Ninety days to predict a year: a Bayesian journey into medication adherence post-myocardial infarction

This editorial refers to ‘Predictive modelling of medication adherence in post-myocardial infarction patients: a Bayesian approach using beta-regression’, by E.E. Tannous et al., https://doi.org/10.1093/eurjpc/zwae327. The management of medication adherence in patients recovering from myocardial infarction (MI) remains a cornerstone of secondary prevention. Medication adherence is often suboptimal, even in after acute serious non-fatal cardiovascular events, and is known to progressively decline during the year following index event. At mid-term follow-up, a Swedish real-world cohort study1 reported <60% of adherence to statin following acute coronary syndrome. Beyond adherence, assessment of risk factors is often delayed. The recommended yearly measurement of LDL cholesterol is only performed in <30% of patients,2 resulting in an obvious suboptimal management, impaired outcome, and increased healthcare costs. The determinants of poor adherence are complex and may markedly vary according to age, social group, countries, and healthcare system or underlying disease. However, it seems associated with older age, adverse social determinants (i.e. unfavourable educational, social, economic, or environmental conditions), and habits and behaviour rather than cognitive status.3,4