A new nomogram for the individualized prediction of children's mortality risk in pediatric intensive care unit.

OBJECTIVE

We developed a new nomogram for the prediction of mortality risk in children in pediatric intensive care units (PICU).

METHODS

We conducted a retrospective analysis using the PICU Public Database, a study that included a total of 10,538 children, to develop a new risk model for mortality in children in the intensive care units (ICU). The prediction model was analyzed using multivariate logistic regression with predictors including age and physiological indicators, and the prediction model was presented as a nomogram. The performance of the nomogram was evaluated based on its discriminative power and was internally validated.

RESULTS

Predictors contained in the individualized prediction nomogram included the neutrophils, platelets, albumin, lactate, oxygen saturation (<0.1). The area under the receiver operating characteristic (ROC) curve for this prediction model is 0.7638 (95% CI: 0.7415-0.7861), which has effective discriminatory power. The area under the ROC curve of the prediction model in the validation dataset is 0.7404 (95% CI: 0.7016-0.7793), which is still effectively discriminative.

CONCLUSION

The mortality risk prediction model constructed in this study can be easily used for individualized prediction of mortality risk in children in pediatric intensive care units.