Construction and validation of a fall risk prediction model in elderly maintenance hemodialysis patients: a multicenter prospective cohort study.

Purpose: To analyze the risk factors for falls in elderly maintenance hemodialysis patients, construct a nomogram prediction model and validate the application.

Background: Elderly maintenance hemodialysis patients face a high risk of falls, and there are fewer and less effective fall-specific assessment tools.

Method: A total of 871 elderly hemodialysis patients from 9 hospitals in Chengdu City from October 2023 to December 2024 were selected as the study objects. Baseline characteristics and fall outcomes of patients in the fall group and non-fall group were recorded and compared through 6-month follow-up. Multivariable logistic regression analysis was employed to identify independent risk factors, and construct the nomogram prediction model and complete the internal verification of the model. 218 elderly maintenance hemodialysis patients from three other hospitals in Chengdu City were selected for a 6-month follow-up of falls from January to February 2024 to complete the external validation of the model.

Result: The incidence of falls in elderly maintenance hemodialysis patients was 31.96%, and logistic regression analysis showed that age, sex, visual impairment, intradialytic hypotension, cognitive impairment and depression were independent risk factors for falls. Both internal and external validation of the model demonstrated area under the curve greater than 0.80. Furthermore, calibration plots, the Hosmer-Lemeshow test, and clinical decision curves all demonstrated that the model had good calibration and clinical utility.

Conclusion: The nomogram constructed based on the above risk factors can provide scientific basis and practical tools for early clinical identification of high-risk groups of falls.