Factors of Post-Prison Medication for Opioid Use Disorder Dispensation Among People Who Inject Drugs

Overview: Decision tree machine learning methodology was used to categorize injection drug use (IDU) and post-prison release medication for opioid use disorder (MOUD) dispensation. We report on predictive factors of IDU in the 12-months prior to individuals' most recent incarceration, and 2) classify pathways to MOUD dispensation 1-month post-release.Proposal text:

Background: The overdose epidemic continues to be one of the most pressing public health crises in US history, with approximately 93,000 overdose deaths in 2020 (Provisional Drug Overdose Data, 2020) and the prevalence of opioid use in the U.S. has increased since the early 2000s (Han et al., 2020). Individuals with opioid use disorder (OUD) are overrepresented in the criminal/legal system, as substance use is criminalized in the U.S. and it is also a contributing factor of recidivism (Victor, Comartin, Willis, & Kubiak, 2021). Most prison facilities are not equipped to address the reentry needs of detainees with OUD (Binswanger, 2013), and as a result, individuals with OUD who are released are at an elevated risk for overdose (Binswanger, 2013; Victor et al., 2021; Victor et al., 2021).

Methods: Data were obtained from a larger prison re-entry project between May 1, 2017, and April 30, 2020. Eligibility criteria required participants to be at least 18 years of age, incarcerated with a release date roughly 3 months from enrollment date, and released to one of the three eligible urban counties. Additionally, individuals must self-report opioid misuse at least once in their lifetime and mental health symptomology at the time of program referral. We used a subset of (Nf160) participants who reported either no lifetime opioid use by injection (non-IDU group; n=89) or opioid use by injection in the year prior to incarceration in their baseline assessment (IDU group; n=71). We applied a decision tree machine learning methodology to categorize PWID and post-release (i.e., 1-month) MOUD dispensation among PWID and those who did not.

Results: Approximately 46.25% of individuals reported IDU. Most of the sample was white (51.88%), male (52.50%), with an average age of 37.46 years old (SD = 9.76), and the majority (88.89%) reported receiving MOUD. The subgroup of participants most likely to report IDU were individuals with a Hep C diagnosis (53.53%) and a history of overdose (92.10%; terminal node 2). The subgroup of participants least likely to report IDU were individuals who were not diagnosed with Hep C (46.47%), identified as a person of color (21.84%), and never overdosed on opioids previously (26.51%; terminal node 3). The subgroup of participants that were most likely to report MOUD dispensation were individuals taking psychiatric medication (55.00%) who had a history of IDU (77.27%; terminal node 1).

Discussion: To protect vulnerable populations and to flatten the overdose mortality curve, an increased focus is required within criminal–legal systems to divert individuals with OUD and other behavioral health concerns away from the criminal legal systems and towards providing access to appropriate care. Further research is needed to demonstrate the efficacy of best practices for the treatment of OUD (e.g., naloxone distribution) to be used in prison settings and immediately upon re-entry.

Learning Objectives: This e-poster will present data on the impact of the overdose crisis on a vulnerable population. Implications for social work practice will be discussed, such as social workers' role in facilitating case management and care coordination during the reentry process.