Question
Which patient groups have the highest non-adherence rates, and should be reached first by adherence programs?
Approach
I computed baseline adherence, ranked non-adherence by age, gender, condition, education, social support and access to care, and estimated odds ratios with logistic regression. The results feed a Shiny dashboard with KPIs, a prioritized targeting table and the odds-ratio drivers.
Limits
The data are observational, so the odds ratios show associations, not causes. The project ranks who to target; it does not measure whether an intervention works.