ABSTRACT
BACKGROUND: To develop and internally validate a prediction model for identifying patients with hematologic diseases of fall risk. RESEARCH DESIGN AND METHODS: This is a prospective cohort study from a prospective collection of data for 6 months. We recruited 412 patients with hematologic diseases in medical institutions and home environment of China. The outcome of the prediction model was fall or not. These variables were filtered via univariable logistic analysis, LASSO, and multivariable logistic analysis. We adopt an internal validation method of K-fold cross validation. The area under the ROC curve and the H-L test were used to evaluate the discrimination and calibration of the model. RESULTS: Five influencing factors were identified multivariable logistic regression analysis. The established model equation is as follows: the H-L goodness-of-fit test of the model p > 0.05. The area under the ROC curve of train is 0.957 (95% CI: 0.936 ~ 0.978), and the area under the ROC curve of test is 0.962 (95% CI: 0.884 ~ 1), so the model calibration and discriminant validity are good. CONCLUSION: Our equation has good sensitivity and specificity in predicting the fall risk of patients with hematologic diseases, and has certain positive significance for clinical assessment of their fall risk. TRIAL REGISTRATION NUMBER: ChiCTR2200063940.