The development of a risk prediction model for fall in patients with low vision: Based on lasso regression.

Falls are a global public health concern, causing millions of injuries and deaths worldwide. Low vision, characterized by moderate to severe visual impairment, poses a high risk of falls due to its multiple risk factors. This study focuses on developing a fall risk prediction model for low vision patients. A total of 162 low vision patients were included in the validation of the prediction model, and the fall rate was found to be 31.48%. The results showed that a history of syncope, activities of daily living (ADL), vision field, distance vision, and collision in unfamiliar surroundings were the most predictive factors. The regression analysis showed (R2=0.95, C-index=0.998). A nomogram was constructed, ROC curves and calibration curves were plotted, giving an AUC of 0.9978 for a well-calibrated and differentiated model. DCA decision curves demonstrated the superior net benefit and predictive accuracy of the model. This model shows great potential in predicting falls in low vision patients and aiding in clinical identification of those at high risk.