Objectives: To assess the incremental cost-effectiveness of drug addiction treatment programmes provided in the UK by the National Health Service and not-for-profit agencies in terms of crime-related outcomes. All costs and crime-related outcomes were implicitly evaluated relative to a 'no treatment' alternative.
Methods: Longitudinal observational data on a national sample of heroin addicts referred to addiction treatment services throughout England were re-analysed. Predictions from a Poisson random-effects model were used to estimate the incremental effectiveness and cost-effectiveness of treatment programmes. Interaction variables were used to assess whether the injecting of heroin on entry to treatment had an impact on cost-effectiveness.
Results: The findings rejected the null hypothesis that increasing time in treatment (and therefore treatment cost) has no mean crime prevention effect on clients referred for community-based methadone treatment, treatment delivered within specialist drug dependency units and residential rehabilitation programmes (P < 0.05). However, the size of the cost per unit of effect based on model predictions was sensitive to the exclusion of a small group of outlying observations. The interaction between client injecting status and time in treatment was found to be statistically significant (P < 0.05), with an estimated reduction in treatment cost-effectiveness across all treatment programmes for clients who reported injecting drugs at treatment intake.
Conclusions: Whilst the analyses did not include an evaluation of the effect of treatment programmes on client health and quality of life and stopped short of providing a social weighting for the predicted reduction in crimes, they do offer a useful starting point for establishing the cost-effectiveness of treating heroin addiction. The onus is on public decision-makers to decide whether the predicted reductions in crime are worth the opportunity costs of investing extra resources in a major expansion of treatment services.