Methods	Design: Randomized clinical trial and cost analysis
	Recruitment: from prisons and jails
	Setting: Residential drug treatment (RDT) facility contracted by the state and Los Angeles County correctional agencies, Los Angeles, CA, US
	Follow‐up: At baseline, 6 months, and 12 months
Participants	Population: Homeless men recently released from county jails with a history of drug use
	Eligibility criteria: (a) had a history of drug use prior to their latest incarceration, (b) were 18–60 years of age, (c) resided in one participating RDT program, and (d) were considered to be homeless prior to discharge from incarceration. Exclusion criteria included not speaking English and being judged to be cognitively impaired by the research staff.
	Sample size: Total n=600; PC‐NCM n=195, PC n=196, Usual care n=209.
	Baseline characteristics: the sample of 60 parolees reported a mean age of 40 (SD=10.4) and 11.5 years of education. The men were predominantly African American (46%) or Latino (33%) and nearly two thirds were never married; yet 62% reported having children. While all participants were screened as being homeless, 88% were living on the street/halfway houses or someone else's apartment 6 months prior to their most recent incarceration, while 12% were transitioning in residential drug treatment programmes or a prior incarceration within the 6‐month period. In terms of health, one‐third reported that they were of fair or poor health. More than two‐thirds (70%) reported having committed a violent crime. These participants were released from county jails (55%) and state prisons (45%), respectively. The vast majority of these participants (84%) had reported a history of lifetime stimulant use, and 85% had used marijuana. Close to 60% reported having more than one sexual partner in the 6 months immediately prior to their most recent incarceration. No program differences were found in any of the demographic variables.
Interventions	Intermediate peer coaching:
	Participants received weekly peer coaching interaction. The main foci of each session included building effective coping skills, personal assertiveness, self‐management, therapeutic nonviolent communication (NVC), and self‐esteem building. Further, the sessions were dedicated to avoidance of health‐risk behaviours, increasing access to medical and psychiatric treatment and improving compliance with medications, skill‐building, and personal empowerment. Discussions also centred on strategies to assist in seeking support and assistance from community agencies as parolees prepare for completion of the residential drug treatment program. Integrated throughout, skill building in communication and negotiation and issues of empowerment were highlighted.
	Usual care:
	For the usual care (UC) program, participants received all recovery and rehabilitation services available at the RDT site, including substance abuse services, assistance with independent living skills, job skills assistance, literacy, various counselling services, and discharge planning.
Outcomes	Housing stability: Percentage of patients in different housing settings using a structured questionnaire.
	Substance use: Alcohol and drug use was assessed using the modified version of the Texas Christian University (TCU) Drug History form.
	Employment: Percentage of patients with full‐time, part‐time employment and percentage of patients who are unemployed using a structured questionnaire.
	Cost/Cost‐effectiveness:
	The amount of cash spent on program activities was about the same for all three groups of participants: 32,583 for the PC‐NCM participants (M = $167.09; SD = $79.51), $33,375 for PC (M = $170.28; SD = $76.20), and $33,293 for UC (M = 159.30; SD = $76.61).
	The PC‐NCM group consumed the most staff time (more than half or 54% of the total recorded staff time), followed by PC group with about 44% of the staff time, while the UC group used the least staff time, with only 2.11% of the staff time.
	On an annualized basis, participants of the PC‐NCM group on average consumed $593.26; participants in the PC group on average consumed $488.92; and participants of the UC group consumed $59.92.
Notes