Skip to main content
Fig. 2 | BMC Medical Research Methodology

Fig. 2

From: Mitigating selection bias in organ allocation models

Fig. 2

Observed vs. Predicted Survival in Testing Cohort. Time-dependent calibration of A) the modified pre-transplant outcome model, B) the existing pre-transplant LAS model, C) the modified post-transplant outcome model, and D) the existing post-transplant LAS model, in the testing cohort. Smooth, solid lines represent predicted survival probabilities; points with vertical error bars represent observed Kaplan-Meier estimates with their corresponding 95% confidence intervals. Estimates were plotted every 30 days to ease plot readability. Three risk groups are shown: low-risk/best survival (darkest lines), medium-risk/intermediate survival (medium-shaded lines), and high-risk/worst survival (lightest lines). A vertical, dashed, red line is placed at one year post-waitlist registration for reference

Back to article page