Before implementing the TADAM project in Belgium (a heroin-assisted treatment trial), our research team studied the trials in other countries. Since 1994, six randomised controlled trials have been developed using the same treatment model of heroin-assisted treatment (HAT). Each trial concluded that HAT had more efficacy than methadone treatment. We analysed those trials in order to find on which levels patients in a HAT treatment are expected to improve. Improvements appeared after at least six months on the level of street heroin use, (physical and mental) health and criminal behaviour. In the longer term, the continuation of treatment had positive but limited effects on the social level. Due to his higher cost, this treatment should remain a second-line treatment for this special target group: severe heroin addicts, using continuously street heroin in spite of a methadone treatment.