Predicting discharge destination of stroke patients using a mathematical model based on six items from the Functional Independence Measure.

Abstract

A mathematical model using selected items from the Functional Independence Measure (FIM) was developed to predict disposition of stroke patients from an acute care hospital. Case series of 279 acute stroke patients admitted to the hospital from 4/91 through 12/93. Data collection remains ongoing; a second series of patients will be analyzed to validate results and determine if the same 6 FIM items are significant. Tertiary care center, general acute care community hospital. Diagnosis of acute stroke; 298 patients whose attending physician referred the patient to the rehabilitation team. An 18-item FIM was administered within 3 days of admission and 24 hours of discharge. FIM scores were analyzed to determine if the initial score could be used to predict disposition. Discriminant analysis identified 6 items as being statistically significant in predicting discharge to home, rehabilitation facility, or nursing home. Patient's actual discharge location was compared with the location predicted by the model developed using the FIM. Bathing, bowel, toileting, social interaction, dressing lower body, and eating were the selected initial FIM items that predicted disposition with 70% accuracy in our patient series. The FIM is effective in the acute care setting to help determine appropriate discharge status, with certain variables being more predictive than others.