PRISM 4-C: An Adapted PRISM IV Algorithm for Children With Cancer.

Area of Science:

• Pediatric critical care medicine
• Biostatistics
• Oncology

Background:

• The Pediatric Risk of Mortality (PRISM) IV score is used to predict mortality in critically ill children.
• Its performance in pediatric cancer patients, a high-risk group, is not well-established.
• Accurate mortality prediction is crucial for clinical decision-making and resource allocation.

Purpose of the Study:

• To evaluate the performance of the PRISM IV score in predicting hospital mortality among pediatric cancer patients.
• To adapt and calibrate the PRISM IV algorithm to improve its predictive accuracy for this specific population.

Main Methods:

• An ambidirectional cohort study included 1003 pediatric patients (500 derivation, 503 validation).
• Logistic regression models were used to test risk variables within PRISM IV components.
• Model performance was assessed using the area under the receiver operating characteristic curve (AUC) and the Hosmer-Lemeshow test.

Main Results:

• The original PRISM IV algorithm showed limited predictive performance (AUC=0.89) in the validation set, predicting only 33 of 65 deaths.
• A modified algorithm demonstrated significantly improved prediction (AUC=0.93), accurately predicting 63.7 of 65 deaths.
• The modified algorithm achieved a standardized mortality rate closer to 1 (1.02 vs. 1.8 for PRISM IV).

Conclusions:

• The standard PRISM IV score is an inadequate predictor of mortality in pediatric cancer patients.
• A modified PRISM IV algorithm significantly enhances the prediction of hospital mortality in this population.
• Further validation of the modified algorithm is warranted for clinical implementation.