Categorizing Risk and Resilience for Homeless Youth: A Typology to Inform Service Delivery

Background/Purpose: Homeless youth are a diverse and vulnerable population with complex service needs. In order to better target interventions for this heterogeneous population, policymakers and researchers have articulated a need for data-driven typologies that delineate subgroups with potentially different service needs, based on identifiable risk and resilience factors. This study used data from a survey of homeless young adults to examine whether distinct subgroups could be identified and whether these subgroups could inform screening and service planning. 

Methods: Survey data were collected at 47 locations over 4 weeks in Houston, Texas.  Youth (18-24) in unstable housing situations were included (N=379).  Trained volunteers assessed eligibility and obtained consent; then youth self-administered the survey.  Youth self-reported risk factors including public system involvement, arrest, substance use, mental disorders, and health. Trauma history was assessed using the Adverse Childhood Experiences scale and mental distress was assessed with the Kessler-6.  Resilience factors included employment, high school graduation and presence of supportive adults.  Latent class analysis was used to examine whether there were subgroups with distinct patterns across these indicators.   Bivariate analyses were used to describe the classes identified.  SBIC, entropy and the Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (LRT) were used to identify the best-fitting model.

Results: A model with four distinct subgroups was the best fit for the data (SBIC=9030.9; Entropy=.86: LRT=122.94, p<.05).  Across these four groups, two were similar in having relatively low rates of substance use, mental distress, and traumatic experiences; however, one of these groups had a lack of adult support (Low Trauma/Low Support, n=61), while the other had the highest rates of adult support across all groups (Low Trauma/High Support, n=74).  The remaining two groups were characterized by high rates of trauma and mental health problems.  One of these groups (Trauma and Mental Health, n=153) had lower rates of substance use and criminal justice involvement compared to the other.  The last group had high rates of substance use, arrest, and prior involvement with foster care and juvenile justice in addition to the highest rates of trauma and mental health problems (Multiple Problems, n=91).  The Low Support/Low Trauma group was disproportionately African American and male while the Trauma and Mental Health group was disproportionately female and the Multiple Problems group was disproportionately white. 

Conclusions/Implications:  The risk and resilience factors successfully distinguished four groups of young people with different service needs.  Young people with low trauma also had low rates of most risk behaviors, so likely would need lower intensity services focused mainly on housing.  Among youth with low trauma, one subgroup clearly needed connections with supportive adults, while the other had adults that could be leveraged as a resource for interventions.  The majority of youth in the sample had greater trauma exposure and indicators of need for other supportive services, especially mental health supports.  The Multiple Problems group had indicators of wide-ranging needs across all risk indicators and this group could benefit from the most intensive service package, including mental health and substance use supports.