Abstract
We analyzed the ADL structure for stroke patients at admission and discharge based on the Functional Independence Measure (FIM).
The subjects were 190 stroke patients who had been admitted to Keio University Hospital and its eight affiliated hospitals in 1993 within 4 months of the occurrence of stroke. The “Independence of the FIM items” (percentage of the patients scoring 6 and 7 of all patients at each item) was calculated from the FIM scores at admission and discharge. The scores were also converted to the interval scale by use of the Rasch analysis program, and the item score distribution by each total score was calculated.
The difficulty pattern for motor FIM items shows almost no differences between admission and discharge between patients with left and right hemisphere lesions, and for age. Bladder management, bowel management and eating were the easiest items, whereas stairs and tub/shower transfer were the most difficult. Concerning cognitive FIM items, expression was easiest for patients with right hemisphere lesions and most difficult for those with left hemisphere lesions. Dementia (Mini Mental State<10) affected the difficulty patterns of motor and cognitive FIM items. The motor FIM items were divided into five groups according to total score. 50>represent the completely assisted group (dependence on a helper for all items or for all items except eating), 50≤represent the incompletely assisted group (independence for eating, grooming, bladder management and bowel management, but dependence on a helper for the other items)<70, 70≤represent the independent group for “self-care-associated items” (dependence on a helper for bathing, tub/shower transfer and walk or wheel chair, but independence for the other items)<80, 80≤represent the independent indoor walking group (independence for all items except stairs)<85, and 85≤represent the independent outdoor walking group (independence for all items). Because the cognitive items do not have the unidimensionality, they fit the Rasch model poorly. We must be careful when using cognitive FIM data converted by Rasch analysis.
