A Review of Interventions for Substance Use Among Homeless Youth

Abstract

This review summarizes existing evidence on interventions to address substance use problems among homeless youth. Fifteen studies met the inclusion criteria of the review. The quality of included studies was analyzed using explicit and well-validated criteria. Interventions evaluated ranged from individual-focused therapies such as brief motivational intervention, community reinforcement approach, and knowledge and skills training, to broader interventions such as family therapy, support groups, and shelter-based health care and housing programs. Participants reported improvements in substance use outcomes over time in most of the studies. However, the superiority of a specific intervention is difficult to determine because of the heterogeneity of the interventions and the few studies conducted on each intervention. Implications for practice and research are discussed.

Keywords

homeless youth substance use interventions

Introduction

High prevalence of substance use, including the use of alcohol and other drugs, is a well-documented problem and a recurring concern for helping professionals working with homeless youth. The term homeless youth as used here refers to unaccompanied youth as it is defined in the McKinney–Vento Homeless Assistance Act—youth who lack a fixed, regular, and adequate residence and are not in the physical custody of a parent or guardian. Estimates of homeless youth in the United States range from 1.6 to 3 million (Sedlak, Finkelhor, Hammer, & Schultz, 2002; Whitbeck, 2009). The current review includes youth between the ages of 12 and 24 who fit the McKinney–Vento definition to allow review of the full array of relevant literature.

Substance use is both a precipitating factor and a consequence of homelessness (Zerger, 2002). High prevalence of substance use among homeless youth has been documented in national and regional representative surveys (Friedman et al., 2009; Greene, Ennett, & Ringwalt, 1997; O’Toole et al., 2004; Van Leeuwen et al., 2004). Greene Ennett, and Ringwalt (1997) analyzed data from three national representative surveys and found that 75% of the street youth reported marijuana use; one third used hallucinogens, stimulants, and analgesics; 25% used crack cocaine, other cocaine, inhalants, and sedatives; and 17% engaged in injection drug use.

Homeless youth with substance use problems often suffer from other mental health problems. In a longitudinal study of comorbidities of substance use among homeless youth in eight cities, Johnson, Whitbeck, and Hoyt (2005) found that 93% of the youth in their sample who met substance abuse criteria also reported one other mental disorder, and half of their sample reported two or more mental disorders. In an intervention study, Slesnick and Prestopnik (2005) found that approximately 75% of the youth met Diagnostic and Statistical Manual of Mental Disorders (4th ed.) criteria for one or more Axis I diagnoses other than substance abuse. Research has also shown that homeless youth with drug addiction are more likely to engage in high-risk sexual practices (Friedman et al., 2009) and delinquent behaviors such as theft, property crimes, and drug trafficking in order to finance their addictions (Farabee, Shen, Hser, Grella, & Anglin, 2001).

To effectively address the high prevalence of substance use among homeless youth, an evidence-based approach is required to identify interventions that result in demonstrable health benefits for substance abusing homeless youth. Nevertheless, to the author’s knowledge, only two comprehensive reviews of interventions for homeless youth exist to date (Altena, Brilleslijper-Kater, & Wolf, 2010; Slesnick, Dashora, Letcher, Erdem, & Serovich, 2009) and neither focuses on interventions for substance use. The primary goal of this review, therefore, is to summarize existing evidence on interventions for substance use among homeless youth. This information will help guide social service providers and agencies as they seek to identify effective treatments to assist this population with their substance use problems. Recognizing the myriad methodological complications associated with recruiting homeless youth to intervention studies (Zerger, Strehlow, & Gundlapalli, 2008), this review uses explicit and well-validated criteria to evaluate the quality of studies. Secondary goals of the review are to draw implications for practice, to provide a critical appraisal of the methodologies in existing literature, and to suggest avenues for future research.

Method

Data Sources

Multiple databases including PubMed, MEDLINE, PsychINFO, ERIC, Social Work Abstracts, Social Service Abstracts, Sociological Abstracts, and the Cochrane Central Register of Controlled Trials were searched from their inception through April 2012. Combinations of the following keywords were used: drug, substance, alcohol, homeless youth, homeless adolescents, street youth, runaways, intervention, treatment, programs, and services. The title and abstract of each article were reviewed to determine whether the publication should be included based on predetermined criteria. The bibliographies of relevant reviews and publications meeting final selection criteria were also searched to identify additional articles. A total of 1,829 articles were identified.

Study Selection

Studies were included if they examined the effectiveness of an intervention to improve substance use problems among homeless youth between the ages of 12 and 24. Interventions were broadly defined to include services provided by primary care providers and to which homeless youth could be referred. Studies had to report data on substance use outcomes. Studies that only examined substance use knowledge and attitude without data on actual substance use outcomes were excluded. Studies that reported multiple health and behavior outcomes were included only if they reported specific substance use outcomes. Given the challenges of adopting rigorous research designs in intervention studies with homeless youth, a wide range of study designs were accepted in this review, including randomized controlled trials (RCTs), prospective longitudinal studies with nonrandomized control groups or no control group, and longitudinal or retrospective studies that compared substance use outcomes among people receiving different types or intensities of treatments. Only articles published in English were included in this review. After reviewing the abstracts and titles, 1,692 were excluded and the remaining 137 were retrieved for a full-text review; 110 articles were further excluded for not meeting the selection criterion. The bibliographies of the retrieved articles and relevant reviews (Altena et al., 2010; Hwang, Tolomiczenko, Kouyoumdjian, & Garner, 2005; Slesnick, et al., 2009) were also searched. Results from articles that examined data from the same study were combined. A total of 15 unique studies (18 articles) met the inclusion criteria and were included in this review. The results of the search and study selection process are presented in Figure 1.

Figure 1.

Summary of study research selection process.

Critical Appraisal

The quality of the studies was assessed using the guidelines developed by the U.S. Preventive Services Task Force Work Group (see Harris et al., 2001, for more information). In a modification of the guidelines, blinded outcome assessment was excluded from the rating criteria because none of the selected studies utilized this assessment procedure. The detailed quality appraisal criteria used in this review is presented as follows (adapted from U.S. Preventive Services Task Force):

Studies were rated “good” if they met all of the following criteria (1) clear definition of interventions; (2) initial assembly of comparable groups, which refers to adequate randomization for RCTs and adjustment for potential confounders in the analysis for other types of design; (3) proper maintenance of comparable groups throughout the study, which means there was ignorable attrition or comparable attrition rates across study groups; (4) a follow-up rate of at least 80% at the end of study (this cutoff criterion was used in Hwang et al., 2005); (5) use of reliable and valid measurements; (6) all important outcomes were considered; and (7) for RCTs, intent-to-treat analysis was performed. Studies received a “fair” rating if they failed to meet one or more of the criteria listed in the “good” category, and had none of the fatal flaws noted in the “poor” category.

Studies were rated “poor” if they had any of the following fatal flaws (1) study groups were very different initially and/or the comparability across groups was not properly maintained throughout the study; (2) the measurements were unreliable or invalid; (3) key confounders were not controlled for in the analysis for studies other than RCTs; (4) follow-up rate was less than 50% at the end of the study (this cutoff criterion was used in Hwang et al., 2005); and (5) sample size was less than 50 per study group (this cutoff criterion was used in Hwang et al., 2005).

Results

The majority (n = 12) of the studies were conducted in the United States, and three were conducted in foreign countries (two in Canada and one in Honduras). Substance use was the primary target behavior in seven studies. In one study, substance use was the secondary outcome to facilitate HIV reduction. The remaining seven studies targeted multiple health outcomes. Of the 15 studies reviewed in this article, six (40%) were rated “fair” and the remaining nine (60%) were rated “poor”; no study received a “good” rating. The most common reasons for a study to receive a poor-quality rating were low retention rates, initial incomparable groups, and unaccounted for potential confounders. Rating results, reasons for fair or poor ratings, and detailed information on the reviewed studies are presented in Table 1.

Table 1.

Summary of Evaluations of Interventions for Substance Abusing Homeless Youth

Study and Design	Quality rating and reasons	Study Settings and Participants	Intervention and Follow-up	Substance use outcomes
Brief motivational intervention (BMI)
Peterson et al. (2006) RCT	Fair Follow up < 80% at end of study	Study settings: field-site offices in Seattle. Treatment: n= 92; Control 1: n= 99; Control 2: n= 94. Age: 14-19 (M=17.4, SD=1.5). 55% Male. 72% White, 16% multiracial, 3% Native American, 3% Black, 3% Latino, 1% other.	Intervention: 1 session BMI; Control 1: same interview assessment protocol as the intervention group, without the feedback; Control 2: assessment at 1-month follow-up only. Follow-up at 1 month (82%) and 3 months (80%), overall follow-up rate 74.4%.	Significant reduction in summed illicit drug use other than marijuana for treatment group compared to control groups at 1-month follow-up. This effect disappeared at 3-month follow-up. No significant benefits for the treatment group on alcohol and marijuana use compared to the control groups. Intervention effects did not differ by age, gender, or ethnicity.
Baer et al. (2007) RCT	Fair Follow up < 80% at end of study	Study settings: field-site offices in Seattle. Treatment: n=75; Control: n=52. Age: 14-19 (M=17.9, SD=1.2). 56% Male. 58% White, 19% multiracial, 9% Native American, 8% Black, 4% Latino, 2% other.	Intervention: 4 sessions BMI; Control: treatment-as-usual. Follow-up at 1 month (82.9%) and 3 months (follow-up rate not reported), overall follow-up rate 76.1%.	Participants in both groups reported overall reductions in alcohol use at 3- but not 1-month follow-up, and reductions in marijuana and other drug use at both follow-ups. No significant benefits for treatment participants compared to the control.
Community reinforcement approach (CRA)
Slesnick, Prestopnik, Meyers, et al. (2007) RCT	Fair Groups were initially different in number of other substance use diagnoses	Study settings: drop-in center Albuquerque, New Mexico. Treatment: n= 96; Control: n=84. Age: 14-22 (M=19.2, SD= 2.1). 60% Male. 41% White, 30% Latino, 13% Native American, 12% multiracial, 3% Black.	Intervention: 12-session CRA and 4-session HIV education/skill practice; Control: regular services provided by a drop-in center. Follow-up at 6 months (86%).	Participants in both groups had reductions in substance use over time including percent days of alcohol and drugs use, and number of categories of drug used. Treatment group had greater reduction in percent days of alcohol and drug use, but not number of categories of drug used, compared to those in the control. Intervention effects did not differ by age, gender, or ethnicity.
Slesnick, Kang, et al. (2008) Longitudinal study of outcomes with different types and intensities of treatment	Fair Follow-up < 80% at end of the study	Study settings: drop-in center in Albuquerque, NM. Recruited: n= 172. Age: 14-20 (M=20, SD not reported). 59% Male. 37% White, 31% Latino, 12% Native American, 12% multiracial, 8% Black.	Intervention: case management and community reinforcement approach (CRA). No control group. Follow-up at 6 months (73%) and 12 months (76%), overall follow-up rate 68%.	Results from random coefficient model suggested that on average, participants' alcohol and drug use significantly decreased at 12-month follow-up. However, time effect disappeared when other covariates were added in the model. Number of treatment sessions attended did not predict participants’ rate of change in alcohol or drug use across time.
Knowledge and skills training
Booth et al. (1999) Crossover design	Fair Groups were initially different in substance use; follow up < 80% at end of study	Study settings: drop-in center in US. Treatment: n= 72; Control: n=75. Age: range not reported (M=17.4, SD=1.5). 51% Male. 73% White, 12% Latino, 8% Black, 5% Native American, 3% other.	Intervention: 8-hour knowledge and skill training; Control: no training. Follow-up at 3 months (60%).	Participants did not report significant reduction in use of heroin/cocaine or number of drugs used.
Ferguson & Xie (2008) Prospective longitudinal study with non-randomized control group	Poor Sample size <50; follow up <80% at end of study; some confounders unaccounted for	Study settings: drop-in center in LA. Treatment: n=16; Control: n=12. Age: 18-24 (M=21, SD= 1.4). 71% Male. 39% Black, 21% Latino, 21% White, 14% multiracial, 4% Asian.	Intervention: social enterprise intervention providing mentoring, job training, and clinical services; Control: usual care services. Follow-up at 9 months (68%).	Treatment group reported an increase in number of days they had been drunk in the past 6 months, whereas those in the control reported a decrease. Both groups reported an increase in the number of days of drug use in the past 6 months with the treatment group reporting a significant greater increase compared to the control group.
Case management (CM)
Cauce et al. (1994, 1998) RCT	Poor Initial resemblance between groups not stated; follow-up <50% at end of study	Study settings: a multi-service center in downtown Seattle. Recruited: n=304, Treatment: n=78; Control: n=72 in the analysis. Age: 13-21 (M=17, SD not reported). 61% Male. 60% White, 16% Black, 6% Latino, 5% Native American.	Intervention: intensive case management; Control: regular case management. Follow-up at 3, 6, 9, and 12 months (specific follow-up rates not reported). 49% of the participants completed at least two follow-up assessments.	Participants reported significant reduction in substance use over time. No significant benefits for the treatment group compared to the control.
Souza et al. (2011) Longitudinal study of outcomes with different types and intensities of treatments	Poor Follow-up <80% at end of the study; key confounders not accounted for	Study settings: drop-in center in Tegucigalpa, Honduras. Recruited: n= 400. Age: <=25, (M=17.5, SD not reported). 62% Male.	Intervention: multidisciplinary CM Follow-up at 12, 24, and 36 months (specific follow-up rates not reported). 59% of the participants had baseline and at least one follow-up visits.	Cessation of substance use defined as more than 50% of follow-up visits without substance use. Survival analysis indicated probability of remaining on substance was .88 at 12-month and .82 at 24-month follow-up.
Peer support intervention
Stewart et al. (2009) Prospective longitudinal study without a control group	Poor Confounders not accounted for; follow-up <50% at the end of study	Study settings: two agencies serving homeless and at-risk youths and one community center in Alberta, Canada Recruited: n = 70. Age: 16-24 (M=19, SD=2.5). 54% Male. 60% Aboriginal, 27% White, 13% other.	Intervention: 20-week pilot peer support intervention. Follow-up at 12 weeks after the first intervention (41%), and at the end of intervention (20%).	Univariate analysis showed that 55% participants at mid-point and 29% at post-test reported decreased use of drugs and alcohol or complete cessation of both.
Family therapy
Slesnick, Bartle-Haring, et al., (2006); Slesnick & Prestopnik (2005, 2009) RCT	Fair Groups might not be comparable because of their different retention rates.	Study settings: runaway shelters in Albuquerque, NM. Treatment 1 (FFT): 40; Treatment 2 (EBFT): n= 102; Control: n=101. Age: 12-17 (M=15, SD = 1.4). 43% Male. 41% Latino, 34% White, 7% Native American, 6% Black, 12% other.	Intervention: 16-session family therapy (FT), one provided in the office (FFT) and one in the home (EBFT) Control: shelter services. Follow-up at 3 months (88%), 9 months (82%), and 15 months (81%) for the treatment groups; at 3 (78%), 9 (81%), & 15 (82%) months for the control group.	Participants reported significant reduction in percent days of overall drug and alcohol use, with the FT groups showing significant greater decrease compared to the control group. Participants' age and gender moderated the effects of FFT. Participants' history of sexual and physical abuse moderated the effects of EBFT.
Milburn et al. (2012) RCT	Poor Groups differed by gender initially; overall follow-up <50%	Study settings: site selected by the family (usually their home) in LA and San Bernardino. Treatment: n= 68; Control: n= 83. Age: 12-17 (M=14.8, SD= 1.4) 34% Male. 62% Latino, 21% Black, 11% White, 10% other.	Intervention: 5-session family-based intervention based on cognitive-behavioral theories; Control: regular care. Follow-up at 3 months (71%), 6 months (58%) and 12 months (46%).	Participants had overall reduction in alcohol and hard drug use over time, with the treatment group having greater reduction compared to the control group. However, treatment group reported an increase in marijuana use, whereas the control group reported a decrease.
Shelter services
Steele & O'Keefe (2001) Longitudinal study of outcomes with different intensities of treatment	Poor Key Confounders not accounted for; follow-up rate not reported	Study settings: a residential comprehensive care center in New Orleans. Recruited: n= 106. Age: 16-21 (Mean and SD not reported). 30% Male. Ethnic composition not reported.	Intervention: residential comprehensive health care services. Specific interventions related to drug abuse were screening and intensive counseling to modify high-risk behaviors. Follow-up at 9 months (follow-up rate not reported).	Percent of participants who had drug dependence decreased from 41% on admission to 3% during follow-up or at completion.
Rotheram-Borus et al. (2003) Prospective longitudinal studies with non-randomized control group	Poor This study meets all the criteria for “good” except for an overall follow-up rate <50%	Study settings: four shelters in New York City. Treatment n= 101; Control: n=86 after matching. Age: 11-18 (M=15.6, SD = 1.7). 51% Male. 59% Black, 26% Latino, 15% other.	Intervention: residential comprehensive care services including access to health resources, skill training; Control: regular services. Follow-up at 3 months (57%), 6 months (58%), 1 months (50%), 18 months (46%), and 24 months (70%).	Marijuana use and the number of drug used were significantly lower in the treatment group among female participants at 6-month and 12-month follow-up. No significant benefits of the treatment for male participants. No significant benefits of the treatment beyond 12 months.
Pollio et al. (2006) Longitudinal study of outcomes with different types of treatments	Poor Follow up <50% at end of study; potential confounders not accounted for	Study settings: youth crisis shelters in a Midwestern city. Recruited: n= 371. Age: range not reported (M=14.7, SD = 1.5). 39% Male. 73% White, 27% other.	Intervention: emergency & crisis shelter services. Follow-up at 6 weeks, 3 months and 6 months (25-30% attrition per time period).	Proportion of youth reporting substance use decreased significantly at all follow-ups compared to the baseline. There was an additional significant decrease from 6 weeks to 3 months, but then a significant increase from 3 months to 6 months, resulting in no significant differences between 6-week and 6-month follow-ups.
Supportive housing
Kisely et al. (2008) Retrospective study comparing outcomes of people using different types of services	Poor Groups initially not comparable on years of education; sample size <50; key confounders not accounted for.	Study settings: a youth program in Halifax, Canada. Treatment: n= 15; Control: n=30. Age: 16-25 (M=19.8, SD=2.52). 71% Male. Ethnic composition not reported.	Intervention: supportive housing with residential supervision, in-home counseling and support; Control: access to services but not housing. No follow-up as it was a retrospective study.	Univariate analysis showed that treatment participants had significantly lower level rates of current substance use compared to those in the control condition.

Brief Motivational Intervention

Brief Motivational Intervention (BMI) is a technique that aims to increase clients’ motivation to change. Utilizing strategies such as therapist empathy and reflective listening, the therapist involves the clients in assessing their behavior and consequences, and provides feedback and direct advice as appropriate (Miller & Rollnick, 2002). BMI is a low-threshold and low-demand intervention, and has shown promising results for reducing substance abuse among adults (Stotts, Schmitz, Rhoades, & Grabowski, 2001). Two RCTs evaluated the effectiveness of BMI in reducing substance use among homeless youth. In the earlier study (Peterson, Baer, Wells, Ginzler, & Garrett, 2006), the treatment group reported a greater decrease in summed illicit drug use (other than marijuana) compared to the control group at the 1-month follow-up despite the small effect size; but this effect disappeared at the 3-month follow-up. No significant benefits of BMI on alcohol and marijuana use were found at either follow-up. As an attempt to improve the results from the earlier study, Baer, Garrett, Beadnell, Wells, and Peterson (2007) conducted a modified BMI featuring more sessions with a different sample of homeless youth. Although study participants reported overall reductions in alcohol, marijuana, and other drug use, there were no significant benefits of receiving the BMI intervention for the treatment group compared to the control group. Considering the relative rigor of the design in both studies, the findings indicate that BMI may be limited in its effectiveness as outreach for substance abusing homeless youth.

Community Reinforcement Approach

Community reinforcement approach (CRA) is a comprehensive behavioral treatment approach for substance abuse problems. CRA is based on the belief that environmental contingencies can play a powerful role in encouraging or discouraging substance use. Therefore, it seeks to rearrange these contingencies such that sober behavior is more rewarding than drinking or using drugs (Meyers & Smith, 1995). Two studies evaluated the effectiveness of CRA and mixed findings were reported. The first study adopted a rigorous RCT design (Slesnick, Prestopnik, Meyers, & Glassman, 2007). The results showed an overall improvement on substance use outcomes for all participants and a significant greater improvement for the treatment group compared to the control group. The study results did not differ by age, gender, or ethnicity. The second study (Slesnick, Kang, Bonomi, & Prestopnik, 2008) examined the joint effectiveness of CRA and case management for homeless youth who requested the services through a drop-in center in Albuquerque, New Mexico. Results from a random coefficient model suggest that on average, participants’ alcohol and drug use significantly decreased at 12-month follow-up. However, the number of treatment sessions attended did not predict participants’ rate of change in alcohol or drug use across time. The findings should be interpreted with caution because of the relative weak design (i.e., no control group).

Knowledge and Skills Training

Two studies evaluated the efficacy of knowledge and skills training on substance use outcomes among homeless youth. One study (Booth, Zhang, & Kwiatkowski, 1999) utilized a crossover quasi-experiment design to examine an 8-hr knowledge and skills training based on a health belief and peer influence model in reducing drug and sex risk behaviors among runaways who were trained to act as peer educators. The training features role-playing, discussion, and practice of refusal skills when drugs were offered. Univariate analysis showed that participants in the treatment group decreased their use of heroin/cocaine by 7% while those in the control group reported no change. However, this change (i.e., 7% reduction) was not statistically significant. Similar results were found on the number of drugs used. Participants in the treatment group decreased the number of drugs used by 3% while those in the control group reported a slight increase of 1%. However, neither change was statistically significant. The authors argued that the changes, albeit statistically insignificant, might be substantially meaningful. The other study (Ferguson & Xie, 2008) pilot-tested the effectiveness of a social enterprise intervention (SEI) in improving the mental health and risk-behaviors among a small sample of homeless youth in a drop-in center in Los Angeles. The SEI model targets homeless youth aged from 18 to 24. It consists of vocational training and supportive mentorship with the aims to improve healthy behaviors, mental health status, service connections, and social support of street youth. While the treatment group experienced improvements in some mental health measures (e.g., depression and self-esteem), they reported a significant greater increase in the number of days they had been drunk and using drugs in the past 6 months compared to those in the control group. It was unclear why the treatment group had a greater increase in substance use.

Case Management

Two studies examined the effectiveness of case management for reducing substance use among homeless youth. In a RCT, Cauce et al. (1994, 1998) evaluated the relative efficacy of intensive mental health case management compared to regular case management delivered in a drop-in center in downtown Seattle. The intensive case management features lower caseloads, more supervision, flexible funds, and higher educational qualifications of the caseworkers. The participants reported significant reductions in substance use over time, but there were no significant benefits for the participants receiving intensive case management compared to those who received regular services. The other study (Souza, Porten, Nicholas, Grais, & the Medecins Sans Frontiere-Honduras team, 2011) evaluated multidisciplinary case management services delivered in a drop-in center in Tegucigalpa, Honduras. Survival analysis showed a lower probability of continuing substance use at the 1- and 2-year follow-ups. The study results should be interpreted with caution considering its low retention rate and weak design. Besides, the authors defined “stopped substance use” in the survival analysis if participants reported no substance use at more than 50% of their visits to the drop-center. The results could be easily changed if a different criterion was used.

Peer Support Intervention

One study (Stewart, Reutter, Letourneau, & Makwarimba, 2009) pilot-tested the effectiveness of a peer support intervention in enhancing healthy behaviors among homeless youth in Alberta, Canada. The support intervention was developed on the theory that social support from peers and professionals can improve homeless youth’s mental and physical health status, their coping skills, and abilities to deal with stressful events (Stewart et al., 2009). The intervention features group and dyad support delivered by both peer mentors who experienced homelessness themselves and professional mentors. The support groups met 3–4 hr per week for 5 months. Over half of the participants reported decreases in drug and alcohol use at 12 weeks after the first intervention, and 29% reported reduction in substance use or complete cessation of both at the end of the intervention. However, the findings should be interpreted with extra caution because the study quality was low as the retention rate was below 50% at end of the study and potential confounders were not accounted for.

Family Therapy

Two studies evaluated the effectiveness of family-based interventions in reducing risk behaviors among homeless youth. One study (Slesnick, Bartle-Haring, & Gangamma, 2006; Slesnick & Prestopnik, 2005, 2009) utilized a rigorous RCT design and compared the effectiveness of office-based functional family therapy (FFT) and home-based ecologically based family therapy (EBFT) to usual shelter services. The interventions consisted of communication and parenting skills training, behavioral, cognitive, and environmental interventions, with the aim to alter dysfunctional family patterns that contribute to substance use, running away, and related individual problem behaviors (Slesnick, et al., 2006). Up to 16 family sessions were provided with each session lasting about 60 min. Participants reported a significant reduction in percent days of overall drug and alcohol use, with the two treatment groups experiencing greater decrease compared to the control group. Subgroup analysis revealed that participants with both sexual and physical abuse did better with EBFT than with treatment as usual on the number of drugs used in the past 90 days, while youth with neither sexual nor physical abuse did not differ between modalities. Age and gender were found to moderate the effect of FFT, with significant reduction in percent of days of drug and alcohol use observed only for male and older (ages: 16–17) participants. The other study (Milburn et al., 2012) examined the effectiveness of a short family intervention in reducing risk behaviors and drug use among a sample of newly homeless youth. The intervention was specifically designed for newly homeless youth, and was based on the belief that running away from home was an ineffective attempt to resolve family conflict. Up to five sessions were provided, aiming to establish a positive family climate, improve family functioning, and help participants to learn problem-solving skills. Participants reported overall reductions in alcohol and hard drug use, and the treatment group reported greater reductions than the control group. However, marijuana use showed the opposite effect, with the treatment participants increasing times of use whereas those in the control group decreasing use.

Shelter Services

Two studies evaluated the effectiveness of residential comprehensive care services addressing substance use, sexually transmitted diseases, and other health issues among homeless youth. In one study (Steele & O’Keefe, 2001), participants who had positive urine drug screens received a 28-day in-house drug treatment and individual counseling, in addition to the general health services provided at a residential comprehensive care center. Univariate analysis showed that percent of drug abusing homeless youth dropped from 41% on admission to 3% during follow-up or at completion. Despite the large percentage decrease in drug use, the efficacy of the treatment should be taken with caution considering the low quality of the study. There was no control group; and the results were merely descriptive, thus leaving out many uncontrolled confounders. The other study (Rotheram-Borus et al., 2003) followed a sample of homeless youth receiving shelter services for 2 years and found significant benefits of the shelter services for female participants, but not for male participants. Female participants in the treatment group reported a significant decrease in marijuana use and the number of drugs used at the 6- and 12-month follow-ups compared those in the control condition.

Another study (Pollio, Thompson, Tobias, Reid, & Spitznagel, 2006) evaluated the effectiveness of emergency shelter and crisis services. Youth crisis shelters generally serve adolescents 12–18 years of age, with the priority of ensuring that youth are provided with the basic necessities. Alcohol/drug treatment was also available in these crisis shelters. Participants reported significant reductions in drug use at the 6-week, 3-, and 6-month follow-ups compared to the baseline. There was an additional significant reduction of substance use from 6-week to 3-month follow-ups, but this change reversed itself between 3- and 6-month follow-ups, resulting in no significant changes in substance use between 6-week and 6-month follow-ups. The findings indicate that benefits of emergency shelter may be limited in the long run.

Supportive Housing

One study (Kisely, et al., 2008) pilot-tested a supportive housing program, a combination of safe, affordable housing with integrated support services, in Halifax, Canada. Participants in the housing program had stable supervised accommodation for a minimum of 3 months and had access to regular services in a drop-in center, whereas those in the control group had the same access to drop-in center services but not the accommodation. Univariate analysis suggested that residents of supportive housing had significantly lower rates of current substance use than those in the control group. However, the study was retrospective and there was no baseline test on substance use. The positive association between the supportive housing and lower rate of substance use could easily be explained away by differences in rates of substance use at the baseline or other systematic differences between the two groups.

Discussion and Applications to Social Work

Evidence on the effectiveness of interventions for substance abusing homeless youth is emerging as more formal evaluations are conducted. This review revealed that a variety of interventions with homeless youth ranging from individual-focused therapies to shelter-based services have been examined. Common effects seemed evident as participants showed overall improvement in substance use outcomes in most of the studies. Differential effects were less common as participants receiving the treatment of interest were rarely found to have greater improvement than those who received regular services. Only family therapy appeared to show signs of relative efficacy.

This review has certain limitations. Only articles published in English were included. There might be relevant studies that either have never been published or published in other languages. Moreover, the working definition of homeless youth used in this review excluded homeless children under the age of 12 or those who are homeless with their families. Findings of the review might not apply to these homeless populations. Another limitation of the working definition is its broad age limit of homeless youth (i.e., age 12–24). Although a broad age limit was needed for a fuller review of evidence, it did not distinguish between the age groups of 12–18 and 18–24, an important determinant of service availability. Nevertheless, it seems that the participant’s age has a limited role in moderating the treatment effects based on several studies reviewed here (Peterson et al., 2006; Slesnick & Prestopnik, 2005; Slesnick et al., 2007).

Descriptive analysis of the sample characteristics of reviewed studies suggests that homeless youth are often troubled prior to becoming homeless. Participants often reported being physical and/or sexually abused (Slesnick & Prestopnik, 2005), attempting suicide (Booth et al., 1999), experiencing school disruptions or dropping out at an early age (Stewart et al., 2009), and having multiple runaway episodes (Pollio et al., 2006) prior to treatment. These findings indicate a need for prevention programs targeting household youth who are at high risk of becoming homeless.

For youth who are currently homeless, comprehensive substance use screening, assessment, and treatment should be make accessible to them given the high prevalence of substance use in this population (Greene et al., 1997). Overall, this review indicates that study participants receiving some type of services (either in the treatment group or in the control group) improved in substance use outcomes over time in most of the studies. When two different interventions were compared, however, differential effects were rarely found. These findings suggest that depending on the availability of resources and practice settings, practitioners could offer a variety of services to homeless youth who abuse substances.

Another important implication drawn from the current review is that more effort should be made to encourage homeless youth to engage and remain in therapy for their substance use and other mental health problems. A majority of the studies reviewed in this article had recruitment and retention issues (Cauce et al., 1998; Milburn et al., 2012; Pollio et al., 2006). The reluctance of homeless youth to engage in therapy has also been well documented in the literature (see Slesnick, Meyers, Meade, & Segelken, 2000, for a review of this literature). Some strategies to engage homeless youth in substance use treatment include using and understanding language from the youth’s culture, presenting the treatment in an appealing and nonthreatening way, emphasizing confidentiality, and identifying youths’ motivations to enter therapy (Slesnick, Meyers, et al., 2000).

Moreover, there is some emerging evidence suggesting moderation effects of certain youth characteristics on treatment outcomes. For example, Slesnick and Prestopnik (2005) found that homeless youth with a history of sexual and/or physical abuse benefited more from family therapy than those without any type of abuse. In addition, another study evaluating the effects of residential comprehensive care services found reductions in substance use among females but not males (Rotheram-Borus et al., 2003). If systematic differences in treatment effects among subgroups of homeless youth are consistently found in future research, service providers should explore intervention alternatives that are sensitive to client characteristics, such as life circumstances, gender, and culture.

Practitioners should note that this review was limited to studies that evaluated interventions to decrease substance use among homeless youth. Many studies have evaluated substance use interventions with household adolescents (see Weinberg, Rahdert, Colliver, & Glantz, 1998, for a review of this literature) or homeless adults (see Hwang et al., 2005, for a review of this literature), and some interventions appear to be effective for these populations. These interventions may also work well for homeless youth. Practitioners are advised to consult these studies and reviews for more information.

The review revealed several limitations in the current body of intervention research on substance use problems among homeless youth. Only 1% of the initial search results met the inclusion criteria. For each specific intervention, there were only two to three studies that examined its efficacy. These results indicate an insufficient number of studies that have evaluated treatment effectiveness among homeless youth who use substances. The paucity of empirical evidence will most likely impede the implementation of evidence-based substance use practice for homeless youth.

Of the 15 studies included, none appeared to have good quality, and the majority received a poor rating. The most common contributors to a poor rating for studies that adopted a design other than RCTs were the initial incomparability of study groups and failure to control for potential confounders caused by lack of randomization. When the study target is homeless youth, randomization is not always possible because of feasibility, ethical issues, and potential contaminations (Booth et al., 1999). As a result, several studies had incomparable study groups. Only one of the studies (Rotheram-Borus et al., 2003) utilized some statistical techniques (i.e., propensity score matching [PSM]) to address the initial incomparability between study groups.

It is also worth noting that three of the studies (Ferguson & Xie, 2008; Kisely et al., 2008; Stewart et al., 2009) were pilot work that tested the feasibility and effectiveness of SEI, peer-support intervention, and support housing respectively. Given the limited scope of pilot work, it is not surprising that all three received a poor rating. The poor ratings, however, do not necessarily translate into lack of effectiveness of those interventions. Nevertheless, more studies with larger scopes are needed to replicate the results from these pilot studies before they can be widely applied in practice settings.

The most significant methodological problem affecting the studies, including RCTs, is retention of participants. Two RCTs (Baer, Garrett, Beadnell, Wells, & Peterson, 2007; Peterson et al., 2006) were not rated “good” only because their retention rates were below 80%. In one study (Cauce et al., 1994, 1998), less than 50% of the participants initially recruited completed all the follow-up assessments. Retaining clients in substance use treatment is an important near-term outcome for service providers, and a potential mediator of subsequent client outcomes. Retaining homeless clients in interventions presents a particular challenge for substance use treatment providers. This is because homeless youth are difficult to engage in treatment initially, and consequently providing services can be more costly (Orwin, Garrison-Mogren, Jacobs, & Sonnefeld, 1999). Attrition may result in original comparable groups becoming incomparable at follow-up assessments, increasing the risk of selection bias that can greatly weaken internal validity. Many factors have been found to be associated with retention of homeless adults in longitudinal studies, including length of the treatment, nature and service components of the intervention, study site, clients’ perceptions of the usefulness of the treatment, client satisfactions, therapeutic relationships, and responsiveness and friendliness of the staff (Orwin et al., 1999; Padgett, Henwood, Abrams, & Davis, 2008). High drop-out rate might also be due to youth’s perceptions that the interventions are not effective.

Another issue that emerged from this review is the measurement of substance use outcomes and its implications for intervention studies. The most common measures of substance use in the studies reviewed here were the number of categories of drug used and percent days of overall alcohol and drug use in a specified period of time. A majority of the studies reported mixed findings regarding the effectiveness of the intervention when different measures of substance use outcomes were examined. In a study testing the effectiveness of CRA, for example, the treatment group had greater reduction than the control group in percent days of substance use but not number of drugs used (Slesnick et al, 2007). The findings also appeared to be different depending on the categories of drugs used. For instance, two studies (Milburn et al., 2012; Peterson et al., 2006) reported a significant treatment effect on the total number of drugs used (excluding marijuana); however, the effect did not extend to marijuana use. More studies are in need to replicate the findings. If differential effects on marijuana and other types of drugs are consistently found, efforts should be made to explore the underlying mechanisms; and new intervention strategies that are based on the types of drug used need to be developed.

Methodological problems set aside, studies rarely examined treatment experiences from the participants’ perspectives with a few exceptions (e.g., Stewart et al., 2009). The effectiveness of a treatment may be of statistical significance when measured by standard instruments, but it might have little substantive significance for substance abusing homeless youth. Participant’s perspectives may offer important insight on the substantive significance of the treatment. Seeking feedback from the participants may also facilitate the interpretations of unexpected findings. In one study (Ferguson & Xie, 2008), participants in the treatment group had more drunk and drug use episodes at follow-ups compared to the baseline. Exploring the participants’ experiences through interviews or focus groups at follow-ups may help us understand these disturbing findings. Future intervention studies should explore clients’ experiences. For example, in what ways are the treatments beneficial or harmful to them? Are they likely to maintain any treatment gains that they had made in stopping or reducing substance use? What part of the treatment could be enhanced to better meet their needs?

Finally, with a few exceptions (Peterson et al., 2006; Rotheram-Borus et al., 2003; Slesnick & Prestopnik, 2005; Slesnick et al., 2007), potential moderation effects of ethnicity, gender, age, and other youth characteristics on the treatment outcomes were rarely explored. Mixed findings were reported in studies that examined moderations effects. In two studies, participants’ outcomes did not differ by age, gender, or ethnicity (Peterson et al., 2006; Slesnick et al., 2007). However, being sexually and/or physically abused appeared to moderate the treatment effects in another study (Slesnick, et al., 2006; Slesnick & Prestopnik, 2005, 2009). These findings have yet to be replicated.

Corresponding to these limitations identified, several suggestions for future research are presented here. To augment effect size in RCT studies, researchers may consider using control groups that receive minimum usual care services or temporarily withholding services for the control groups until after the study is finished. Another way to improve study findings is to maintain comparable study groups and high retention rates throughout the study. When dropout occurs, follow-up data should be collected to explore the reasons for dropout and to allow intent-to-treat analysis to be performed. For studies conducted in naturalistic settings or secondary analysis of routinely collected data from agencies, efforts should be made to address the threat of selection bias by examining the comparability of study groups and describing the differences between groups (if any). Statistical techniques such as PSM could be utilized when appropriate. Potential cofounders should always be controlled in order to discern true treatment effect. Moreover, studies that explore systematic differences in treatment effects among subgroups of homeless youth are in great demand. Homeless youth is a diverse population. The effects of gender, ethnicity, age, and history of abuse have been explored in a few studies, but the findings are not conclusive. The potential impact of sexual orientation and duration of homelessness on treatment effects are rarely explored. More research is needed to examine these systematic differences in order to inform interventions that are sensitive to the specific characteristics of subgroups of homeless youth. This review also highlights the need for exploring participants’ experiences of the treatment, their satisfaction, and how their perceptions influence their treatment outcomes. A mixed method approach that embeds focus groups and/or interviews within the larger quantitative experiment, as it was adopted in Ferguson and Xie’s (2008) pilot study of SEI, may be utilized to help develop recruitment strategies, improve treatment procedures, and better understand participants’ perspectives. Additionally, multiple measures of substance use outcomes should be examined in a single study as a way to strengthen the robustness of the treatment effects. Separate analyses should be conducted for different categories of substance used (e.g., alcohol, cannabis, and narcotics) when appropriate.

In conclusion, although existing evidence on the effectiveness of substance use interventions for homeless youth is emerging, it is quite limited in informing evidence-based practice. Mixed findings are often reported for the same intervention. The heterogeneity of the interventions and the few studies conducted on each intervention makes it difficult to determine the relative efficacy of a specific intervention. The quality of the studies was often poor, which further impedes drawing firm conclusions about the efficacy of interventions. Hopefully, the directions presented will assist the development of more high-quality intervention studies to strengthen the evidence base of substance use interventions for homeless youth.

Footnotes

Acknowledgments

The author would like to thank Dr. Mary Keegan Eamon for her comments that helped improve the article.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

References

Altena

A. M.

Brilleslijper-Kater

S. N.

Wolf

J. R. L. M.

(2010). Effective interventions for homeless youth: A systematic review. American Journal of Preventive Medicine, 38, 637–645. doi:10.1016/j.amepre.2010.02.017

Baer

J. S.

Garrett

S. B.

Beadnell

Wells

E. A.

Peterson

P. L.

(2007). Brief motivational intervention with homeless adolescents: Evaluating effects on substance use and service utilization. Psychology of Addictive Behaviors, 21, 582–586. doi:10.1037/0893-164X.21.4.582

Booth

R. E.

Zhang

Y. M.

Kwiatkowski

C. F.

(1999). The challenge of changing drug and sex risk behaviors of runaway and homeless adolescents. Child Abuse and Neglect, 23, 1295–1306. doi:10.1016/S0145-2134(99)00090-3

Cauce

A. M.

Morgan

C. J.

Wagner

Moore

Wurzbacher

Blanchard

(1994). Effectiveness of intensive case management for homeless adolescents: Results of a 3-month follow-up. Journal of Emotional and Behavioral Disorders, 2, 219–227.

Cauce

A. M.

Paradise

Embry

Morgan

C. J.

Lohr

Theofelis

Wagner

(1998). Homeless youth in Seattle: Youth characteristics, mental health needs, and intensive case management. In Epstein

M. H.

Kustash

Duchnowski

(Eds.), Outcomes for children and youth with behavioral and emotional disorders and their families: Programs and evaluation best practices (pp. 611–632). Austin, TX: Pro-ed.

Farabee

Shen

Hser

Y. I.

Grella

C. E.

Anglin

M. D.

(2001). The effect of drug treatment on criminal behavior among adolescents in DATOS-A. Journal of Adolescent Research, 16, 679–696. doi:10.1177/0743558401166009

Ferguson

K. M.

Xie

(2008). Feasibility study of the social enterprise intervention with homeless youth. Research on Social Work Practice, 18, 5–19. doi:10.1177/1049731507303535

Friedman

M. S.

Marshal

M. P.

Stall

Kidder

D. P.

Henny

K. D.

Courtenay-Quirk

Holtgrave

D. R.

(2009). Associations between substance use, sexual risk taking and HIV treatment adherence among homeless people living with HIV. AIDS Care, 21, 692–700. doi:10.1080/09540120802513709

Greene

J. M.

Ennett

S. T.

Ringwalt

C. L.

(1997). Substance use among runaway and homeless youth in three national samples. American Journal of Public Health, 87, 229–235. Retrieved from http://ajph.aphapublications.org/doi/pdf/10.2105/AJPH.87.2.229

10.

Harris

R. P.

Helfand

Woolf

S. H

Lohr

K. N.

Mulrow

C. D.

Teutsch

S. M.

Atkins

(2001). Current methods of the U.S. preventive services task force: A review of the process. American Journal of Preventive Medicine, 20, 21–35. doi:10.1016/S07493797(01)00261-6

11.

Hwang

S. W.

Tolomiczenko

Kouyoumdjian

F. G.

Garner

(2005). Interventions to improve the health of homeless: A systematic Review. American Journal of Preventive Medicine, 29, 311.e1–311.e75. doi:10.1016/j.amepre.2005.06.017

12.

Johnson

K. D.

Whitbeck

L. B.

Hoyt

D. R.

(2005). Substance abuse disorders among homeless and runaway adolescents. Journal of Drug Issues, 35, 799–816.

13.

Kisely

S. R.

Parker

J. K.

Campbell

L. A.

Karabanow

Hughes

J. M.

Gahagan

(2008). Health impacts of supportive housing for homeless youth: A pilot study. Public Health, 122, 1089–1092. doi:10.1016/j.puhe.2008.01.009

14.

Meyers

R. J.

Smith

J. E.

(1995). Clinical guide to alcohol treatment: The community reinforcement approach. New York, NY: Guilford Press.

15.

Milburn

N. G.

Iribarren

F. J.

Rice

Lightfoot

Solorio

Rotheram-Borus

M. J.

Duan

(2012). A family intervention to reduce sexual risk behavior, substance use, and delinquency among newly homeless youth. Journal of Adolescent Health, 50, 358–364. doi:10.1016/j.jadohealth.2011.08.009

16.

Miller

W. R.

Rollnick

(2002). Motivational interviewing: Preparing people to change addictive behavior (2nd ed.). New York, NY: Guilford Press.

17.

Orwin

R. G.

Garrison-Mogren

Jacobs

M. L.

Sonnefeld

L. J.

(1999). Retention of homeless clients in substance abuse treatment: Findings from the National Institute on Alcohol Abuse and Alcoholism Cooperative Agreement Program. Journal of Substance Abuse Treatment, 17, 45–66. doi:10.1016/S0740-5472(98)00046-4

18.

O’ Toole

T. P.

Gibbon

J. L.

Hanusa

B. H.

Freyder

P. J.

Conde

A. M.

Fine

M. J.

(2004). Self-reported changes in drug and alcohol use after becoming homeless. American Journal of Public Health, 94, 830–835. doi:10.2105/AJPH.94.5.830

19.

Padgett

D. K.

Henwood

Abrams

Davis

(2008). Engagement and retention in services among formerly homeless adults with co-occurring mental illness and substance abuse: Voices from the margins. Psychiatric Rehabilitation Journal, 31, 226–233. doi:10.2975/31.3.2008.226.233

20.

Peterson

P. L.

Baer

J. S.

Wells

E. A.

Ginzler

J. A.

Garrett

S. B.

(2006). Short-term effects of a brief motivational intervention to reduce alcohol and drug risk among homeless adolescents. Psychology of Addictive Behaviors, 20, 254–264. doi:10.1037/0893-164X.20.3.254

21.

Pollio

D. E.

Thompson

S. J.

Tobias

Reid

Spitznagel

(2006). Longitudinal outcomes for youth receiving runaway/homeless shelter services. Journal of Youth and Adolescence, 35, 859–866. doi:10.1007/s10964-006-9098-6

22.

Rotheram-Borus

M. J.

Song

Gwadz

Lee

Rossem

R. V.

Koopman

(2003). Reductions in HIV risk among runaway youth. Prevention Science, 4, 173–186. doi:10.1023/A:1024697706033

23.

Sedlak

A. J.

Finkelhor

Hammer

Schultz

D. J.

(2002). National estimates of missing children: An overview. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

24.

Slesnick

Bartle-Haring

Gangamma

(2006). Predictors of substance use and family therapy outcome among physically and sexually abused runaway adolescents. Journal of Marital and Family Therapy, 32, 261–281. doi:10.1111/j.1752-0606.2006.tb01606.x

25.

Slesnick

Dashora

Letcher

Erdem

Serovich

(2009). A review of services and interventions for runaway and homeless youth: Moving forward. Children and Youth Services Review, 31, 732–742. doi:10.1016/j.childyouth.2009.01.006

26.

Slesnick

Kang

M. J.

Bonomi

A. E.

Prestopnik

J. L.

(2008). Six- and twelve-month outcomes among homeless youth accessing therapy and case management services through an urban drop-in center. Health Services Research, 43, 211–229. doi:10.1111/j.1475-6773.2007.00755.x

27.

Slesnick

Meyers

R. J.

Meade

Segelken

D. H.

(2000). Bleak and hopeless no more: Engagement of reluctant substance-abusing runaway youth and their families. Journal of Substance Abuse Treatment, 19, 215–222. doi:10.1016/S0740-5472(00)00100-8

28.

Slesnick

Prestopnik

J. L.

(2005). Ecologically based family therapy outcome with substance abusing runaway adolescents. Journal of Adolescence, 28, 277–298. doi:10.1016/j.adolescence.2005.02.008

29.

Slesnick

Prestopnik

J. L.

(2009). Comparison of family therapy outcome with alcohol-abusing, runaway adolescents. Journal of Marital and Family Therapy, 35, 255–277. doi:10.1111/j.1752-0606.2009.00121.x

30.

Slesnick

Prestopnik

J. L.

Meyers

R. J.

Glassman

(2007). Treatment outcomes for street-living, homeless youth. Addictive Behaviors, 32, 1237–1251. doi:10.1016/j.addbeh.2006.08.010

31.

Souza

Porten

Nicholas

Grais

, & the Medecins Sans Frontiere-Honduras Team. (2011). Outcomes for street children and youth under multidisciplinary care in a drop-in centre in Tegucigalpa, Honduras. International Journal of Social Psychiatry, 57, 619–626. doi:10.1177/0020764010382367

32.

Steele

R. W.

O’Keefe

M. A.

(2001). A program description of health care interventions for homeless teenagers. Clinical Pediatrics, 40, 259–263. doi:10.1177/000992280104000504

33.

Stewart

Reutter

Letourneau

Makwarimba

(2009). A support intervention to promote health and coping among homeless youth. Canadian Journal of Nursing Research, 41, 54–77.

34.

Stotts

A. L.

Schmitz

J. M.

Rhoades

H. M.

Grabowski

(2001). Motivational interviewing with cocaine-dependent patients: A pilot study. Journal of Consulting and Clinical Psychology, 69, 858–862. doi:10.1037/0022-006X.69.5.858

35.

Van Leeuwen

J. M.

Hopfer

Hooks

White

Peterson

Pirkopf

(2004). A snapshot of substance abuse among homeless and runway youth in Denver, Colorado. Journal of Community Health, 29, 217–299. doi:10.1023/B: JOHE.0000022028.50989.aa

36.

Weinberg

N. Z.

Rahdert

Colliver

J. D.

Glantz

M. D.

(1998). Adolescent substance abuse: A review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 252–261. doi:10.1097/00004583-199803000-00009

37.

Whitbeck

L. B.

(2009). Mental health and emerging adulthood among homeless young people. New York, NY: Psychology Press.

38.

Zerger

(2002). Substance abuse treatment: What works for homeless people? A review of the literature. Nashville, TN: National Health Care for the Homeless Council.

39.

Zerger

Strehlow

A. J.

Gundlapalli

A. V.

(2008). Homeless young adults and behavioral health: An overview. American Behavioral Scientist, 51, 824–841. doi:10.1177/0002764207311990