In this paragraph the treatment participation of patients will be described. Patients were defined as 'treatment completer' if they were still participating in the intended treatment condition in month 12 of the study as evidenced by the drug accountability data. Thus, patients in the control group A were considered to be treatment completer if they still received oral methadone in month 12. Similarly, patients in the experimental group B were considered to be treatment completer if they still received medically co-prescribed heroin in month 12.

As is shown in Figure 8 all patients in the methadone group started their methadone treatment offer. This is because these patients were already taking part in a methadone maintenance treatment program at the time of randomization. In the 12th month after baseline, 15 patients were no longer receiving methadone maintenance treatment, i.e., 15.3% of the patients in the control condition had 'dropped out' the treatment regimen offered at the beginning of the trial. The remaining 84.7% of the control patients had completed treatment.
It can be seen from Figure 8 that seven patients (9.2%) never started the co-prescribed heroin treatment. In the 12th month after baseline, 14 more patients (18.4%) were no longer treated with co-prescribed heroin. For the majority of patients the termination of the heroin treatment was non-voluntary. They were expelled from the program (five patients), had become incarcerated (one patient), had died (one patient) or were forced to stop the treatment for medical reasons (one patient) or because of problems with a residence permit (one patient). The other five patients ended the treatment with co-prescribed heroin voluntary for various personal reasons. The remaining 55 patients in group B were still treated with co-prescribed heroin at the end of the study and hence were considered as 'treatment completer' (72.4%).
The difference in treatment completion between the 12 months methadone condition (84.7%) and the 12 months co-prescribed heroin condition (72.4%) is not statistically significant (Odds-Ratio=0.48; 95%-CI: 0.22-1.02; p=0.0575).

Figure 8. Disposition of randomized patients

Treatment completing patients in the control condition had been prescribed methadone on 336 days (median) during the experimental treatment year. The mean dosage of methadone prescribed to treatment completers in the control group was 71 mg (sd=23.6 mg).
The mean methadone dosage in the experimental group was 60 mg (sd=17.0 mg), which is significantly lower compared to the control group (t-value=3.12; df=135; p=0.002). In addition, treatment completing patients in the experimental group visited the heroin dispensing units on 352 days (median) and 2.1 times per day (median). The mean heroin dosage prescribed was 254 mg per visit (sd=62.5 mg) and 549 mg per day (sd=193 mg).

Patients in the trial on injectable heroin had the opportunity to switch to medically prescribed inhalable heroin (see paragraph 2.2.2). Of the 55 treatment completers in the trial on injectable heroin, 18 had switched from injectable heroin to inhalable heroin by the end of the experimental study period (32.7%). The patients who switched to inhalable heroin did not differ significantly from the patients who stayed on injectable heroin with respect to their frequency of visiting the heroin dispensing units (2.1 versus 2.2 visits per day, respectively) or the heroin dosage prescribed (575 mg versus 536 mg per day, respectively). Lastly, it should be emphasized that patients remained in the trial they were assigned to at the start of the trial, regardless possible switches in their route of heroin administration. Thus, patients in the injectable heroin trial who had switched to co-prescribed inhalable heroin, were still analyzed as participants in the injectable heroin trial.