Abstract
We collected and analyzed large-scale data concerning the fall of stroke inpatients in convalescent rehabilitation wards. Three hundred seventy-four of 1,107 inpatients experienced one fall or more, and 16 factors associated with falls were clarified by the chi-square test. To extract the significant item from a multifactor, the logistic regression analysis of 16 factors was carried out, and we developed an assessment sheet for the risk degree of first fall prediction in stroke inpatients. We selected eight variables as the items on the assessment sheet : history of previous falls, central paralysis, visual impairment, sensory disturbance, urinary incontinence, use of psychotropic medicines, mode of locomotion, and cognitive impairment. The total score of the assessment sheet was ranged from 0 to 10 and the mean score of fallers (6.4±1.5) among subjects was significantly higher than that of non-fallers (5.1±1.9) (p <0.001). When the subjects were classified into three groups, a significant difference (p<0.001) in the tendency of fall incidence in term of days after admission was found among the three groups on the basis of the Kaplan-Meier survival curve.
