Table 4 Summary of model performance for the developed model by apparent, split sample, bias-corrected internal validation and external validation estimates
Model performance measure | Number of studies (%, 95% CI) | Median [p25-p75] | Range |
|---|---|---|---|
R-squared reported | 11 (9%, 95% CI: 5–16%) | ||
Apparent* | 1 | 0.06 [-] | - |
Split sample | 6 | 0.31 [0.15–0.34] | 0.12 to 0.45 |
Bias corrected | 5 | 0.35 [0.11–0.64] | 0.01 to 0.89 |
External validation* | 1 | 0.88 [-] | - |
C-statistic reported | 116 (97%, 95% CI: 92–99%) | ||
Apparent | 52 | 0.84 [0.76–0.91] | 0.63 to 0.99 |
Split sample | 43 | 0.81 [0.74–0.88] | 0.61 to 1.00 |
Bias corrected | 53 | 0.84 [0.75–0.90] | 0.59 to 1.00 |
External validation | 16 | 0.81 [0.73–0.88] | 0.67 to 0.96 |
Calibration-in-the-large reported | 10 (8%, 95% CI: 4–15%) | ||
Apparent | 10 | 0.01 [-0.02–0.18] | -0.26 to 1.82 |
Split sample* | 1 | -0.01 [-] | - |
Bias corrected | 0 | - | - |
External validation | 0 | - | - |
Calibration slope reported | 9 (8%, 95% CI: 4–14%) | ||
Apparent | 9 | 0.99 [0.98 to 1.00] | -0.22 to 1.02 |
Split sample* | 1 | 1.02 [-] | - |
Bias corrected | 0 | - | - |
External validation | 0 | - | - |