Milburn 2012.

Methods	Study name: STRIVE Study design (e.g. RCT, cluster RCT): cluster RCT at the level of the family Intervention arm(s): received the Support To Re‐Unite, Involve, and Value Each other (STRIVE) intervention Comparator arm(s): standard care, based on referring agency. If family was not referred, a referral was made based on each respective family's needs. Sample size calculation performed: N/S Subgroups prespecified: N/A; no subgroup analysis conducted Subgroup analyses: no Start date: March 2006 Duration of follow‐up: 12 months post intervention Number of follow‐ups: 3 Follow‐up time points: 3, 6, and 12 months post intervention ICC (if reported): not reported
Participants	Number of schools randomised: N/A Number of participants randomised (total and by arm): 68 to intervention, 83 to control Age (range or mean (SD)) or grade at the start: intervention (14.7 years, 1.3 SD), control (14.9 years, 1.5 SD), overall (14.8 years, 1.4 SD) Gender: baseline: intervention (male: 22.1%, female: 77.9%), control (male: 43.4%, female: 56.6%), overall (male: 33.8%, female: 66.2%) Ethnicity: baseline: intervention (Hispanic: 61.8%, white: 11.8%, African American: 17.6%, other: 8.8%); control (Hispanic: 61.4%, white: 10.8%, African American: 22.9%, other: 4.8%); overall (Hispanic: 61.6%, white: 11.3%, African American: 20.5%, other: 6.6%) SES: N/S Inclusion criteria: away from home for 2 or more nights in the past 6 months (but not more than 6 months), having the potential to return home Exclusion criteria: no current abuse/neglect, no active psychosis, no current substance intoxication
Interventions	Randomisation before or after baseline survey: after Duration of the intervention (excluding follow‐up): 5 weekly sessions, each lasting 1.5 to 2 hours (< 3 months' duration in total) Description of the intervention STRIVE Consisted of 5 sessions, each lasting 1.5 to 2 hours. All sessions re‐inforced the importance of establishing a positive family climate, resolving family issues via conflict resolution (to improve family functioning), recognising and managing feelings, increasing positive affirmations, and learning and practising problem‐solving skills. Learning is cumulative, and all sessions complement each other, with each progressing from the last and building upon previous content delivered. Brief description of the theoretical model: session content based upon cognitive‐behavioural theories, designed to improve families' problem‐solving and conflict resolution skills Description of the comparator: standard care by the agencies that referred them. For families not receiving any services, appropriate referrals were given based on their needs.
Outcomes	Primary outcomes: self‐reported alcohol use, substance use, risky sexual behaviour, and delinquent behaviour Secondary outcomes: N/S
Setting	Country: USA; State: California Setting: family (intervention took place 'usually' within a family's home, but this was decided upon by family members themselves) Focus: targeted; newly homeless youth and their families
Process measures	Process data reported: yes; all quantitative estimates Method (qualitative or quantitative): quantitative Description Acceptability of the intervention: 76% of families completed all 5 sessions; 16% attended 1 to 2 sessions, and only 1 family attended no sessions. The Working Alliance Inventory (range 12 to 84) measured satisfaction in both adolescents and parents/guardians. The mean adolescent score was 72.0 (12.5 SD), and the mean parent score was 76.4 (10.0 SD), indicating high satisfaction. Adherence to the intervention: An intervention manual was created to increase intervention fidelity. 98% of tasks were implemented with fidelity (according to a 13‐item bespoke scale assessing fidelity of session delivery.) Intensity of the intervention: intervention sessions typically lasted 1.5 to 2 hours and occurred once weekly.
Statistics	Sample size: 151 families Unit of randomisation: adolescent (and subsequent attached parent) Unit of analysis: individual Method to promote equivalence between groups: not reported Statistical models: random‐intercept regression models Baseline differences adjustment: analyses adjusted for gender, as this was not balanced between study groups Repeated measures methods in analysis: random intercept model accounted for repeated measures taken in each individual.
Notes	Equity: sexual orientation and demographic variables collected and reported for baseline participants in intervention and control groups and overall Funding: study funded by the National Institute of Mental Health (NIMH R01‐MH070322) Randomisation method, e.g. block, stratification, computer: computerised coin toss Clustering accounted for in sample size calculation (if relevant): N/A Cluster randomisation methods to account for clustering in analysis: N/A
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerised coin toss used to randomise participants
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and intervention team not blinded to study allocation; self‐reported outcome measures
Blinding of outcome assessment (detection bias) All outcomes	High risk	Self‐reported outcomes
Incomplete outcome data (attrition bias) All outcomes	High risk	Although baseline differences were non‐significant between those who did and did not complete follow‐up, overall rates of attrition were very high (up to 54% at 12‐month follow‐up).
Selective reporting (reporting bias)	Unclear risk	No published protocol
Other bias	Low risk	No other forms of bias evident