Needs assessment surveys of homeless adults, while rarely based on representative samples, have revealed consistent findings: that they are paradoxically a multi-problem but undeserved population; that they order their service needs differently from the professionals order them; that being mentally ill and/or an abuser of drugs or alcohol is associated with not getting into services. However, no study has yet examined the relative contribution of expressed needs and other characteristics of the individual or of service provider or service system factors, to utilization of services. This study will examine the """"""""interactive fit"""""""" of these three types of variables in help- seeking for mental health or substance abuse problems. Specifically, the aims of the research are to understand (1) how expressed needs are determined and structured. (2) what contributes to getting into mental health, alcohol, or drug treatment. We also aim to disseminate these findings to relevant city, state, federal and private agencies. To pursue these aims, the study will use a data-set based on a random study, directed in 1987 by the Principal Investigator, of over 1000 men and women in the New York City adult shelter system. The survey used a 59-page interview protocol to obtain detailed information in areas such as health, mental health, substance abuse, housing, treatment, etc. It included standardized scales, with established reliability and modified by use for a homeless population. In addition, twenty areas of need were rated by both responders and interviewers. The study will also use ecological-level data to indicate presence of a """"""""drug culture"""""""" in and around the shelters, and availability of services. Factor analytic procedures will be applied to identify dimensions of need and hierarchical multiple regression will be used to identify variables within variable sets hypothetically linked to measures of need. The help-seeking models of service utilization will be tested with logistic regression analyses, entering the variables hierarchically. All analyses will be run separately on men and women, as gender differences are known to affect expression of needs and help-seeking. The study will be the first to analyze individual and contextual level variables in understanding how needs and service utilization are related in a homeless population.
