| Outreach workers of the Department of Veterans Affairs (VA) Comprehensive Homeless Center engage homeless veterans and refer them to services. For this study, selected predictor variables were correlated with the dependent variable, service utilization. Based on the Andersen and Newman (1973) model, this study theorized that utilization of services is the consequence of predisposing, enabling and need factors. Variables predicted to relate to service use were past hospitalization, knowledge of VA services, accessibility of services, veteran-perceived need and worker-observed need. An instrument was used to gather sociodemographic characteristics, residential, military history, workers' treatment plans and veteran use of services. A six-month sample of 414 homeless veterans interviewed in soup kitchens, shelters and drop-in centers was selected. Variables were measured as follows: past hospitalization was a sum of past hospitalizations, knowledge of VA services was VA public support and recent use of VA services, accessibility was income received in the last 30 days, veteran-perceived need was the sum of services needed and worker-observed need was the sum of diagnoses. The dependent variable was utilization of service, the sum of medical, mental health, financial, vocational, alcohol and drug treatment and residential services. Multiple regression using a path analytic framework was used to test for relationships between independent predictor variables and the dependent variable, utilization of services. Of the 414 study veterans, 175 (42.3 percent) used 228 of the 576 services for a utilization rate of 39.6 percent. Significant predictors were worker-observed need and knowledge of VA services for use of all services, military service before Vietnam and lack of housing for medical services, veteran-perceived and worker-observed need for mental health services, being female and not previously hospitalized for financial services, having housing and no access for vocational services, worker-observed need and knowledge for alcohol and drug treatment and knowledge of services for residential services. Conclusions were that the predictive model is limited, need for comprehensive services is substantiated, outreach is effective and the VA is responsive to homeless veterans who then use services. Recommendations were made for the social work profession. |
