Economic, cultural, and social inequalities in potentially inappropriate medication: A nationwide survey- and register-based study in Denmark.

Abstract

Background: Potentially inappropriate medication (PIM) is associated with negative health outcomes and can serve as an indicator of treatment quality. Previous studies have identified social inequality in treatment but often relied on narrow understandings of social position or failed to account for mediation by differential disease risk among social groups. Understanding how social position influences PIM exposure is crucial for improving the targeting of treatment quality and addressing health disparities. This study investigates the association between social position and PIM, considering the mediation effect of long-term conditions.

Methods and findings: This cross-sectional study utilized data from the 2017 Danish National Health Survey, including 177,495 individuals aged 18 or older. Data were linked to national registers on individual-level. PIM was defined from the STOPP/START criteria and social position was assessed through indicators of economic, cultural, and social capital (from Bourdieu's Capital Theory). We analyzed odds ratios (ORs) and prevalence proportion differences (PPDs) for PIM using logistic regression, negative binomial regression, and generalized structural equation modeling. The models were adjusted for age and sex and analyzed separately for indicators of under- (START) and overtreatment (STOPP). The mediation analysis was conducted to separate direct and indirect effects via long-term conditions. Overall, 14.7% of participants were exposed to one or more PIMs, with START PIMs being more prevalent (12.5%) than STOPP PIMs (3.1%). All variables for social position except health education were associated with PIM in a dose-response pattern. Individuals with lower wealth (OR: 1.85 [95% CI 1.77, 1.94]), lower income (OR: 1.78 [95% CI 1.69, 1.87]), and lower education level (OR: 1.66 [95% CI 1.56, 1.76]) exhibited the strongest associations with PIM. Similar associations were observed for immigrants, people with low social support, and people with limited social networks. The association with PIM remained significant for most variables after accounting for mediation by long-term conditions. The disparities were predominantly related to overtreatment and did not relate to the number of PIMs. The study's main limitation is the risk of reverse causation due to the complex nature of social position and medical treatment.

Conclusions: The findings highlight significant social inequalities in PIM exposure, driven by both economic, cultural, and social capital despite a universal healthcare system. Understanding the social determinants of PIM can inform policies to reduce inappropriate medication use and improve healthcare quality and equity.