Development and validation of a nomogram for predicting depression risk in older adults with cognitive impairment

To develop and validate a clinically practical nomogram for predicting depression risk in older adults with cognitive impairment.

Cross-sectional data from the 2018 Chinese Longitudinal Healthy Longevity Survey (CLHLS) included 3254 adults ≥65 years. Cognitive impairment was defined by Mini-Mental State Examination (MMSE) scores ≤24; depression by Center for Epidemiologic Studies Depression Scale (CESD) ≥10. Participants were divided into training (n = 2278) and test (n = 976) sets. Variables included demographics, lifestyle, and medical history. LASSO regression (10-fold cross-validation) identified key predictors; multivariate logistic regression constructed the nomogram. Model performance was evaluated using area under the curve (AUC), calibration curves, and internal/external validation.

Eight independent predictors were identified: younger age (OR = 0.981), living alone (OR = 1.584), smoking history (OR = 0.650), poor sleep quality (e.g., "very poor" OR = 50.326), infrequent outdoor activities (OR = 2.272), rare reading (OR = 5.558), limited TV/radio exposure (OR = 1.905), and cataract history (OR = 1.587). The nomogram demonstrated strong discrimination (AUC training: 0.807; test: 0.716) and good calibration (Hosmer-Lemeshow P = 0.491, internal validation; bootstrap MAE = 0.011 confirming stability).

This nomogram integrates modifiable lifestyle and clinical factors for depression risk stratification in cognitively impaired older adults. Its simplicity and accuracy suit resource-limited settings, enabling early intervention and personalized care. Limitations include the cross-sectional design, reliance on a single data source which may affect external validity beyond the tested sample, the assumption of linearity for some lifestyle factors, and the absence of a control group of depressed individuals without cognitive impairment for comparative analysis.