PEPSA 2006.

Methods	Allocation: Permuted blocks of two, four, and six were used (not generated by a list, but ‘manually’ (see risk of bias documents). Patients were assigned to receive diacetylmorphine or methadone in a 50:50 ratio. Randomization was not stratified (information provided by the authors on request). Blindnessof patients and or care providers in respect to treatment: given the administration routes and different treatment schedules in each group, it was impossible to blind health care personnel to the treatment condition. Nevertheless, the professional who made the assessments and the statistical analysis was blind to the treatment condition. Duration: 12 months
Participants	Diagnosis: opiate dependency for more than 2 years in line with International Statistical Classification of Diseases, 10th Revision criteria; ongoing intravenous opioid habit; have been in MT in the past at least twice according to official certificates issued by authorized centers N= 62 Age= mean 37.2 (SD 5.5) Sex= male 90.3% History= opiate dependency for more than 2 years , resident in Granada over the preceding year. Criminal activity:number of days in the prior month M=9,8 (SD=12,2) (information provided by the authors on request) Mental State:mental health problems, and social maladjustment (according to scores of the assessment of severity by the interviewerQ in the social/family situation and legal Addiction Severity Index [ASI] subscales)
Interventions	Experimental group: injected DAM, twice a day, plus oral methadone, once a day. The average DAM dosage was 274.5 mg/day (range: 15–600 mg), and average methadone 42.6 mg/day (range: 18–124 mg). Control group: daily methadone (once a day) 105 mg/day (range:40–180 mg). Comprehensive clinical, psychological, social, and legal support was given to both groups. Psychosocial interventions: see above
Outcomes	General state of health, Quality of life, Severity of the addiction. Consumption of illegal opiate. Consumption of cocaine.  Consumption of other psychoactive substances, illegal or legal, not prescribed. Behavior that puts the patient at risk of contracting HIV and hepatitis C Psychological adjustment Symptoms of depression Symptoms of anxiety Family situation Social support Rate of retention.  Level of utilization of the psychosocial services of the trial
Notes	Country: Spain 2001‐2004 Website: http://www.easp.es/pepsa/inicio/ensayo_english.htm#Design
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	no information
Allocation concealment (selection bias)	Low risk	Treatment name ‘Metadona’ or ‘Heroína’ were introduced in identical opaque envelopes at a 50/50 ratio (i.e. six patients, 3 envelopes will say methadone, 3 will say heroin).  There were shuffled in the presence of the participants and each participant will pick an envelope. Our guys had ‘trust’ issues.  We early realized that a physician calling the center and tell the participant the treatment he or she randomly got, would generate problems.  We knew this from our contact with them, some of them where convinced we were going to ‘cheat’ in the randomisation.  Therefore, we had to show them that they had a fifty fifty chance to enter the heroin or methadone arm. (information provided by the authors)
Incomplete outcome data (attrition bias) All outcomes	Low risk	Intention to treat and per protocol analysis were performed (completers in the experimental group: 23/27( 85.1%), and in the control group: 21/23 (91.3%).
Selective reporting (reporting bias)	Low risk	Protocol available and study registered, details on table 3.
Blinding (objective outcomes: drop out, use of substances measured by urine analysis)	Unclear risk	"Independent interviewers of the research and clinical teams were responsible for applying the assessment instruments".
Blinding (subjective outcomes: use of substances as measured by self report, side effects)	Unclear risk	Our outcomes were based on the ASI, administered by independent interviewers.  This instrument is based on self‐report. We tested the use of street heroin with the acetylcodeine test.  It did not work; 95% of the test came back ‘negative’ regardless of the allocation group and self‐reported use of street heroin.