Approximately one third of the patients presenting to University College Hosptial (London) were considered suitable for the trial. All but 4 percent were followed up to 12 months, and some anecdotal information was available for the remainder. The heroin group were generally more likely to have continued regular use of heroin than the group who were refused heroin from the clinic. The methadone group tended to polarize into "very good" or "very bad" outcomes in terms of drug use and criminality. It seems that refusal of heroin to confirmed addicts is more therapeutic, in terms of discouraging continued drug use, but leaves a group heavily involved in drugs outside of clinical control.