Step 1. Registration system
In the first step of the selection procedure, the available information from the local methadone maintenance treatment registration data bases was analyzed in order to define the source population, which served as the basis for subsequent steps of the recruitment process. The source population amounted to approximately 3,000 methadone patients. These potential candidates were at least 25 years old and were registered in the city area of the local treatment site for at least three years. In addition, they were heroin dependent for at least five years, had been prescribed an effective methadone dose for at least four consecutive weeks in the past five years and had had regular contact with the methadone maintenance program in the preceding six months. All potential candidates in the source population received a randomly assigned patient identifier number from the independent monitoring organization.

Step 2. Initial meeting
In each local treatment site, potential candidates were invited by the local study co-ordinator in the order of the patient identifier number for an initial meeting with the physician. Given the closing date of the randomization and the low prevalence of injecting in the remaining source population, approximately 1,000 potential candidates with a high patient identifier number were not invited anymore for the initial meeting of the selection procedure. Approximately 2,000 methadone clients from the source population were invited for this initial meeting. Around 500 of the invited candidates were never seen at the initial meeting. The major reasons for not attending the initial meeting were that potential candidates did not use heroin on a (near) daily basis, had left the methadone (maintenance) program, were incarcerated or hospitalized, had left the city or region, or had died. Almost half of all the methadone clients seen in the initial meeting ( 1,500 of the 2,000 candidates invited), did not meet the inclusion criteria or were barred by one of the exclusion criteria. Of the candidates that were excluded at this phase of the selection process, 45% did not meet the criterion of (near) daily heroin use. After this first initial meeting, 853 candidates were eligible for the next step in the selection procedure, the initial screening by the local research team and the physician.

Step 3. Initial screening
All 853 candidates that passed the screening at the initial meeting were assessed by the local research team. Eighty-two potential candidates did not meet one or more of the inclusion criteria of poor functioning in the area of physical health, mental health or social functioning that were assessed in the interview by the research team or decided themselves not to continue the selection process and were not screened by the local physician. From the remaining 771 candidates, 254 were eligible for the trial on injectable heroin and 517 for the trial on inhalable heroin. In the remaining of this chapter the focus will be on the 517 candidates that were eligible for further screening in the inhalable heroin trial (see Figure 18).

Figure 18. Flow diagram of selection process of patients in the trial on inhalable heroin

Of the 517 candidates for whom the initial screening was completed, the majority met the criteria for inclusion and was not barred by any of the exclusion criteria (88.4%). The most common reason for not including candidates at this step of the selection procedure was that candidates did not meet at least one of the inclusion thresholds indicative of poor functioning in the area of physical health, mental health or social functioning. Thus, 457 patients passed the initial screening and entered the qualification phase at the end of which they were invited for the second and final screening.

Step 4. Final screening
Four to eight weeks after the initial screening candidates were invited for their final screening. Thirty potential patients did not show up for their final screening and nine did not complete the final screening. Of the 418 candidates with a completed final screening the majority was eligible for randomization in the trial (85.3%). The major reason for not being included, was again related to not meeting the thresholds of poor functioning.
Three-hundred-and-ninety candidates were eligible for the trial investigating the effectiveness of inhalable heroin. However, 15 eligible candidates were excluded from the intention-to-treat population. At the beginning of the study, four patients were transferred from the control group (A) to the experimental group (B) in order to have a sufficient number of patients in the treatment site to be treated with co-prescribed heroin. These four patients were therefore excluded from the intention-to-treat population (CCBH, 1999b). Eleven patients were not randomized into the study because they were partner of candidates that were randomized into the trial already. Instead, they were offered to participate in the same type of treatment condition as their randomized partner (see also chapter 4). Thus, the intention-to-treat population for the inhalable trial consisted of 375 inhaling heroin dependent patients meeting all criteria for inclusion in the trial.

Step 5. Randomization and ITT-sample
In Table 15, an overview is given of the randomization of all 375 patients in the intention-to-treat population, divided by treatment site. One-hundred-and-thirty-nine inhaling heroin addicts were randomized in the control group (A), receiving an offer of 12 months treatment with methadone alone, 117 heroin addicts were randomized in the experimental group (B) receiving an offer of 12 months treatment with inhalable heroin plus oral methadone, and 119 patients were randomized into the group (C) receiving an offer of an initial period of six months treatment with methadone alone followed by six months treatment with inhalable heroin plus oral methadone.

Table 15. Intention-to-treat population and number of randomized patients per site (inhalable heroin)