Following the study protocol, all study subjects were selected from the local methadone maintenance treatment registration systems, according to a predefined set of inclusion and exclusion criteria. These criteria are described in Table 3.

From the inclusion criteria, the items 1a-1f refer to the severity and chronicity of the addiction problems, despite the patient's regular participation in a methadone maintenance program in which methadone has been prescribed in adequate dose levels. With regard to inclusion criterion 1a, the cut-off point of five years was chosen on the basis of a systematic review by Cramer and Schippers (1994) regarding the course and outcome of heroin addiction. They found that a major shift from drug use towards abstinence occurred in the first five years of the addiction career. Generally, after five years approximately 20% of the population of problematic drug users had reached abstinence, and this percentage did not change much in the following five to 10 years (see also: Hser et al., 2001). Regarding inclusion criterion 1b, it is important to note that a methadone dose of 80 milligrams plus or minus 20 milligrams per day has been recommended as an effective dose in the American State Methadone Maintenance Treatment Guidelines (Parrino, 1992). Therefore, a daily dose of 60 milligrams can be regarded as the minimum dose level for methadone maintenance to become effective and a minimum requirement for the decision regarding treatment resistance. For methadone maintenance patients who inhale their heroin, given their usually lower methadone dose level (Buster and Van Brussel, 1996) and the lower bioavailability of inhaled heroin compared to that of injected heroin, the methadone dose level required for entering the study was adapted to a minimum of 50 milligrams per day.
In inclusion criterion 1c, the "regular" attendance of the methadone program was operationalized in terms of at least 50 visits during the previous six months for patients in a five-day per week methadone program, and at least 30 visits for patients in a three-day methadone program.
In inclusion criterion 1e, daily or nearly daily use of illicit heroin was operationalized analogous to the definition used in the Addiction Severity Index (McLellan et al., 1980, 1992; Hendriks et al., 1989) for regular heroin use. To be eligible for entering the study, subjects had to use illicit heroin at least three days a week during the previous month.
With regard to the operationalization of inclusion criteria 1f, the physical condition of the patient was assessed by means of the Maudsley Addiction Profile Health Symptoms Scale (MAP-HSS; Marsden et al., 1998). For entering the study, the patient had to have a MAP-HSS total score of at least 8 (see paragraph 2.9.1). The psychiatric status of the patient was measured by means of the Symptoms Checklist (SCL-90; Derogatis, 1983; Arrindel and Ettema, 1986). Based on the (different) distributions of the SCL-90 in male and female populations, a cut-off score of at least 41 was used for male patients and at least 60 for female patients for entering the study (see paragraph 2.9.1). With respect to the domain of social functioning, the patient should have had at least six days in the previous month in which he was involved in criminal activities, and/or at least six days without personal contact (of at least half an hour duration a day) with a non-drug-using person, in order to be eligible for the study. This criterion of "at least six days in the previous month" in the social domain was based on the face value of a threshold of at least 20% of the maximum possible score on these items (i.e. 30 days).

With regard to the exclusion criteria, it is important to note that the items 2, 3, and 4 were purposely described in general terms, i.e. without explicit operationalization, because these items were intended to exclude only those patients who a priori would be very likely to jeopardize a successful conduct of the study. On the basis of information about the target population, no more than 2-3% of the recruited patients were expected to be excluded from participation in the study. With respect to exclusion criterion 4, the history of aggressive behavior should have been indicated by the patient having repeatedly been expelled from a methadone program prior to the start of the trial. In criterion 5, the pregnancy should have been confirmed by urine screening. Exclusion criterion 6 included those patients who - on the basis of sentences that still have to be executed - anticipated to be incarcerated for a duration of at least three months during the study period. Exclusion criterion 10 was specifically aimed at excluding patients from participating in the study, who still had a genuine possibility to terminate their heroin addiction. Lastly, in exclusion criterion 11, the maximum methadone dosage was based on the maximum dose levels allowed in most Dutch methadone programs, which generally range from 70-120 mg a day, but occasionally involve higher dosages.