RESUMEN
OBJECTIVES: Individuals in recovery from opioid use disorder (OUD) are vulnerable to the impacts of the COVID-19 pandemic. Recent findings suggest increased relapse risk and overdose linked to COVID-19-related stressors. We aimed to identify individual-level factors associated with COVID-19-related impacts on recovery. METHODS: This observational study (NCT04577144) enrolled 216 participants who previously partook in long-acting buprenorphine subcutaneous injection clinical trials (2015-2017) for OUD. Participants indicated how COVID-19 affected their recovery from substance use. A machine learning approach Classification and Regression Tree analysis examined the association of 28 variables with the impact of COVID-19 on recovery, including demographics, substance use, and psychosocial factors. Ten-fold cross-validation was used to minimize overfitting. RESULTS: Twenty-six percent of the sample reported that COVID-19 had made recovery somewhat or much harder. Past-month opioid use was higher among those who reported that recovery was harder compared with those who did not (51% vs 24%, respectively; P < 0.001). The final classification tree (overall accuracy, 80%) identified the Beck Depression Inventory (BDI-II) as the strongest independent risk factor associated with reporting COVID-19 impact. Individuals with a BDI-II score ≥10 had 6.45 times greater odds of negative impact (95% confidence interval, 3.29-13.30) relative to those who scored <10. Among individuals with higher BDI-II scores, less progress in managing substance use and treatment of OUD within the past 2 to 3 years were also associated with negative impacts. CONCLUSIONS: These findings underscore the importance of monitoring depressive symptoms and perceived progress in managing substance use among those in recovery from OUD, particularly during large-magnitude crises.