Objectives: To curb opioid-related adverse events, the Centers for Medicare and Medicaid Services (CMS) endorsed several quality measures of high-risk opioid use: 1)high-dose use, 2)multiple providers, and 3)concurrent benzodiazepine use. We examined geographic patterns of high-risk opioid use among disabled Medicare beneficiaries given their high prevalence of musculoskeletal disorders and opioid use.

Methods: In multi-year (2011-2018), cross-sectional analyses of disabled, non-cancer beneficiaries with >2 opioid prescriptions, we examined geographic variation in high-dose use (≥120 daily morphine milligram equivalents for ≥90 consecutive days), multiple providers (receiving opioid prescriptions from ≥4 prescribers and ≥4 pharmacies), and concurrent benzodiazepine use (≥30 cumulative days). Using multivariable logistic regression (marginal effects), we obtained annual rates of high-risk opioid use across 306 Dartmouth Atlas of Health Care hospital referral regions(HRRs), adjusting for relevant confounders. We examined trends using the Mann-Kendall test and conducted spatio-temporal analyses to identify 3 types of HRR hotspots: 1)persistent (hotspots for ≥90% of study period without any apparent trend), 2)consecutive (newly identified in the past 3 years), and 3)intensifying (≥90% of study period with increasing trends).

Results: In 2018, rates of high-dose use, multiple providers, and concurrent benzodiazepine use were 8.6%, 2.1%, and 31.5%, respectively. Multiple providers measure decreased significantly (5.2% to 2.1%, p<0.001) from 2011 to 2018, while we observed no significant trend (all p>0.05) in high-dose use(~9%) and concurrent benzodiazepine use(~32%). We identified 20%-30% of HRR as persistent, consecutive, and intensifying high-risk opioid use hotspots concentrated in: Rocky Mountain states/Alaska (high-dose; multiple providers); northeast (high-dose; concurrent benzodiazepine use); southeast (multiple providers; concurrent benzodiazepine use); and the Florida peninsula(all 3 measures). Hotspots of the multiple providers measure were often neighboring regions of high-dose use.

Conclusions: Among disabled Medicare beneficiaries, we identified high-risk opioid use hotspots in 20-30% of HRRs, which may aid in developing targeted regional-level interventions and policies.