Understanding the outcomes of a trial of contingency management targeting abstinence from heroin: a process evaluation.
BACKGROUND
Contingency Management (CM) is a behavioural intervention which involves positive reinforcement (e.g. monetary vouchers, prizes) being offered to service users contingent on them achieving a pre-defined behavioural target consistent with treatment goals. Based on international evidence, UK clinical guidance recommends CM to target reduction in illicit drug use and clinic attendance. However, UK drug services have little experience of applying CM. The UK PRAISE trial examined the effectiveness of CM in promoting abstinence from street heroin among people in opiate agonist treatment. The trial compared two pragmatically adapted CM schedules (targeting attendance and abstinence respectively) against a control intervention (no CM). CM targeting attendance was moderately effective in encouraging abstinence compared with control only. However, CM targeted at abstinence was not effective. Neither schedule was associated with sustained abstinence. To investigate the mechanisms of action contributing to these outcomes we completed a qualitative process evaluation.
METHODS
The PRAISE trial used a cluster randomised design. Purposively sampling clinics (clusters) in each CM treatment arm, we completed qualitative focus groups seperately with practitioners and service users. Topic guides explored how the intervention was received, and what mechanisms were perceived to be associated with change (or the absence of it) in the primary (abstinence from street heroin) and secondary trial outcomes. All focus groups were audio recorded, professionally transcribed and subject to a thematic analysis.
RESULTS
We completed 11 service user and 8 clinician focus groups (+ 3 individual clinician interviews). Clinicians and service users in both arms were positive about CM and reported that it supported engagement and positive therapeutic relationships.
CM targeting abstinence reinforced provision of a 'clean' urine drug screen. Service users often reported reductions in use in response to CM. However, because such reductions were not reinforced, service users often experienced this negatively. In contrast, CM targeting attendance allowed clinicians providing keywork the flexibility to reinforce attendance, while being empathic and non-judgemental about continued use. Improved attendance enabled personal goals to be defined which promoted the service user's sense of agency in achieving change. The absence of sustained change was associated with abrupt termination of reinforcement but also withdrawal of regular keywork where services lacked the resources for this to continue.
CONCLUSIONS
By targeting attendance, CM promoted engagement with keywork, and helped establish strong therapeutic relations that promoted reductions in heroin use. CM directly targeting abstinence proved less helpful, in part because it failed to reward reductions in drug use. Implications for service configuration, CM implementation and the design of future trials will be discussed.
