Results of Binary Logistic Regression Analysis.

Objective
Compared with those without such impairment, middle-aged and older adults with sensory impairment (SI) demonstrate a greater prevalence and severity of depressive symptoms, significantly affecting their mental health. We aimed to develop and validate a depression risk prediction model for middle-aged and elderly individuals with SI.
Methods
Data from the 2018 China Health and Retirement Longitudinal Study were randomly partitioned into training and validation sets at a 7:3 ratio. Within the training set, least absolute shrinkage and selection operator (LASSO) regression analysis and binary logistic regression were used to identify predictor variables, and a risk prediction column‒line graph was subsequently developed, with depression status among middle-aged and elderly individuals with SI as the dependent variable. Predictive performance of the training and validation sets was assessed via receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis.
Results
In total, 5308 middle-aged and older adults with SI were included, with 50.1% (n = 2657) developing depression. Multifactorial logistic regression analysis identified several depression predictors, including sex, education level, place of residence, marital status, self-rated health, life satisfaction, pension insurance status, nighttime sleep duration, functional impairment status, and pain (all P < 0.05), which were incorporated into a column–line graph that demonstrated good consistency and accuracy. The areas under the ROC curves for the predictive models in the training and validation sets were 0.797 (95% CI = 0.783–0.811) and 0.778 (95% CI = 0.755–0.800), respectively. The Hosmer–Lemeshow values were P = 0.176 and P = 0.606 (P > 0.05), and the calibration curves revealed significant agreement between the model and actual observations. ROC and DCA curves indicated good predictive performance for the column‒line graph.
Conclusion
This study presents a reliable, validated, and acceptable predictive model for depression risk in middle-aged and elderly individuals with SI, and the identified predictors have potential applications in public health policy and clinical practice.