Street Victimization and Mental Health Consequences Among Homeless Youth: A Latent Class Analysis

Background: Exposure to street victimization is common among homeless youth, with more than 80% reporting exposure.  Continuous exposure to trauma without treatment can lead to the development of posttraumatic stress disorder (PTSD) and depression.  The transient lifestyle of homeless youth includes fewer stable relationships and inconsistent mental health services, thus increasing risk of psychopathology following victimization exposure. Despite recognition of the scope of this problem, little work has investigated whether subgroups of homeless youth exist with different victimization profiles and mental health service needs. This study sought a large sample of homeless youth from three U.S. cities to examine victimization profiles. Two research questions addressed whether 1) distinct subgroups (or latent classes) exist based on type and frequency of victimization experiences, and 2) subgroups differ in PTSD and depression symptomatology.

Methods: Data were analyzed from a multi-site, cross-sectional study of youth (N=601) ages 18-25 seeking homeless services in Denver (n=201), Austin (n=200) and Los Angeles (n=200). Face-to-face quantitative interviews assessed demographics, transience (number of intercity moves), symptoms of depression and PTSD (Mini International Neuropsychiatry Interview), and both direct (e.g., physical assault) and indirect (e.g., witnessed violence) victimization on the streets (Traumatic Life Events Questionnaire). Latent class analysis was used to identify and describe classes of youth based on their victimization experiences. Multinomial regression was used to examine mental health symptoms associated with latent class membership, controlling for demographics.

Results: Latent class analysis revealed that a three-class solution best fit the data, according to BIC/adjusted BIC fit indices. A large class of youth (n=204), labeled non-victims, reported rarely experiencing any type of victimization; a second class (n=136), labeled victims, experienced extremely high rates of direct and indirect victimization; and a third class (n=250), labeled witnesses, appeared to be on the periphery of violence, reporting high rates of threats and witnessed violence, but rarely reporting direct victimization. Multivariate analyses indicated that, compared to non-victims, witnesses and victims were more likely to have been homeless longer, to be more transient, and to meet criteria for depression and PTSD (c² [14]=142.6, p<.001, R2=.21). Specifically, compared to non-victims, witnesses and victims were nearly 2 times more likely to experience depressive symptoms and 2.5 times more likely to experience PTSD symptoms. Although victims and witnesses experienced similar levels of mental health symptoms, victims reported greater levels of transience than witnesses.  

Conclusions: Our findings indicate youth exposed to violence as witnesses experience mental health problems equal to youth who are directly victimized. This suggests that youth should be screened for experiences of direct and indirect victimization, and both groups should be engaged in mental health services. Youth transience creates a challenge in applying these recommendations, as transient youth were more likely to witness traumatic events, and highly transient youth were at greatest risk for direct victimization. Furthermore, transience often disrupts service engagement. Service providers should consider innovative technologies (cell phone, texting, email, social networking) as modalities for providing mobile services for transient homeless youth in conjunction with efforts to stabilize youth in permanent housing.