Mortality and morbidity related to suicidal behavior and opioid use disorder (OUD) have increased significantly over the past decade. These two public health crises are intertwined at multiple levels. Medications for OUD, especially buprenorphine, have the potential to decrease illicit opioid use and reduce the multiple negative consequences of OUD, including fatal and nonfatal overdose, criminal justice involvement, infectious complications, and misuse of other substances. In addition, several small randomized trials of buprenorphine treatment in treatment-resistant depression (with or without co-occurring OUD) suggest that buprenorphine may reduce depressive symptoms and suicidal ideation. No conceivable randomized trial, however, would be large enough to assess effects of buprenorphine on suicidal behavior. Consequently, we propose a large observational study to evaluate the effects of initiating buprenorphine treatment on subsequent suicidal behavior among people with OUD, including those with and without co- occurring mental health conditions or other known risk factors for suicidal behavior. We propose to assemble data from four MHRN health systems serving a combined member/patient population of approximately 11 million. This work will take advantage of methods successfully implemented in previous MHRN research and research by our HCSRN Addictions Research Network. Valid causal inference will require appropriate methods to account for confounding by indication (i.e. people with OUD treated with buprenorphine have higher or lower pre-existing risk of suicidal behavior than people not treated with buprenorphine). Choice of the optimal method will depend on patterns observed in actual data (e.g. rates and correlates of buprenorphine use), but we anticipate considering the following: ? Propensity-score matching using machine learning-derived propensity scores ? Propensity-score weighting using machine learning-derived propensity scores ? Disease risk score adjustment using machine learning-derived suicide risk prediction scores Using this data resource and these analytic methods, we will address the following specific questions: 1) Overall effect: Among people with recognized OUD, how does initiation of buprenorphine treatment affect risk of suicidal behavior over the following 90 days compared to risk among otherwise similar people with OUD not initiating buprenorphine treatment? 2) Heterogeneity of effect: Does any effect of initiating buprenorphine vary according to: a. Means or mechanism of suicidal behavior (opioid overdose vs. other overdose vs. self-injury)? b. Presence/absence of co-occurring mood disorder or severe mental illness? 3) Specificity of effect: Are effects observed for buprenorphine also observed for alternative medications for treatment of OUD (e.g. naltrexone)?
This large observational study will evaluate the effects of initiating buprenorphine treatment on subsequent suicidal behavior among people with opioid use disorder, including those with and without co-occurring mental health conditions or other known risk factors for suicidal behavior. We will use comprehensive health records data from four large health systems serving a combined member/patient population of approximately 11 million. Analyses will examine overall effect of buprenorphine treatment on subsequent suicide attempt or death, heterogeneity of effects in patient subgroups, and specificity of effects to buprenorphine vs other medications.