Establishment and validation of a model for predicting depression risk in stroke patients

Area of Science:

• Neurology
• Psychiatry
• Biostatistics

Background:

• Post-stroke depression is a common complication impacting recovery.
• Accurate prediction of depression risk is crucial for timely intervention.
• Existing prediction models may not integrate diverse clinical and lifestyle factors effectively.

Purpose of the Study:

• To develop and validate a clinically applicable nomogram for predicting depression risk in stroke patients.
• To integrate rehabilitation, biochemical, and lifestyle data for enhanced prediction.
• To provide a tool for early identification and intervention of at-risk individuals.

Main Methods:

• Utilized data from 767 stroke patients.
• Employed Least Absolute Shrinkage and Selection Operator (LASSO) regression to identify key predictors: Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), sleep duration, uric acid, and Triglyceride-Glucose-Body Mass Index (TyG-BMI).
• Constructed a nomogram using multivariable logistic regression and validated it with ROC analysis, calibration curves, Decision Curve Analysis (DCA), and SHapley Additive exPlanations (SHAP).

Main Results:

• The nomogram achieved good discrimination with AUC values of 0.731 (training), 0.663 (testing), and 0.748 (external validation).
• Calibration plots indicated high predictive accuracy, and DCA demonstrated clinical utility.
• SHAP analysis highlighted sleep duration (protective) and ADL (risk) as significant predictors. Lower uric acid and TyG-BMI were associated with increased depression risk.

Conclusions:

• The developed nomogram offers a rapid, cost-effective method for stratifying depression risk in stroke patients using routine clinical data.
• It facilitates early intervention by prioritizing high-risk individuals.
• The model emphasizes the importance of sleep hygiene, metabolic health, and functional rehabilitation in post-stroke mental health.