Abstract
ABSTRACT
Introduction
Optimizing treatment adherence behaviour has become a focus in healthcare, as improved adherence enhances clinical outcomes, patient experiences and health system sustainability. For this optimization endeavour scholars have posited diverse variables divided into two clusters. The first variable cluster emphasizes treatment priority, treatment concerns and health consciousness. The second variable cluster highlights acceptance, control and subjective health experience. Although some of these variables have been tested in isolation, few comprehensive models have been established. Therefore, this study aims at establishing a comprehensive model for explaining treatment adherence behaviour.
Method
An online questionnaire was administered to 4081 participants from six different countries. This questionnaire covered sample characteristics as well as instruments measuring health consciousness, treatment priority, treatment concerns, acceptance, control, subjective health experiences, medication compliance, health behaviour and treatment adherence behaviour. Data were analysed using descriptive, reliability, validity and model statistics.
Result
All concepts in the established model were measured reliably and showed sufficient validity. Most associations between concepts were stable and statistically significant across the overall and segmented models, except for those associating treatment concerns with adherence behaviour and control with subjective health experience in the previous month. The analysis shows that health consciousness is modestly to moderately positively associated with treatment adherence indirectly via 12 intermediary pathways, four of which were consistently significant, two were largely nonsignificant, and the remaining pathways exhibited mixed patterns. Between 22.6% and 38.5% of variance was explained by the models.
Conclusion
This model offers valuable insights into treatment adherence behaviour yet remains open to further refinement and development, either by striving for greater parsimony or by embracing relational complexity through data‐driven machine learning approaches.