Purpose: We evaluate the cost-effectiveness of heroin assisted treatment (HAT) versus optimized MMT (standard practice) for chronic, treatment-refractory opioid addicts recruited into the North American Opiate Medication Initiative (NAOMI) randomized controlled trial.
Method: A societal perspective was considered in our analysis, which spanned the lifetime of a hypothetical cohort of patients eligible for HAT (two prior failed attempts at treatment). We developed a Semi-Markov cohort simulation model to determine the incremental cost per quality-adjusted life year gained of treatment with HAT as opposed to MMT. The model tracked hypothetical cohorts assigned to MMT or HAT from treatment entry through relapse, abstinence, re-entry and death in monthly cycles. The time to discontinuation of episodes of treatment, relapse and abstinence were estimated with Weibull regression models. Durations of successive episodes of treatment and relapse were adjusted based on prior results of proportional hazards frailty models. HIV seroconversion was modeled explicitly. Self-reported and linked administrative data collected alongside the NAOMI trial as well as population-level administrative data on all MMT patients in British Columbia from 1996-2007 was used. Costs of treatment, healthcare utilization and criminality were considered, as well as quality-adjusted life years (QALYs) in each model state. One-way and probabilistic sensitivity analyses were conducted.
Result: 251 patients, aged 39.7 years, with 38.5% female, were recruited into the NAOMI study. Participants had injected drugs an average of 16.5 years, used heroin 26.5 of the previous 30 days and had at least two prior attempts at MMT at baseline. Median estimated duration in the initial HAT episode was 24.7 months, and 10.0 months in MMT. Costs and QALYs outcomes were similar in the treatment states (HAT and MMT), but statistically significantly better than in the relapse state. The HAT cohort accumulated an average of 13.33 discounted QALYs over their lifetime versus 12.40 QALYs in the MMT cohort. Average annual costs were $20,355 for the HAT cohort and $20,152 for the MMT cohort. HAT had an ICER of $11,866 per QALY gained - below commonly-cited willingness to pay thresholds. The result was robust to sensitivity analyses assessing key assumptions on the structure of the simulation model.
Conclusion: HAT was found to be a cost-effective option for the treatment of chronic, treatment-refractory opioid addicts.
See more of: The 32nd Annual Meeting of the Society for Medical Decision Making