PDG72 DETERMINANTS OF AN EFFICIENT MODEL OF EXTENDED-RELEASE NALTREXONE INITIATION FOR THE TREATMENT OF OPIOID USE DISORDER

Abstract

A recent eight-site randomized effectiveness trial compared buprenorphine-naloxone (BUP-NX) to extended-release naltrexone (XR-NTX) to prevent opioid-use relapse. Participants were recruited from inpatient-detoxification or short-term residential treatment programs. Current treatment protocols require persons initiating XR-NTX, but not BUP-NX, to be fully detoxified from opioids. This detoxification requirement resulted in fewer persons initiating XR-NTX, higher detoxification costs, and a higher relapse rate. However, among those who successfully initiated pharmacotherapy, relapse rates were not significantly different between XR-NTX and BUP-NX. We used trial data to identify factors associated with an efficient model of residential/inpatient opioid detoxification; i.e., factors that simultaneously increase the likelihood of XR-NTX initiation, and minimize detoxification duration. Given the many potential combinations of supportive medications that can be used in the detoxification process, latent class analysis (Stata 15.1) was used to categorize participants randomized to XR-NTX. A multivariable generalized structural equation model was then estimated to explore determinants of XR-NTX initiation and detoxification duration, while controlling for the simultaneous influence that many factors have on both measures. Five latent pharmacotherapy classes were identified. Neither pharmacotherapy class, nor detoxification duration was a significant determinant of XR-NTX initiation. Site effects were the largest determinants of both initiation and duration, and both factors varied by site, as did efficiency. Across the four sites associated with a significantly higher likelihood of initiation, predicted probabilities of success/day spent in detoxification ranged from 0.12–0.16. The predicted cost/day of detoxification varied from $115–$348, largely due to different staffing models. The most efficient site relied on physician assistants or nurse practitioners, and counselors; had a predicted probability of success/day=0.16; and had a predicted cost/day=$115. Identifying and adopting efficient site models of XR-NTX initiation could improve its relative economic value, thereby increasing its attractiveness to payers for patients who prefer this treatment.