Allocative efficiency analysis of strategies to reduce overdose deaths among people with opioid use disorder and history of incarceration in Connecticut.

Abstract

Background: While Connecticut has successfully slowed overdose death (ODD) rates, additional progress is necessary. We examined policies that allocate resources with maximal efficiency to reduce ODDs.

Methods: We developed a mechanistic simulation of overdose policies, including medications for opioid use disorder (OUD) among people involved in the criminal justice system both during incarceration (MOUD-INC) and post-release in the community (MOUD-COM), and naloxone in the community (NLX) to determine how maximally scaling all permutations compared to current MOUD and NLX levels (i.e., status quo) would impact five-year cohort ODDs, discounted life-years (LYs), discounted quality-adjusted life-years (QALYs), and incremental cost-effectiveness ratios (ICER). A simulated cohort of 2748 people with OUD incarcerated in Connecticut moved between settings and OUD statuses. Costs were assessed in 2021 $US, employed health and public health sector perspective in base-case analyses and limited-societal perspective in sensitivity analyses, 3% discount rate, a cost-effectiveness criterion of < $100,000 per QALY gained. Analyses were performed over life-year and lifetime horizons.

Results: At status quo, the simulated cohort experienced 345 five-year ODDs, 16.1 per-person discounted LYs, and 12.5 per-person discounted QALYs. Individually, maximally scaling MOUD-INC reduced five-year ODDs 7%, added 0.3 LYs and 0.3 QALYs at an ICER $78,000/QALY gained; MOUD-COM reduced five-year ODDs 23%, added 1.2 LYs and 1.2 QALYs at $18,000/QALY gained; NLX reduced five-year ODDs 18%, added 0.4 LYs and 0.3 QALYs at $15,000/QALY gained. Considering all permutations, compared to status quo and each other, not all met the cost-effectiveness criterion. Maximally scaling MOUD-COM and NLX together was the most beneficial option meeting cost-effectiveness criterion, reducing five-year ODDs 39%, adding 1.5 LYs and 1.5 QALYs versus baseline at $18,000/QALY gained compared to the next-best option. Maximally scaling all had similar effects but an unfavorable ICER. In sensitivity analyses using a limited societal perspective, all options were cost-saving, and maximally scaling all three interventions was most beneficial and most cost-effective, adding 1.6 LYs, 1.5 QALYs, and averting 41% of ODD while saving society $363,000.

Conclusion: Maximally scaling community MOUD and naloxone can reduce ODDs among people who are incarcerated by 39%. Considering societal costs, maximally scaling all three decreases ODDs while saving money.