Variations in Schedule III prescription patterns in a Medicaid population pre- and post-policy

Abstract

Abstract The present study investigated variations in patient movement patterns between prescribers before and after House Bill 1 (HB1) implementation in Kentucky using network abstractions (PPN: prescriber-prescriber networks) from a one-month cross-sectional Schedule III prescription data in a Medicaid population. Network characteristics such as degree centrality distribution of PPN was positively skewed and revealed Dental Practitioners to be the highly connected specialty with opioid analgesic hydrocodone-acetaminophen to be the most commonly prescribed drug. Taxonomy enrichment of the prescriber specialties in PPN using chi-square test revealed a reduction in the enriched taxonomies Post-HB1 compared to Pre-HB1 with Dental practitioners being constitutively enriched (p < 0.05). PPNs were also found to exhibit rich community structure revealing inherent clustering of prescribers as a result of patient movement, and were markedly different from those generated by random graph models. The magnitude of deviation from random graphs decreased Post-HB1 relative to Pre-HB1. The proposed network approach provides system-level insights into prescribers with potential to complement classical reductionist approaches and aggregate statistical measures used in assessing changes in prescription patterns pre- and post- policy implementation. It can provide preliminary cues into drug seeking behavior, and facilitate targeted surveillance of prescriber communities.