Fig 1.
Flow chart of the study.
Table 1.
Baseline Characteristics of the Derivation, Internal, and External Validation Cohorts.
Table 2.
Univariate and multivariate Cox proportional hazards regression analysis for variables selection.
Table 3.
Predictors of Mortality in Advanced Schistosomiasis and Model Performance in the Derivation Cohort.
Fig 2.
Model 1–4 Reclassification Performance Across Different Cohorts.
Model 1 included Age, Frequency of ascites, Course of disease, ALT, AST, ALP, DBil, ALB, AFP and HBsAg. Model 2 included DBil, ALT, AST, ALB, HBsAg, ALP and AFP. Model 3 included Age, Frequency of ascites, Course of disease, HBsAg, AFP. Model 4 included Age, Course of disease, Frequency of ascites, ALT, AST, ALP, ALB and DBil. a. ROC Curves and AUC for the Derivation Cohort. Models 1 through 4 had AUCs of 0.759, 0.697, 0.693, and 0.751, respectively. b. ROC Curves and AUC for the Internal Validation Cohort. Models 1 through 4 had AUCs of 0.763, 0.703, 0.694, and 0.755, respectively. c. ROC Curves and AUC for the External Validation Cohort. Models 1 through 4 had AUCs of 0.774, 0.704, 0.704, and 0.763, respectively.
Table 4.
Evaluating the Predictive Accuracy of Models for Mortality in Internal and External Validation Cohorts.
Fig 3.
Risk Index for Post-Discharge Patients.
a. Point values for each variable. To calculate a patient’s total risk score, sum the points assigned to each variable value, which will indicate the predicted 6-year mortality risk. b. Estimated 6-Year Mortality Risk After Discharge.
Fig 4.
Comparison of Predicted and Observed 6-Year Mortality Risk Based on the Risk Index in Internal and External Validation Cohorts.
Data are shown for the internal validation cohort (n = 1400) and the external validation cohort (n = 1208). Error bars reflect 95% confidence intervals. Risk scores ranged from 1 to 15 in both cohorts.