Association for Public Policy Analysis & Management

Background: Opioid use disorders (OUD) are a U.S. public health crisis. Rates of OUD are sharply rising, and deaths from overdose have tripled since 2000. Yet, only one in five individuals with OUD receive treatment. Medication-assisted treatment (MAT) with buprenorphine and naltrexone is highly efficacious and can be delivered in primary care settings. Yet, these medications have been slow to diffuse into practice. While there is evidence of limited diffusion of buprenorphine and naltrexone treatment into publicly-funded programs, little is known about factors that influence MAT use among commercially-insured populations.

Methods: Using 2008-2015 claims data from a large national commercial insurer, we examined MAT use among individuals age 10-64 with an OUD diagnosis. We assigned patients to a tax identification number (TIN) based on the quantity and intensity of OUD service use and then estimated hierarchical logistic regression models of the likelihood of buprenorphine or naltrexone use among individuals with an OUD diagnosis and frailty models of the time until discontinuation of MAT medications after initiation. Variables included patient (e.g., age, sex, employee status, comorbid medical and mental health conditions, region), provider (e.g., number and specialty of clinicians in the TIN, academic medical center affiliation), and community-level (e.g., number of buprenorphine-waivered clinicians, provider capacity) factors.

Results: The number of person-years in each calendar year ranged from 9.7 million to 10.9 million. The proportion with an OUD diagnosis increased, from 0.15% in 2008 to 0.39% in 2015. The proportion of individuals with an OUD diagnosis who received any OUD medications remained low, with 14.0% in 2008, 20.9% in 2010, and 15.6% in 2015. Buprenorphine formulations accounted for the majority of use, although the proportion receiving naltrexone increased over time. There was considerable variation in medication use by region, with differences increasing over the period (e.g., in 2015, the proportion with medication use was 11.9% in the South, 13.2% in the West, 16.3% in the Midwest, and 22.4% in the Northeast). Results of the logistic and frailty models will be completed this summer.

Conclusions: While the proportion of individuals diagnosed with OUD more than doubled in this commercially-insured population, the proportion of diagnosed individuals who received MAT remained relatively flat—no more than one in five, although there was considerable variation across regions. While the federal parity law has helped address barriers due to insurance restrictions, concerns have been raised about prior authorization requirements for OUD medications.⁶ In an effort to address shortages in providers who deliver MAT, the Department of Health and Human Services recently increased federal limits on the number of patients who can receive buprenorphine per qualified physician, from a maximum of 100 per prescriber to 275. Given the lethality of OUD, it is clear that additional health systems efforts must be made to increase MAT utilization. An understanding of patient, provider, and community level factors that influence diffusion will inform future policy efforts to expand access and use of these evidence-based treatments.