Pmh21 Cost-Effectiveness of the Predict Test: Results and Lessons Learned From a European Multinational Depression Trial

Abstract

In a US randomized clinical trial testing the effectiveness of preventing opioid relapse among individuals initiating extended-release naltrexone (XR-NTX) compared to buprenorphine-naloxone (BUP-NX) in an inpatient detoxification setting, the additional time required to detoxify from opioids prior to initiating XR-NTX resulted in fewer persons initiating XR-NTX, leading to a higher opioid relapse rate, and higher detoxification costs compared to BUP-NX. The objective of this study was to use trial data to estimate whether an efficient model of inpatient opioid detoxification would improve the economic value of XR-NTX compared to BUP-NX. Methods: We identified efficient models of inpatient detoxification for trial participants assigned to XR-NTX using 1) latent class analysis to identify detoxification pharmacotherapy use patterns, 2) a multivariable generalized structural equation model to explore determinants of XR-NTX initiation and detoxification duration, while controlling for endogeneity, and 3) data from the trial on detoxification daily cost by site. We then estimated trial cost-effectiveness outcomes from the healthcare sector perspective assuming alternative detoxification models for trial participants. Results: Five latent pharmacotherapy classes were identified and included in the multivariable model. Site effects were the largest determinants of both XRT-NTX initiation and detoxification duration. The predicted probabilities of successful initiation per detoxification-day ranged from 0.13-0.15 at the four sites with significant effects, and the predicted durations ranged from 5.5 to 6.9 days. The predicted cost/detoxificationday varied from $115-$348, largely due to different staffing models. We estimated that the most efficient site model would result in non-significant cost and effectiveness differences between initiating XR-NTX or BUP-NX. Conclusions: Adopting an efficient model of XR-NTX initiation could result in XR-NTX and BUP-NX having similar economic value from the healthcare sector perspective for the average patient requiring residential detoxification before initiating XR-NTX. The feasibility of implementing more efficient detoxification models needs to be explored.