Development of models to predict medication non-adherence based on a new typology

Medication non-adherence, the extent to which a person's behavior does not coincide with medical or health advice, is a serious public health issue.

Objectives: 1) Develop a new typology of medication non-adherence, 2) Develop models to predict different types of non-adherence based on Andersen Behavioral Model (ABM) and Leventhal's Common Sense Model (CSM), and 3) Test the models across two different medications used in treating disease conditions with varying symptomatology.

Methodology: A new typology of medication non-adherence was developed through literature review of the frequently reported reasons for non-adherence based on the possibility of a cognitive process intervention directed towards patients and the mutability of interventions. The typology was analyzed qualitatively and quantitatively. A new self-reported scale to measure non-adherence was developed from the frequently reported reasons and compared to the Morisky scale. The conceptual models developed using ABM and CSM were tested using regression techniques to identify significant predictors of non-adherence.

Results: Qualitative analysis supported the typology from the literature review, yet the quantitative exploratory factor analysis did not support it. Instead, four types of non-adherence each for cholesterol lowering (non-adherence due to managing issues, multiple medication issues, belief issues with medications, forgetfulness due to busy schedule) and asthma maintenance medications (non-adherence due to managing and availability issues, beliefs and convenience issues, cost issues, forgetfulness due to busy schedule) were identified. Predisposing factors such as concern beliefs in medications, enabling factors such as self efficacy, and need factors such as self health and illness perceptions, and severity of disease were significant predictors of medication non-adherence. The Reasons scale had moderate levels of agreement with the Morisky scale based on kappa coefficients.

Conclusion: No one typology of medication non-adherence fit cholesterol lowering and asthma maintenance medications, and the typology was driven by type of disease condition and reasons for non-adherence. The Reasons scale measured and categorized non-adherence better than the Morisky scale. Adding CSM to ABM facilitated in identifying predictors of medication non-adherence.