Table 2.

Characteristics of Included RCTs

Study	Population	Sample Size (n)	Setting & Country	Intervention	Control	Outcome Measures and Follow-Up Time Intervals (Follow up rates)	Conclusions of the paper
Baer 2007	Youth, ages 13 to 19, vulnerably housed Mean age 17.9 Males 56%- Females 44% Ethnicity was reported as 58% Caucasian, 19% multiracial, 9% Native American, 8% African American, 4% Hispanic or Latino, and 2% Asian or Pacific Islander.	n = 117	Community Drop in Center- USA	Brief Motivational Intervention; up to 4 sessions; Average session length was 17 mins for 1st and 35 mins for 2nd session (n = 66)	Service as Usual (n = 51)	1. Substance use 2. Service use 3. Counsellor ratings of engagement 4. Treatment exposure and satisfaction Measurements were conducted at 1 month (82.9%) and 3 months (76.1%) post intervention.	The purpose of this study was to build upon previous mixed findings. However, the Brief Motivational Intervention did not lead to any improved outcomes in youth compared to those in the treatment as usual group.
Bender 2015	Youth, ages 18–21, Mean age 19 homeless 68.9%- housed 31.1% Males 60.8%- Female 36.5%- Other 2.7% Ethnicity was reported as White 41.9%- Black 20.3%- Latino 5.4%- other 32.4%	n = 97	Youth homeless Shelter -USA	SAFE (Safety Awareness for Empowerment); 3 day group intervention of 6–8 youth; focus areas include mindfulness, skill-building (n = 56)	Shelter services as usual, which includes case management services (n = 41)	1. Mindfulness scores (total, observing, describing, acting with awareness, accepting without judgement) Measurements completed as posttest at the end the intervention. The control group youth completed the interview approximately 5 to 7 days after their baseline interview (F/U for control 90.2% and for intervention 94.9%)	The SAFE intervention led to a significant increase in mindfulness, defined as observation skills, compared to those receiving services as usual. This suggests that youth experiencing homelessness are likely to engage in mindfulness training in shelters.
Greeson 2015/ Courtney 2008	youth age 17; in out of home care- Males 41.15%- Females 58.85%- Ethnicity was reported as White 8.97%- Black 40.17%- Hispanic 43.38%- Other 7.48%	n = 482	Independent living programs for youth in foster care- USA	Life skills training course (LST); two 3 h sessions per week for 5 weeks at community college (n = 234)	Services as usual aimed at preparing youth at risk of aging out of foster care (n = 248)	1. Interventions and service use 2. Job preparedness and preparedness 3. Education and employment 4. Economic well-being 5. Housing 6. Delinquency 7. Pregnancy 8. Documentation and accounts 9. Social support Measured over three time points (baseline, 1 year (91%), and 2 years (88%))	The in-class life skills training course did not appear to be more effective than services as usual to improve social support and other reported outcomes in youth. More research is required to determine the types of youth based interventions that lead to an improvement of the desired outcomes.
Guo 2016 / Slesnick 2013a / Slesnick 2013b)	youth 12–17 years; met DSM-IV criteria for alcohol/drug abuse; Mean age 15.4 Males 47.5%- Female 52.5% Ethnicity reported as White non- Hispanic 25.7%- African American 65.9%- Hispanic 1.7%- Native American 1.1%- Asian American 0.6%- Other 5%	n = 179	Short term Crisis Center for Run-away adolescents- USA	1. Community Reinforcement Approach (CRA) - 12 sessions, operant conditioning approach to teach methods of addressing life problems without alcohol or drugs (n = 61) 2. Ecologically-based Family Therapy (EFT) - 12 sessions, works with youth and family to identify dysfunctional interactions and improve social interactions (n = 57) 3. Motivational Enhancement Therapy (MET) - 2 therapy sessions, improve intrinsic motivation to change drug and alcohol use (n = 61)	Three arm study, see interventions	1. Family cohesion and conflict 2. Internalizing behaviours and externalizing behaviours 3. Substance use Measurements were conducted at baseline and at 3, 6, 9, 12, 18, and 24 months post intervention (F/U rates ranged from 69 to 79% across 6 time points and did not differ statistically between groups)	Ecologically-based Family Therapy is more effective and has longer lasting effects on family dynamics compared to individual therapies. While there are many challenges in the implementation of family-based therapies, overcoming these barriers will lead to improved family outcomes in youth.
Hyun 2005	male youth; residence in shelter; mean age 15.5; without psychiatric disorders- 100% male- 0% female Ethnicity not specified	n = 27	Shelter for runaway and homeless youth-South Korea	Cognitive behavioural therapy - 8 sessions over 8 weeks (n = 14)	Service as usual, no cognitive behavioural therapy (n = 13)	1. Self-esteem and self-efficacy 2. Depression Measurement interval was a Pretest- posttest design (F/U rates 87.5% for experimental group and 81.2% for control group)	This study shows that CBT is a useful intervention to increase self-efficacy and decrease depression in youth compared to no treated. These results are in agreement with previous studies showing the effectiveness of CBT in youth to improve mental health outcomes.
Kozloff 2016	youth aged 18–24; homeless or vulnerably housed; with mental disorder Mean age 21.5 Gender reported as Non- male 39% Ethnicity was reported as White 38% -Ethnoracial 36%- Aboriginal 26%	n = 156	Participants were recruited from community agencies that serve homeless people, institutions, including health care facilities and prisons, and directly from the street- Canada	Housing first with Assertive Community Treatment or Intensive Case Management (n = 87)	Service as usual (n = 69)	1. Housing stability 2. Quality of life 3. Employment Measured at baseline, 6, 12,18 and 24 months post intervention (89.7% F/U for the intervention and 72.3% for the control)	Housing First significantly improved housing stability in homeless youth with mental illness compared to those in the treatment as usual group. This is an effective intervention to improve the stability of homeless youth and reduce the long-term negative outcomes of this population. However, since the intervention did not have a significant effect on the other measured outcomes, it is recommended to adjust the intervention to better meet the needs of youth.
Krabbenborg 2017	Homeless youth aged 17 to 26 Average age 20 68.1% male- 31.9 female Ethnicity reported as 51% had a Dutch background	n = 251	Shelters for Homeless young adults- Netherlands	Houvast: A strengths-based intervention focusing on improving quality-of-life of homeless youth (n = 134)	Services as usual, such as housing, social network, education and finances (n = 117)	1. Mental and physical health 2. Quality of life 3. Violence 4. Income security 5. Satisfaction with family relations 6. Substance use 7. Autonomy 8. Competence 9. Resilience Measured in two waves at baseline as youth enter shelter and as the youth existed the homeless shelter, between 27 and 238 days – mean 156 days post baseline. (F/U 77.6% for control and 80.3% for intervention group)	Both the strength-based intervention and care as usual improve outcomes of homeless youth. No significant differences were found between the two groups. This suggests that youth benefit from receiving care services in general.
Milburn 2012	Families with youth ages 12 to 17; vulnerably housed; no current abuse or neglect Mean age 15.6 Males 33.8%- Females 66.2% Ethnicity reported as White 11.3%- African American 20.5%- Hispanic 61.6%- Other Mixed 6.6%	n = 151	Community based organizations- USA	STRIVE: 5 weekly home-based sessions focused on family conflict resolution and problem solving (n = 68)	Standard care received from the agencies that referred them (n = 83)	1. Substance use 2. Delinquent behaviour 3. Risky sexual behaviours Measured at baseline, 3 (71%), 6 (58%), and 12 months (46%) post intervention	Youth receiving the STRIVE intervention had a significantly decreased number of sexual partners and decreased usage of substances, excluding marijuana, compared to those receiving standard care. Youth receiving the intervention may have increased their marijuana use to replace alcohol and hard drugs.
Peterson 2006	youth; 14–19 years; vulnerably housed; recent binge drinking episode without recent alcohol or drug treatment Mean age 17.4 Males 54.7%- Female 45.3% Ethnicity reported as Caucasian 72.3%- African American 3.2%- Native American 3.2% -Hispanic/Latino 3.2% mixed race 15.9%- Asian/Pacific Islander or other race less than 1%	n = 285	Street or community agencies -USA	Brief Motivational Intervention: 1 session lasting on average 30 mins, provide information about patterns and risks (n = 92)	2 control groups: 1. Assessment only (n = 99) 2. Assessment at follow up only (n = 94)	1. Alcohol and drug use 2. Stage of change for substance use Measured at baseline, 1 month (82%) and 3 months (80%) post intervention	The Brief Motivational Intervention led to a decrease in illicit drug use, apart from marijuana, after one month of follow-up compared to those in the control group. Other results of the study were inconclusive and future research should focus on how and when desired outcomes are achieved.
Slesnick 2016	youth; 14–24 years; recent alcohol use; homeless; did not receive drop in, mental health, substance use services in past 3 months Mean age 20.8 Male 53.2% Female46.8% Ethnicity reported as White, not of Hispanic origin 57.0% Other 43.0%	n = 79	Drop in Center and Shelter- USA	1. 6 months of strengths-based outreach approach linked with drop-in center (n = 40)	1. 6 months of strengths-based outreachapproach linked with crisis center (n = 39)	1. Contact with services 2. Alcohol use 3. Personal control/self-efficacy 4. Depressive symptoms 5. Health (physical and mental) Measured at baseline 3,6,9 months post intervention (3,6,9 month F/U rates were 87,87,90% for the shelter linkage and 88,90,93% for the drop-in linkage conditions, respectively)	This study showed that the drop-in center intervention was more effective to link youth to services and led to an overall increase in service usage than the crisis center intervention. Youth in both groups reported an improvement in mental health and substance use outcomes, with no significant difference between the groups. However, youth in the intervention group demonstrated a reduction in drinking to the point of intoxication.
Slesnick 2015	youth; 14–20 years, vulnerably housed; met DSM IV criteria for abuse or substance disorder Mean age 18.74 Males 52.59% Females 47.41% Ethnicity reported as White non Hispanic 19.6% African American 65.56% Hispanic 2.22% Native American 0.74% Asian American 0.37% Other 11.48%	n = 270	Drop in Center- USA	1. Community reinforcement approach provided through a drop-in center 2. Motivational enhancement technique – two 1 h sessions through a drop-in center 3. Case management - 12 1 h sessions through a drop-in center	Three arm study, see interventions	1. Substance use 2. Depressive symptoms 3. Internalizing and externalizing problems 4. Coping 5. Victimization during the last 3 months 6. Homelessness (12 months) Measured at baseline 3,6, and 12 months post intervention (F/U 58.1% for CRA, 88.4% for MET, and 63.7% for case management)	Youth receiving the community reinforcement approach had improved substance use outcomes compared to those in the other two groups. However, while youth in all three arms had an improvement in the other reported outcomes, there was no significant difference between groups.
Slesnick 2009	Youth; 12–17 years; primary alcohol problem; family reside within 60 miles of research site; parents must have agreed to the possibility of family therapy. Mean age 15.1 years males 45% females 55% Ethnicity reported as African American 5% Anglo 29% Hispanic 44% Native American 11% Other 11%	N = 119	Runaway shelters- USA	1. Home-based ecologically based family therapy (EBFT) (n = 37), 2. Office-based functional family therapy (FFT) (n = 40)	Service as usual case management through a drop-in centre (n = 42)	1. Substance use 2. Psychological functioning 3. Family functioning Measured at baseline, 3 (75%), 9 (76%),15 (76%) months follow up post intervention. There were no statistically significant differences between groups in attrition	Youth in all three groups showed improvement in substance use, psychological functioning and family functioning. Family therapy has a greater impact on decreasing days of substance use compared to service as usual. Mixed results were obtained in the comparison of home-based family therapy compared to office-based. Therefore, more research is necessary to identify the most effective context of family therapy.
Slesnick 2007	youth; 14–22 years; vulnerably housed; met DSM-IV criteria for Alcohol or other Psychoactive Substance Use Disorders Mean age 19.21 Males 66% Females 34% Ethnicity reported as Native American 13% Asian 1% African American 3%, Hispanic 30% Anglo 41%, and mixed ethnicity/race 12%	n = 180	Drop in Center- USA	Community Reinforcement Approach: 16 treatment sessions offered, average 6.8 per participant (n = 96)	Service as usual through the drop-in center. The center offered a place to rest, food, showers, clothing and case management (n = 84)	1. Substance use 2. Mental Health (Individual functioning, depression) 3. Social stability Measured at baseline and at 6 months post intervention. (F/U 84% for CRA, and 88% for control)	Youth who received the community reinforcement approach had statistically significant improvements in mental health and substance use outcomes compared to those receiving treatment as usual. While youth in the control group also demonstrated improvements in certain areas, the effects of the intervention were more significant and long-lasting since it aimed to improve the relationship between homeless youth and their environments.
Thompson 2017	youth; 17–22 years; engaged in unprotected sex or heavy drinking Mean age 19.3 Females 58.3% Ethnicity reported as Hispanic 47.5% African American 36.1% Other race/ethnicity 16.4%	n = 61	Crisis center- USA	Two session individual brief intervention (45–60 min): focused on changing alcohol and HIV risk behavior (n = 30)	Two session educational comparison (n = 31)	1. Alcohol use 2. HIV sexual risk behaviours 3. Alcohol related sexual risk 4. Readiness to change alcohol use 5. Readiness to change HIV sexual risk behaviors 6. HIV preventive knowledge Measured at baseline and at 1 month (87.1%) post intervention	The brief intervention did not improve alcohol use outcomes in youth compared to those in the educational comparison group. However, it did improve the willingness of youth to change their alcohol behaviour. Future research is necessary to demonstrate how to translate willingness to change behaviour to an actual change in behaviour.
Tucker 2017	youth 18–25 years Mean age 21.81 Male 73% Female 27% Ethnicity reported as 31% non-Hispanic white 31% African American 25% Hispanic 24% multiracial/other 21%	n = 200	Drop in Centers-USA	AWARE: 16 weekly 45-min sessions of group motivational interviewing (n = 100)	Service as usual which includes access to food, hygiene services, case management and other programs available at the drop-in center (n = 100)	1. Alcohol, marijuana and drug use 2. Sex related outcomes Measured at baseline and 3 months post interventions. (95% F/U for intervention and 86% for control)	Youth in the AWARE group had decreased frequency in alcohol use and unprotected sex compared to those in the treatment as usual group. While there was an improvement of willingness to reduce marijuana and other drug use, there were no improvements in the frequency of use. This may be because the intervention did not make specific references to marijuana or other drugs.