students’ Publisher (FADL’s Forlag). The remaining authors declare that they have no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.van den Akker M, Vaes B, Goderis G, Van Pottelbergh G, De Burghgraeve T, Henrard S. Trends in multimorbidity and polypharmacy in the Flemish-Belgian population between 2000 and 2015. PLoS One. 2019;14(2):e0212046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Prior A, Vestergaard CH, Vedsted P, Smith SM, Virgilsen LF, Rasmussen LA, et al. Healthcare fragmentation, multimorbidity, potentially inappropriate medication, and mortality: a Danish nationwide cohort study. BMC Med. 2023;21(1):305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sinnott C, Bradley CP. Multimorbidity or polypharmacy: two sides of the same coin? J Comorb. 2015;5:29–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Soley-Bori M, Ashworth M, Bisquera A, Dodhia H, Lynch R, Wang Y, et al. Impact of multimorbidity on healthcare costs and utilisation: a systematic review of the UK literature. Br J Gen Pract. 2021;71(702):e39–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Makovski TT, Schmitz S, Zeegers MP, Stranges S, van den Akker M. Multimorbidity and quality of life: systematic literature review and meta-analysis. Ageing Res Rev. 2019;53:100903. [DOI] [PubMed] [Google Scholar]
- 6.Medication Safety in Polypharmacy. Geneva: World Health Organization; 2019 (WHO/UHC/SDS/2019.11). Licence: CC BY-NC-SA 3.0 IGO.
- 7.Monégat M, Sermet C, Perronnin M, Rococo E. Polypharmacy: definitions, measurement and stakes involved. Rev Lit Meas Tests. 2014;8.
- 8.Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is polypharmacy? A systematic review of definitions. BMC Geriatr. 2017;17(1):230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Davies LE, Spiers G, Kingston A, Todd A, Adamson J, Hanratty B. Adverse outcomes of polypharmacy in older people: systematic review of reviews. J Am Med Dir Assoc. 2020;21(2):181–7. [DOI] [PubMed] [Google Scholar]
- 10.Wells GA, Russell AS, Haraoui B, Bissonnette R, Ware CF. Validity of quality of life measurement tools–from generic to disease-specific. J Rheumatol Suppl. 2011;88:2–6. [DOI] [PubMed] [Google Scholar]
- 11.Moller A, Bissenbakker KH, Arreskov AB, Brodersen J. Specific measures of quality of life in patients with multimorbidity in primary healthcare: a systematic review on patient-reported outcome measures’ adequacy of measurement. Patient Relat Outcome Meas. 2020;11:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bissenbakker K, Siersma V, Jonsson ABR, Moller A, Christensen KB, Brodersen JB. Measuring needs-based quality of life and self-perceived health inequity in patients with multimorbidity: investigating psychometric measurement properties of the MultiMorbidity Questionnaire (MMQ) using primarily Rasch models. J Patient Rep Outcomes. 2023;7(1):94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Schmidt M, Schmidt SAJ, Adelborg K, Sundboll J, Laugesen K, Ehrenstein V, et al. The danish health care system and epidemiological research: from health care contacts to database records. Clin Epidemiol. 2019;11:563–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pottegard A, Schmidt SAJ, Wallach-Kildemoes H, Sorensen HT, Hallas J, Schmidt M. Data resource profile: the danish National Prescription Registry. Int J Epidemiol. 2017;46(3):798–798f. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jorgensen CV, Larsen HHP, Siersma V, Holm A. The association between multimorbidity and needs-based quality of life in primary care: a cross-sectional questionnaire study. Scand J Prim Health Care. 2025. 10.1080/02813432.2025.2527853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Holm A, Lyhnebeck AB, Rozing M, Buhl SF, Willadsen TG, Prior A, et al. Effectiveness of an adaptive, multifaceted intervention to enhance care for patients with complex multimorbidity in general practice: protocol for a pragmatic cluster randomised controlled trial (the MM600 trial). BMJ Open. 2024;14(2):e077441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Larsen HH, Willadsen TG, Prior A, Lyhnebeck AB, Waldorff FB, Holm A. Methods and baseline results of the cohort of health-related outcomes in chronic illness care in general practice in Denmark (CHRONIC-GP). BMJ Open. 2025;15(11):e103807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Sweeney K, Bissenbakker K, Siersma V, Jonsson A, Donaghy E, Henderson D, et al. The multimorbidity questionnaire (MMQ1): English translation and validation of a Danish patient reported outcome measure for quality of life in people with multiple long-term conditions in a cross-sectional survey. Qual Life Res. 2025. 10.1007/s11136-025-03901-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Franchi C, Cartabia M, Risso P, Mari D, Tettamanti M, Parabiaghi A, et al. Geographical differences in the prevalence of chronic polypharmacy in older people: eleven years of the EPIFARM-Elderly Project. Eur J Clin Pharmacol. 2013;69(7):1477–83. [DOI] [PubMed] [Google Scholar]
- 20.R Core Team. R: A language and environment for statistical computing. Austria: R Foundation for Statistical Computing V; 2021. [Google Scholar]
- 21.Akkala S, Iqbal M, Hansen R, Akkula J. Association of polypharmacy and health-related quality of life among US adults: a cross-sectional analysis of the 2022 MEPS data. J Ageing Longev. 2025. 10.3390/jal5040052. [Google Scholar]
- 22.Vonderhaar JM, Ernst ME, Fravel MA, Orchard SG, Owen AJ, Woods RL, et al. Prescription and non-prescription medication pill burdens and their associations with health-related quality of life in older adults: a cross-sectional study. Drugs Aging. 2025;42(5):457–67. [DOI] [PubMed] [Google Scholar]
- 23.Zarinfar Y, Panahi N, Shojaei R, Hosseinpour M, Nabipour I, Larijani B, et al. The association between polypharmacy and quality of life in elderly population in Southern Iran: Bushehr Elderly Health (BEH) Program. BMC Public Health. 2025;25(1):4146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Van Wilder L, Devleesschauwer B, Clays E, Pype P, Vandepitte S, De Smedt D. Polypharmacy and health-related quality of life/psychological distress among patients with chronic disease. Prev Chronic Dis. 2022;19:E50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Schenker Y, Park SY, Jeong K, Pruskowski J, Kavalieratos D, Resick J, et al. Associations between polypharmacy, symptom burden, and quality of life in patients with advanced, life-limiting illness. J Gen Intern Med. 2019;34(4):559–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bjork E, Rentsch CT, Desai RJ, Andersen JH, Lundby C, Pottegard A. Polypharmacy’s association with mortality: confounding from underlying morbidity. J Am Geriatr Soc. 2025;73(8):2485–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gurung A, Ogden E, Chen WS. Experience of the burden of using multiple medicines and the associated impact on health-related quality of life. J Patient Exp. 2025;12:23743735251330353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bissenbakker KH, Møller A, Brodersen J, Jønsson AR. PROMs og livskvalitet: Hvordan måles livskvalitet hos mennesker med multisygdom?. Tidsskrift for Forskning I Sygdom Og Samfund. 2020;17(32). 10.7146/tfss.v17i32.120975.
- 29.Larkin J, Foley L, Smith SM, Harrington P, Clyne B. The experience of financial burden for people with multimorbidity: a systematic review of qualitative research. Health Expect. 2020;24(2):282–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Holm A, Lyhnebeck AB, Buhl SF, Bissenbakker K, Kristensen JK, Moller A, et al. Development of a PROM to measure patient-centredness in chronic care consultations in primary care. Health Qual Life Outcomes. 2025;23(1):4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional file 1: The association between unique prescription medicine redemptions at the 4th ATC level and the MMQ1 domains worries, limitations in everyday life, social life, self-image, and economy strata by sex, age, education, and cohabitation status; description of data: Figure S1. Stratified visualization. Figure S2. Stratified visualization. Figure S3. Stratified visualization. Figure S4. Stratified visualization. Figure S5. Stratified visualization. Higher scores indicate a more burdened QoL. ATC = Anatomical Therapeutic Chemical; MMQ1 = MultiMorbidity Questionnaire 1; QoL = Quality of Life.
Data Availability Statement
The CHRONIC-GP datasets analyzed are not publicly available due to privacy regulations, as they include personal health data. Access is thus restricted to authorized researchers. Anonymized data are available from the corresponding author (HHPL) to authorized researchers only upon reasonable request and with permission from the relevant Danish Authorities.