Regional Differences in Procedural Care of Hemorrhoid Management in Medicare Patients; Unwarranted Variation in Clinical Practice for a Medical-First Problem.

INTRODUCTION: Hemorrhoid disease is very common problem in the Medicare population. Prior work has shown significant variation in county-level practices of hemorrhoidectomy; however, regional variation of rubber band ligation (RBL) has yet to be assessed. This is important as many different practitioners from different specialties can perform this procedure repeatedly in an office-based setting. We aim to evaluate the variation of RBL and hemorrhoidectomy over a 7-y period. METHODS: Using Medicare part B claims data, we identified all beneficiaries >65 y seen for hemorrhoid disease between 2006 and 2013. Current Procedural Terminology (CPT) codes were used to identify all events for hemorrhoidectomy (46083, 46250, 46255, 46257, 46260, and 46261) or RBL (46221) by hospital referral region (HRR). We determined HRR-level rates of hemorrhoidectomy and RBL per 1000 beneficiaries adjusted for age, sex, and race. We calculated annual coefficients of variation (SD × 100/mean) for hemorrhoidectomy and RBL. RESULTS: 1.2 to 1.3 million fee-for-service Medicare beneficiaries were seen annually for evaluation of hemorrhoid disease. Mean-adjusted annual rates for hemorrhoidectomy by HRRs varied from 4.34 to 63.03 per 1000 beneficiaries. Mean-adjusted rates of RBL by HRRs varied from 7.06 to 163 per 1000 beneficiaries. Annual procedural coefficients of variation over the study period were 41-48 (high) for hemorrhoidectomy and 69-74 (very high) for RBL. CONCLUSIONS: While continued high variation exists for hemorrhoidectomy, there is very high variation for RBL between HRRs in treating hemorrhoid disease among Medicare beneficiaries. There are substantial Medicare expenditures in this high-volume population that are likely unwarranted.