Predictive features analysis and nomogram construction for predicting depression in elderly patients.

Introduction: In elderly populations, depression is highly prevalent among those with chronic diseases and cognitive impairment, leading to distress, disability, and poor medical outcomes. With the aging of the population, the prevalence of geriatric depression is rising rapidly. The Comprehensive Geriatric Assessment (CGA), a multidimensional approach, evaluates medical, psychological, and functional capacities to identify highrisk individuals and may be correlated with depression in the elderly.

Methods: From 2021 to 2023, a total of 219 geriatric patients were recruited. These patients were divided into two groups: a modeling group of 153 patients and a validation group of 66 patients. We collected patients' basic information and CGA results and analyzed them using univariate and multivariate regression. Independent variables influencing depression were identified.

Results: Multivariate regression analyses revealed that several factors had an impact on depression in these patients, including social support level (SSRS), Pain, Anxiety, Basic Activities of Daily Living (BADL) and Gender. By integrating these factors into the nomogram, we found good predictive performance in the training set (AUC 0.867, 95% CI: 0.799-0.936) and in the test set (AUC 0.724, 95%CI:0.5919-0.894). The calibration and discrimination accuracy of the nomograms for predicting depression risk in the elderly were satisfactory, and the decision curve analysis demonstrated significant clinical utility.

Discussion: The model demonstrated robust performance in our study and may constitute a valuable tool for clinical screening.