Predictors of future overdose among people who inject drugs in Baltimore, Maryland

Background: Longitudinal studies of future overdose risk among people who inject drugs (PWID) are needed to inform planning of targeted overdose preventions in the United States.

Methods: The Integrating Services to Improve Treatment and Engagement (INSITE) study followed 720 PWID between June 2018 and August 2019 to evaluate the delivery of mobilized healthcare services in Baltimore, Maryland. The present analyses used logistic regression to identify baseline characteristics predictive of non-fatal or fatal overdose during the 6-month follow-up among 507 participants with overdose information. Non-fatal overdoses were self-reported and fatal overdoses were identified through the National Death Index.

Results: At baseline, 121 (23 %) reported an overdose in the prior 6 months. Between baseline and follow-up, 66 (13 %) participants reported a non-fatal overdose and 6 (1 %) experienced a fatal overdose. Overdose during follow-up was positively associated with overdose in the 6 months prior to baseline (6.70 aOR; 95 % CI: 3.51, 12.78) and more than 6 months prior to baseline (2.49 aOR; 95 % CI: 1.52, 4.08) versus no prior overdose. Overdose during follow-up was also positively associated with buprenorphine treatment (2.37 aOR; CI: 1.08, 5.21) and negatively associated with non-prescribed methadone at baseline (0.59 aOR; 0.38, 0.93).

Conclusions: Identifying and intervening with PWID who experienced a recent overdose could reduce short-term elevated risk of future overdose. However, as other PWID reported never experiencing an overdose at baseline nonetheless experienced an overdose during follow-up, targeted approaches should be complemented with population-level interventions. Overdose risk implications of buprenorphine treatment and non-prescribed methadone are also discussed.