Abstract
Fall Risk Index (FRI), consisting of 21 simple yes/no questions, has been shown to successfully screen community-dwelling older Japanese for fall risk. To identify the fall-prone inpatients effectively, we examined additional risk factors besides FRI. We retrospectively investigated 253 inpatients in the University of Tokyo Hospital, discharged from April 2016 to March 2017, mainly hospitalized for cognitive impairment or acute illness such as pneumonia. The data include patients’ characteristics, FRI and a history of falls in the year before admission. T-tests, chi-square tests, linear and logistic regressions and receiver operating characteristic (ROC) curves were used for evaluating strong fall risk factors in addition to FRI. Eighty-eight inpatients (46.7% in 165 patients) fell in the year preceding admission. The FRI ranged from 0 to 20 (mean 11 ± 4). Between fallers and non-fallers, the differences were significant in age, serum albumin, FRI, Mini Mental State Examination (MMSE), Specific Activity Scale, Barthel index and Lawton Instrumental Activities of Daily Living Scale. In logistic regression with these variables and sex, cognitive impairment by low MMSE (≦23/30 points) and men showed positive trends toward fall. In the stepwise linear regression with age, sex and FRI, only MMSE was selected in the model. Adjusted for age, sex, serum albumin and FRI, cognitively impaired patients showed fall odds of 2.57 (95%CI 1.10–6.02). The area under the ROC curve was the largest for FRI (0.77) in the cognitively impaired population. Therefore, additionally to FRI, cognitive impairment should be taken into account to assess fall risk for inpatients.