Visualizing nationwide variation in Medicare Part D prescribing patterns

Objective: To characterize the regional and national variation in prescribing patterns in the Medicare Part D program using machine learning and dimensional reduction visualization methods.

Methods: Using publicly available Medicare Part D claims data, we identified regional and national provider prescribing profile variation with unsupervised clustering and t-distributed stochastic neighbor embedding (t-SNE) dimensional reduction techniques. Additionally, we examined differences between regionally representative prescribing patterns for major metropolitan areas.

Results: Distributions of prescribing volume and medication diversity were highly skewed among over 800,000 Medicare Part D providers, and medical specialties had characteristic prescribing patterns. Although the number of Medicare providers in each state was highly correlated with the number of Medicare Part D enrollees, some states were enriched for providers with >10,000 prescription claims annually. Hierarchical clustering and t-SNE dimension-reduction visualization of drug- or drug-class prescribing patterns revealed that providers cluster strongly based on specialty and sub-specialty, with large regional variations in prescribing patterns. Major metropolitan areas had distinct prescribing patterns that tended to group by major geographical divisions.

Conclusions: There is substantial prescribing variation among providers in Medicare Part D both between and within specialties. Large regional variations in prescribing patterns, particularly among major metropolitan areas, were also seen. Unsupervised clustering and t-SNE dimension-reduction are an effective means to examine variation in provider prescribing patterns, including substantial regional and medical specialty variation.