Factors associated with clinician treatment recommendations for patients with a new diagnosis of opioid use disorder.

ABSTRACT

BACKGROUND: This study examined factors associated with treatment recommendations for patients with a new diagnosis of opioid use disorder (OUD), comparing recommendations for patients with clear signs of OUD versus those with lower likelihood of OUD. METHODS: The study conducted a retrospective medical chart review in a randomly selected national sample of 520 Veteran Health Administration patients with a new opioid-related electronic health record (EHR) diagnosis from 2012 to 2017. The study categorized patients as having "high likelihood" or "lower likelihood of OUD" based on the presence or absence of clinician documentation in medical records of specific qualifying criteria (e.g., clinician documentation of patient meeting diagnostic criteria for OUD, etc). Analyses examined the association between baseline demographic and clinical characteristics with recommendations for medication and other treatments for OUD. RESULTS: Among patients with a new diagnosis of OUD, 28.7 % (n = 149) were recommended medication treatment, 52.5 % (n = 273) were recommended specialty substance use disorder (SUD) treatment, and 41.9 % (n = 218) were recommended treatment in non-SUD mental health settings. In adjusted models, high likelihood of OUD (AOR 8.31, 95 % CI 4.81-15.03) was strongly associated with the clinician recommending medications for OUD, while age 56-75 (compared to <35, AOR 0.36, 95 % CI 0.18-0.69), stimulant use disorder (AOR 0.28, 95 % CI 0.15-0.53), and rural residence (AOR 0.51, 95 % CI 0.30-0.85) were associated with lower likelihood of being recommended medication treatment. CONCLUSIONS: Differentiating among patients with EHR diagnoses of OUD to identify the subset with higher likelihood of underlying OUD is important to accurately understand OUD treatment rates and disparities. However, even among patients with a clear diagnosis of OUD, medication treatment is still recommended less often than other treatments, suggesting interventions are needed to encourage clinicians to prioritize medication treatment as a first-line treatment, especially for older, rural patients and those with polysubstance use.