The Cost Effectiveness of Buprenorphine in a Primary Care Setting: A Randomised Controlled Trial

Introduction Buprenorphine offers an alternative to methadone in the treatment of heroin addiction, and has the advantage of allowing alternate day dosing. However the comparative effectiveness and cost effectiveness of buprenorphine in practice has not been extensively studied. This study is the first to examine the cost effectiveness of buprenorphine as maintenance treatment for heroin addiction in a primary care setting using economic and clinical data collected within a randomised trial. Study design and methods The study was a randomised open-label trial with data collection at baseline, 3, 6 and 12 months following treatment commencement. Patients were dosed buprenorphine or methadone at community pharmacies and the pharmacy of a specialist drug and alcohol centre. The final total sample size was 139. Patients were assessed and treated by their general medical practitioners according to clinical guidelines for methadone and buprenorphine. Individual treatment regimes were used and doses tailored to individuals. Participants were aged between 18 and 65, heroin dependent, and able to give informed consent. Two subgroups were identified before the analysis. Those who were currently on a methadone program n=57 (continuing therapy subgroup) were analysed separately from new treatment recipients n=82 (initial therapy subgroup). The trial took place in Melbourne in 1999/2000. The study took a broad societal perspective and included health, crime and personal costs. Data on resource use and outcomes were a combination of clinical records and self report at interview. The incremental costs per additional day free of heroin use and per QALY were chosen as the outcomes in the study. An analysis of uncertainty calculated the likelihood of net benefits for a range of acceptable money values of outcomes. All costs were in 1999 Australian dollars. Results There was a small difference in average quality of life of 0.03 QALYs over the 52 weeks of the trial that favoured methadone. In the initial therapy subgroup those randomised to buprenorphine had an average of 0.65 QALYs (95% CI 0.56 to 0.73) while those randomised to methadone had an average of 0.61 QALYs (95% CI 0.52 to 0.70). During the year of the trial the estimated mean number of heroin free days did not differ significantly between those randomised to buprenorphine, 222 (95% CI 194 to 250), or methadone, 225 (95% CI 91 to 266). The total economic cost during the year of the trial was $17,736 (95% CI-2,981 to 38,364) and $11,916 (95% CI $7,697 to $16,135) in those randomised to methadone or buprenorphine. If crime is excluded, the costs were $4,513 (95% CI 3,495 to 5,531) and $5,651 (95% CI 4,202 to 7,100) Discussion The trial found no significant differences in costs or outcomes between methadone and buprenorphine maintenance. Although some of the results suggest that methadone may have a cost advantage, it is difficult to infer from the trial data that offering buprenorphine as an alternative would have a significant effect on total costs or outcomes. The point estimates of costs and outcomes suggest that buprenorphine may have an advantage in those initiating therapy although the confidence intervals were wide and the likelihood of net benefits from substituting one treatment for another was close to 50%.