JMIR Publications

ABSTRACT

Medication adherence involves many factors and it assumes that the patient is in agreement. Thus, the term that simply describe the extent to which the patient is taking the medication correctly, is called medication compliance. When analyzing numerous factors for medication adherence and compliance simultaneously, multicollinearity and variable selection can be problematic. We addressed these issues using a regularization. In addition, we further investigated the importance of using LightGBM to avoid the limited representation of the relationship between predictor variables and the response variable. The data were obtained through a questionnaire survey of adult patients taking medication in Japan (n=638). The regularization model showed significant differences for variables such as “using the drug at approximately the same time each day (coefficient 0.479, p=0.016)”, “taking meals at approximately the same time each day (coefficient 0.407, p=0.016)”, and “I would like to have my medication reduced (coefficient -0.410, p=0.012)”. The LightGBM calculated the Feature Importance for each variable, highlighting the importance of timing consistency in medication and mealtime, and attitudes towards accepting/refusing medication, for medication compliance. Our study is the first to suggest that lifestyle habits is the most important factor in medication compliance in general patient population.