Development of a new predictive model for falls among inpatients using the Japan MHLW bedriddenness rank in daily living - Development of predictive model for inpatients' fall

7858 adult patents were available. Using model 1 consist of 13 factors, multivariate logistic regression analysis showed that the significant related factors were male, emergency admission, Department of Neurosurgery, use of hypnotic, previous falls, independent to eat, and bedriddenness rank. The AUC of model 1 was 0.808 in the test set, 0.789 in the validation set. For the model consisting of above 7 factors and age, the AUC was 0.806 in the test set, 0.787 in the validation set.