Medications Used to Treat Opioid Use Disorder and the Prevention of Suicide: A Target Trial Emulation

There is an urgent clinical need to identify effective interventions to prevent suicidal behaviors among individuals at high suicide risk such as patients with opioid use disorders. Little is known about the potentially preventative role of Medical Treatments for Opioid Use Disorders (MOUD), such as buprenorphine, buprenorphine/naloxone, methadone, and naltrexone – largely because inclusion of suicidal individuals in randomized trials has traditionally entailed ethical difficulties and securing large enough sample sizes of such patients may not be feasible. Lack of evidence regarding use of prescribed MOUD among individuals with opioid use disorder has particularly important clinical implications because there is conflicting evidence as to whether MOUD initiation temporarily increases risk of suicidal ideation and behaviors. In the absence of randomized trials, we can use high-quality observational data to evaluate the benefits and risks of clinical interventions – using expert knowledge coupled with advanced data science methods for causal inference to emulate a (hypothetical) pragmatic randomized trial. This approach, formally known as Target Trial Emulation, can be considered the gold-standard for guiding decision-making if/when randomized trials cannot be implemented. Using administrative and clinical data from the US Department of Veterans Affairs, the largest integrated healthcare system in the US, we will apply cutting-edge causal inference methods to longitudinal data on diagnoses, prescriptions, clinical outcomes, disease and suicide risk severity, as well as sociodemographic characteristics. We will emulate two target trials examining (1) initiation vs no initiation of MOUD, and (2) the comparative effectiveness of pharmacological treatment strategies for individuals with opioid use disorders (i.e., head-to-head comparisons). In this set of target trials, all patients will receive all other psychosocial treatment components included as treatment as usual in the VA health system. The outcomes of interest will be a first episode of suicidal self-directed violence including (non-fatal) suicide attempts and deaths from suicide. This AFSP Early Career Researcher Innovation Grant would advance my early career as a physician-scientist by funding an innovative causal inference and data science framework to address a critical unanswered research question with major implications for clinical decision-making. This grant will provide the support for me to develop and adapt cutting-edge causal inference methods for psychiatric research, generate readily applicable findings for clinical practice, and provide the basis for future grant proposals.