OBJECTIVETo assess the performance of pediatric risk of mortality (PRISM), pediatric index of mortality 2 (PIM2) in predicting mortality in critically ill pediatric patients via a prospective study.

METHODThe outcome and the variables required to calculate PRISM and PIM2 were collected. The efficiency of PRISM and PIM2 in differentiation between death and survival by calculating the area under the receiver operating characteristic curve (ROC). Calibration across deciles of risk was evaluated using the Hosmer-Lemeshow goodness-of-fit χ(2) test.

RESULTA total of 412 critically ill pediatric patients transferred to Hunan Children's Hospital during August 1, 2012 and May 31, 2013 were enrolled in the study, and more than two-thirds of the children were suffering from respiratory and miscellaneous diseases; 45 (10.9%) inter-hospital transport of critically ill pediatric patients died at the time of hospitalization. The expected number of deaths were 45.01 by PRISM, and the expected number of deaths were 44.99 by PIM2. The expected mortality rate was 10.9% for PRISM or PIM2. The standardized mortality rate was 1.000 (0.710-1.290) for PRISM and 1.000 (0.710-1.290) for PIM2. The Hosmer-Lemeshow test gave a chi-square of 8.75 (P=0.364) for PRISM and 22.75 (P<0.05) for PIM2, PRISM had better fitting with the actual mortality than PIM2. The area under the receiver-operating characteristics (ROC) curve (95% confidence intervals, CI) were 0.829 (0.768-0.890) for PRISM and 0.758 (0.667-0.849) for PIM2.

CONCLUSIONAlthough the PIM2 test is less well calibrated overall, both PRISM and PIM2 can offer a good capacity for discriminating between survivors and moribund patients. The good performance of PRISM and PIM2 are demonstrated in predicting mortality probability in critically ill pediatric patients.