Non-adherence to oral antidiabetic medications among patients with type 2 diabetes: evidence from a nationwide multicentre study in a lower-middle-income country

Abstract

Background: Non-adherence to oral antidiabetic medications is a critical, multifactorial challenge that undermines disease management nationwide. This study aimed to determine the prevalence and predictors of non-adherence among patients with type 2 diabetes (T2DM) across Pakistan. We conducted a nationwide cross-sectional study between December 2023 and March 2025 in 225 healthcare facilities across all provinces of Pakistan. A total of 41,095 participants were selected using a multistage random sampling approach. Medication adherence was evaluated with self-reported pill counts and the Self-Efficacy for Appropriate Medication Use Scale (SEAMS). Multivariable logistic regression using a generalized linear model (GLM) logit approach was employed integrated with Interaction analysis to discover predictors of non-adherence. Result: Non-adherence was present in 59.9% (n=24,616) of cases (n=24,616). Factors associated with reduced adherence included being female (AOR = 0.229, 95% CI: 0.067–0.779, P = 0.02), having only primary or secondary education (AOR = 0.186, 95% CI: 0.166–0.208, P < 0.001), absence of hospitalization, taking fewer medications per day, less frequent dosing schedules, and uncontrolled HbA1c levels. In contrast, factors linked to increased adherence were being married (AOR = 4.118, 95% CI: 3.611–4.695, P < 0.001), being a non-smoker (AOR = 12.98, 95% CI: 11.53–14.62, P < 0.001), having no family history of diabetes (AOR = 1.074, 95% CI: 1.021–1.129, P = 0.005), and having more than one comorbid condition. Conclusion: Non-adherence to oral antidiabetic medication is prevalent and influenced by diverse personal, socioeconomic, and health-related factors. Targeted interventions addressing modifiable risk factors could improve adherence, ultimately reduce the healthcare burden and enhancing outcomes for patients with diabetes in Pakistan.