Comparison and validation of machine learning-based screening models for elevated depressive symptoms in peritoneal dialysis patients

Figures

• 

  Figure 1. Heatmap of Spearman’s rank correlation coefficients for continuous variables.

• 

  Figure 2. Presentation of the results of the LASSO regression analysis. (A) LASSO regression model screening variable trajectories; (B) LASSO Regression Model Factor Selection: Left dashed line represents the optimal lambda value (lambda.min), while the right dashed line marks the lambda value within one standard error of the optimal (lambda.1se).

• 

  Figure 3. Comprehensive evaluation of ML models across Training, Validation, and External cohorts. Row 1: ROC curves with AUC values and 95% confidence intervals. Row 2: Calibration curves with Brier scores. Row 3: Decision curve analysis (DCA) with Treat All and Treat None reference lines. (A) Training ROC, (B) validation ROC, (C) external ROC, (D) training calibration, (E) validation calibration, (F) external calibration, (G) training DCA, (H) validation DCA, and (I) external DCA.

• 

  Figure 4. SHAP of the model. (A) Characteristic attributes in SHAP. The abscissa is the SHAP value, and each line denotes a feature. Higher eigenvalues are indicated by red dots, and lower eigenvalues are indicated by blue dots. (B) Importance ranking plot of features of the XGBoost model. (C,D) Interpretability analysis of 2 independent samples.

• 

  Figure 5. R Shiny-Based Online Visualization Tool for predicting depression risk in peritoneal dialysis patients based on the XGBoost model.

• 

  Figure 6. Stratified prevalence of elevated depressive symptoms in end-stage renal disease patients undergoing peritoneal dialysis. (A) Stratified by age (<60 years vs. ≥60 years); (B) stratified by history of peritonitis (no vs. yes); (C) stratified by peritoneal dialysis vintage (PDV, <5 years vs. ≥5 years); (D) stratified by history of catheter-related complications (no vs. yes); (E) stratified by Social Support Rating Scale (SSRS) score (<20 vs. 20–29 vs. ≥30); (F) stratified by Self-Rating Anxiety Scale (SAS) score (<50 vs. ≥50). Bar heights indicate the prevalence of elevated depressive symptoms (SDS ≥ 53) within each subgroup. Numbers above bars indicate the exact prevalence and event count (events/total). Statistical significance of between-group differences is assessed by chi-square test and indicated by p-values in each subfigure.