CATON ET AL. RESPOND

We appreciate Alexander-Eitzman’s interest in our article and welcome the opportunity to respond to his letter. We take issue with his assertion that our approach in this study of risk factors for long-term homelessness is one where the “tools define the problem,”1(pXXX) with the implication that our predictor variables were heavily tilted toward the assessment of psychopathology. Although our assessment battery did include standardized assessments of health and mental health status, the tools did not define the results. Our overall findings underscored the heterogeneity of the contemporary homeless, but a Kaplan–Meier survival analysis showed that participants who were younger and those who had better psychosocial adjustment, recent or current employment, earned income, and adequate family support experienced shorter durations of homelessness than those without these characteristics. In the Discussion section of our article we concluded that these findings suggest that new shelter admissions presenting with such characteristics might be offered assistance with job searches, temporary income support, or other supportive services that might provide a fast track out of homelessness. We reject the argument by Meyer and Schwartz2 that epidemiological analyses of social problems such as homelessness should be avoided. The identification of individual-level risk factors for chronic homelessness has practical implications for service delivery in that it suggests new preventive intervention approaches. Other findings we reported, specifically the characteristics of living arrangements prior to shelter entry and upon exit from homelessness, illustrate the scarcity of housing options available to this economically disadvantaged population. Although not predictive of duration of homelessness, widespread and long-standing mental health and substance abuse problems among our study subjects indicate a substantial unmet need for public health and mental health services in the disadvantaged neighborhoods from which these newly homeless men and women emerged. We agree completely with Alexander-Eitz-man’s recommendation that the longitudinal study of homelessness be broadened to include multiple levels of analysis. A strategy for the prevention of chronic homelessness should encompass interventions at both individual and population levels and should address housing, health, and mental health policy and program development issues.