Table 3.
Likelihood and discriminative properties of different models
| Model | BIC | R2 Nagelkerke | Hosmer-Lemeshow test | AUC (95% CI) | Mean predicted probability cases/non-cases | Integrated discrimination improvement (p value) |
|---|---|---|---|---|---|---|
| ESPOIR | 265.89 | 0.17 | 0.95910 | 0.75 (0.70 to 0.84) | 0.20/0.10 | |
| mESPOIR | 239.59 | 0.30 | 0.54740 | 0.82 (0.76 to 0.89) | 0.29/0.09 | −0.101 (p=0.3) |
| SONORA | 264.55 | 0.00 | 0.00709 | 0.76 (0.68 to 0.83) | 0.33/0.22 | −0.008 (p=0.5) |
| SONORArecalculated | 262.76 | 0.00 | 0.00019 | 0.76 (0.69 to 0.83) | 0.06/0.04 | 0.079 (p=0.7) |
| BEST | 1234.82 | 0.00 | 0.00023 | 0.73 (0.65 to 0.81) | 0.90/0.78 | −0.010 (p=0.5) |
| BESTrecalibrated | 449.76 | 0.00 | 0.00008 | 0.73 (0.64 to 0.80) | 0.61/0.40 | −0.102 (p=0.4) |
| ASPIRE ESR | 269.79 | 0.00 | 0.00635 | 0.54 (0.45 to 0.64) | 0.17/0.15 | 0.091 (p=0.7) |
| ASPIRE ESRrecalibrated | 464.10 | 0.00 | 0.00085 | 0.54 (0.46 to 0.64) | 0.45/0.43 | 0.082 (p=0.6) |
| ASPIRE CRP | 268.91 | 0.00 | 0.24880 | 0.62 (0.54 to 0.69) | 0.21/0.18 | 0.067 (p=0.6) |
| ASPIRE CRPrecalibrated | 574.03 | 0.00 | 0.00052 | 0.62 (0.41 to 0.60) | 0.59/0.52 | 0.035 (p=0.6) |
Integrated discrimination improvement is calculated as (average predicted probability cases—average predicted probability controls)new model—(average predicted probability cases—average predicted probability controls)original ESPOIR model. A positive value indicates that the new model provides an improvement over the original ESPOIR model.
AUC, area under the receiver operating characteristic curve; BIC, Bayesian Information Criterion; CRP, C reactive protein; ESR, erythrocyte sedimentation rate.