Research That Matters (January 17 - 20, 2008)

Background and purpose: Victimization is a common problem among homeless, seriously mentally ill populations. Previous research has suggested that recent victimization is negatively associated with service use, and is differentially impacted based on variables such as age and mental illness. However, research to date has not modeled relationships among victimization, mental illness, and mediating factors. The purpose of this study was to test influence of victimization on depressive symptoms among homeless persons with serious mental illness. It was hypothesized that victimization will be associated with increased recent depressive symptoms with perceived safety as a partial mediating factor.

Method: Data for this study were derived from the Access to Community Care and Effective Services and Supports (ACCESS) program. ACCESS was an 18-site, five-year project intended to improve individual functioning, quality of life, and housing outcomes for homeless persons with serious mental illness (analytic n=10,221). The hypothesized model was initially tested for the entire sample using structural equation modeling (SEM). Depression was measured by total symptoms reported during the previous two months using items from the Diagnostic Interview Survey and Psychiatric Epidemiology Research Interview. Victimization (physical and non-physical) was measured using the sum of five different types of victimization for the same period. Perception of safety was measured using items from the Lehman Quality of Life Instrument. Subsequently, the model was tested using an SEM multiple group framework against five different variables (age, gender, homeless duration, substance use disorder, and psychotic disorder) hypothesized to influence the parameter estimates of the model.

Results. Approximately 42% of the subjects reported at least one form of victimization. The overall model exhibited a good fit with the data (RMSEA=.028, GFI=1.0). Non-physical victimization was positively associated with depression (â = .12, p <.001), and physical victimization was negatively associated with perception of safety (â = -.28, p < .001). Perception of safety was negatively associated with depression (â = -.34, p < .001). The model explained approximately 13% of the variance in depression. The multiple group analyses showed that the direction of effects was consistent across subgroups, with small changes in effect sizes.

Conclusions and Implications. This study supports the hypothesized model; victimization influences depressive symptoms, with perceived safety as a mediating factor. The findings form this study affirms the importance of facilitating a perception of safety among the homeless population. However, given the consistent finding by this study and others of disproportionate rates of victimization and trauma, it is clear that beyond advocating for increasing the perception of safety, it is equally (or if not more) important to provide safe living spaces. Additionally, the findings from this study point to the need to focus services specifically on the impact of victimization and trauma on the homeless population. Particularly, it highlights the need to offer these services specifically to the population with severe mental illness, most specifically with depression or depressive symptoms.