Discharge planning for hospitalized older veterans

A random sample of 175 patients in six nursing and personal care settings at the Veterans Administration Medical Center, Little Rock, Arkansas was drawn in order to develop a model for the classification of patients for the purpose of effecting more efficient and effective hospital discharge planning. Data that might effect significant changes in the placement dispositions and outcomes of patients were gathered and subjected to stepwise discriminant function and multiple regression analyses. It was found that the patient's socioeconomic status and the extent of social worker involvement in disposition rather than the patient's medical/psychiatric condition, constituted the best predictors of placement and success in placement. The model accurately predicted 90 percent of all dispositions and could provide a useful basis for assigning to placements a significant minority of patients who are not placed in accordance with the correlates of optimum placement and placement outcome as determined by the model.