Ta,ble2 presents the goodness-of-fit test performed on the developmental dataset. The p value for this test was 0.883. Since this is large (much greater than 0.05), the model very closely reflects the true mortality experience in the developmental dataset. In the validation dataset, the p value for the goodness-of-fit test was 0.104, suggesting that the model reflected the mortality experience in a group of patients independent of those on whom the model was developed.
The areas under the receiver operating characteristic (ROC) curve for SAPS II were 0.88 (95 per cent confidence interval, 0.87-0.90) in the developmental dataset and 0.86 (95 per cent confidence interval, 0.84-0.88) in the validation dataset. Areas this large are generally acknowledged to constitute excellent discrimination.
Was this article helpful?