Paul A. Toro - US grants

The proposed longitudinal research project, in the context of an ecological/resource framework, attempts to understand the conditions under which the severely mentally ill and other homeless persons are able to secure and maintain stable housing (vs. remain homeless). Probability sampling methods will be used to obtain a representative sample of homeless adults in the Buffalo metropolitan area. A broad range of sites will be sampled, including psychiatric facilities, special programs for the homeless, food pantries, and the streets, in addition to the shelters and soup kitchens commonly included in past research on the homeless. Based on our framework, data will be collected in three domains: Personal Attributes, including psychiatric diagnoses and background characteristics; Resources, in personal, social, economic, and service areas; and Outcomes, including detailed information on the course of homelessness and on subsequent housing characteristics. The data will be obtained from four sources: the homeless participant; two significant others, including at least one family member; interviewer ratings; and agency records. The study is especially distinctive among existing studies on the homeless in considering a wide range of social resources, including social support, family social climate, and family violence as predictors of exiting homelessness and subsequent housing stability. The study also includes a longer follow-up period than has so far been attempted (five interviews over two years) and would allow a longitudinal comparison of a variety of important subgroups among the homeless, such as the severely mentally ill (defined based on DSM-III-R criteria for schizophrenia or major affective disorder), substance abusers (also defined based on DSM-III-R criteria), and homeless women (including those with children). The two principal research goals are to: (1) Determine factors associated with obtaining stable housing; and (2) Identify causal mechanisms involved in exiting homelessness. In addition, the study will: (a) provide information on the possible effectiveness of a full range of services currently used by the homeless and whether certain combinations of services may be particularly helpful; (b) compare the utility of several commonly used methods for assessing the mental health status of the homeless; and (c) develop new measures needed for the study of homelessness (e.g., to assess characteristics of homelessness and subsequent housing quality). Through extensive preliminary research, procedures have been refined to ensure a low attrition rate among the homeless sample in this longitudinal research.

DESCRIPTION: The proposed longitudinal research project will compare the outcomes of homeless adolescents (ages 13-17) to the outcomes of a matched housed group of similar adolescents. The homeless sample will come from throughout the Detroit metropolitan area and will include all subgroups (e.g., "runaways" and "throwaways") to the extent that they exist in the total homeless youth population. The probability sample of 250 homeless adolescents will be sampled from 5 different types of settings: shelters for runaway/homeless youth; outreach programs for street youth and drop-in centers; in-patient and out-patient substance abuse and mental health agencies; the juvenile justice system; and soup kitchens and various street sites. The 250 matched housed adolescents will be matched on gender, age, race, and neighborhood socioeconomic characteristics. Measures will be assessed longitudinally in the following domains: Family and context (e.g., parenting, violence, poverty), psychological processes (e.g., competence, self-efficacy, alcohol expectancies), homelessness and stress, utilization of social services, peer relations (e.g., deviant associations, peer support), and outcomes (e.g., alcohol and drug use/abuse, delinquency, academic achievement, mental health and suicidal proneness, physical health and HIV-risk behaviors). The principal research goals are to: (1) describe a representative sample of homeless adolescents and compare them to matched housed adolescents; (2) investigate the longitudinal consequences of adolescent homelessness; (3) test theoretical models on longitudinal pathways producing vulnerability and resilience in homeless and housed at-risk youth; (4) improve understanding of family dynamics among homeless adolescents (most of whom return home); (5) assess longitudinal differences in outcome patterns for African- vs. European-Americans and for girls vs. boys; and (6) document the effectiveness of various services used by homeless adolescents. Through extensive preliminary research, measures and sampling techniques have been refined and procedures have been developed to ensure a low attrition rate in the project's broad sample.

The proposed 5-year continuation research project will assess longitudinal outcomes during the transition from adolescence into early adulthood for a large 2-group sample of 401 at-risk urban youth: (1) a probability sample of 251 homeless adolescent from throughout the Detroit metropolitan area and (2) 150 initially housed adolescents matched on gender, age, race, and SES. Measures assessed longitudinally include family and contextual factors, psychological processes (e.g., competence, self-efficacy), homelessness and stress, peer relations, and outcomes that often begin in late adolescence and coalesce in early adulthood (e.g., alcohol and drug use/abuse, criminal behavior, education and employment, mental and physical health, HIV-risk behaviors). The proposed research builds on the original study by extending the follow-up period from adolescence (initial-final mean ages= 15.5-17.0) into early adulthood (initial-final mean ages=19.5- 21.5). We believe that is during this period of early adulthood that the consequences of problematic experiences in adolescence will become most clearly visible. Relatively few studies have tracked at-risk youth through the important transition. Specifically we will: (1) investigate risk and protective factors predicting longitudinal outcomes and test theoretical models on longitudinal pathways in our at-risk sample; (2) improve understanding of family dynamics among at-risk urban youth during the transition to early adulthood; (3) assess the impact of peer relationships; and (4) assess longitudinal differences in outcomes for African-vs. European-Americans and for girls vs. boys. In the original 4- year project, the rate of attrition has been low and we expect that 179 (81%) of the initially homeless and 141 (94%) of the housed ill receive at least one follow-up interview during early adulthood in the continuation project. Preliminary findings over 1.5 years suggest that our at-risk youth show some short-term improvements (perhaps due to service, family, or other factors) but are beginning to show signs of negative longitudinal outcomes (e.g., alcohol and drug problems) as they turn 16-17. W e are also finding some support for out theoretical perspectives. For example, supporting a social learning perspective, we have found that family relations, school problems, and deviant peers all have roles in the development of alcohol use/abuse.