Exploring the Multidimensional Relationship Between Medication Beliefs and Adherence to Medications Among Older Adults Living With Multimorbidity Using Polynomial Regression: An Observational Cohort Study

Abstract

Abstract Background People living with multimorbidity may hold complex beliefs about medicines, potentially influencing adherence. Polynomial regression offers a novel approach to examining the multidimensional relationship between medication beliefs and adherence, overcoming limitations associated with difference scores. Purpose To explore the multidimensional relationship between medication beliefs and adherence among people living with multimorbidity. Methods Secondary analysis was conducted using observational data from a cohort of older adults living with ≥2 chronic conditions, recruited from 15 family practices in Ireland in 2010 (n = 812) and followed up in 2012 (n = 515). Medication beliefs were measured with the Beliefs about Medicines Questionnaire-Specific. Adherence was assessed with the medication possession ratio using prescription data from the national primary care reimbursement service. Polynomial regression was used to explore the best-fitting multidimensional models for the relationship between (i) beliefs and adherence at baseline, and (ii) beliefs at baseline and adherence at follow-up. Results Confirmatory polynomial regression rejected the difference-score model, and exploratory polynomial regression indicated quadratic models for both analyses. Reciprocal effects were present in both analyses (slope [Analysis 1]: β = 0.08, p = .007; slope [Analysis 2]: β = 0.07, p = .044), indicating that adherence was higher when necessity beliefs were high and concern beliefs were low. Nonreciprocal effects were also present in both analyses (slope [Analysis 1]: β = 0.05, p = .006; slope [Analysis 2]: β = 0.04, p = .043), indicating that adherence was higher when both necessity and concern beliefs were high. Conclusions Among people living with multimorbidity, there is evidence that the relationship between medication beliefs and adherence is multidimensional. Attempts to support adherence should consider the combined role of necessity and concern beliefs.