The Cost-Effectiveness of Financial Incentives to Achieve Heroin Abstinence in Individuals With Heroin Use Disorder Starting New Treatment Episodes: A Cluster Randomized Trial-Based Economic Evaluation.

OBJECTIVES: Cost-effectiveness analysis of two 12-week contingency management (CM) schedules targeting heroin abstinence or attendance at weekly keyworker appointments for opioid agonist treatment compared with treatment as usual (TAU). METHODS: A cost-effectiveness analysis was conducted alongside a cluster randomized trial of 552 patients from 34 clusters (drug treatment clinics) randomly allocated 1:1:1 to opioid agonist treatment plus weekly keyworker appointments with (1) CM targeted at heroin abstinence (CM abstinence), (2) CM targeted at on-time attendance at weekly appointments (CM attendance), or (3) no CM (TAU). The primary cost-effectiveness analysis at 24 weeks after randomization took a societal cost perspective with effects measured in heroin-negative urine samples. RESULTS: At 24 weeks, mean differences in weekly heroin-negative urine results compared with TAU were 0.252 (95% confidence interval [CI] -0.397 to 0.901) for CM abstinence and 0.089 (95% CI -0.223 to 0.402) for CM attendance. Mean differences in costs were £2562 (95% CI £32-£5092) for CM abstinence and £317 (95% CI -£882 to £1518) for CM attendance. Incremental cost-effectiveness ratios were £10 167 per additional heroin-free urine for CM abstinence and £3562 for CM attendance with low probabilities of cost-effectiveness of 3.5% and 36%, respectively. Results were sensitive to timing of follow-up for CM attendance, which dominated TAU (better outcomes, lower costs) at 12 weeks, with an 88.4% probability of being cost-effective. Probability of cost-effectiveness remained low for CM abstinence (8.6%). CONCLUSIONS: Financial incentives targeted toward heroin abstinence and treatment attendance were not cost-effective over the 24-week follow-up. Nevertheless, CM attendance was cost-effective over the treatment period (12 weeks), when participants were receiving keyworker appointments and incentives.

Using a pragmatically adapted, low-cost contingency management intervention to promote heroin abstinence in individuals undergoing treatment for heroin use disorder in UK drug services (PRAISE): a cluster randomised trial.

INTRODUCTION: Most individuals treated for heroin use disorder receive opioid agonist treatment (OAT)(methadone or buprenorphine). However, OAT is associated with high attrition and persistent, occasional heroin use. There is some evidence for the effectiveness of contingency management (CM), a behavioural intervention involving modest financial incentives, in encouraging drug abstinence when applied adjunctively with OAT. UK drug services have a minimal track record of applying CM and limited resources to implement it. We assessed a CM intervention pragmatically adapted for ease of implementation in UK drug services to promote heroin abstinence among individuals receiving OAT. DESIGN: Cluster randomised controlled trial. SETTING AND PARTICIPANTS: 552 adults with heroin use disorder (target 660) enrolled from 34 clusters (drug treatment clinics) in England between November 2012 and October 2015. INTERVENTIONS: Clusters were randomly allocated 1:1:1 to OAT plus 12× weekly appointments with: (1) CM targeted at opiate abstinence at appointments (CM Abstinence); (2) CM targeted at on-time attendance at appointments (CM Attendance); or (3) no CM (treatment as usual; TAU). Modifications included monitoring behaviour weekly and fixed incentives schedule. MEASUREMENTS: Primary outcome: heroin abstinence measured by heroin-free urines (weeks 9-12). SECONDARY OUTCOMES: heroin abstinence 12 weeks after discontinuation of CM (weeks 21-24); attendance; self-reported drug use, physical and mental health. RESULTS: CM Attendance was superior to TAU in encouraging heroin abstinence. Odds of a heroin-negative urine in weeks 9-12 was statistically significantly greater in CM Attendance compared with TAU (OR=2.1; 95% CI 1.1 to 3.9; p=0.030). CM Abstinence was not superior to TAU (OR=1.6; 95% CI 0.9 to 3.0; p=0.146) or CM Attendance (OR=1.3; 95% CI 0.7 to 2.4; p=0.438) (not statistically significant differences). Reductions in heroin use were not sustained at 21-24 weeks. No differences between groups in self-reported heroin use. CONCLUSIONS: A pragmatically adapted CM intervention for routine use in UK drug services was moderately effective in encouraging heroin abstinence compared with no CM only when targeted at attendance. CM targeted at abstinence was not effective. TRIAL REGISTRATION NUMBER: ISRCTN 01591254.