To facilitate timely access to social services in hospitals, high social risk (HSR) screening mechanisms have been implemented throughout the country. The purpose of these mechanisms is to identify patients upon admission who are most likely to need service. Because these screening mechanisms often identify patients who do not need service, this study determined how well the patient characteristics used as screening factors actually predicted service utilization. A multivariate logistic regression model was developed using 1,659 discharges from two midwestern hospitals. The most powerful predictors of social service use were age, multi-system illness, not having a spouse, and psychiatric co-morbidity. Despite the significant contribution of each predictor, their combined ability to predict who receives service was modest.