Validity of pediatric risk of mortality score in prediction of mortality in North Indian pediatric intensive care unit

Background: The pediatric risk of mortality (PRISM) score allows assessment of the severity of illness and mortality risk adjustment in heterogeneous groups of the patients in an objective manner. It has been developed and validated mostly at pediatric intensive care unit (PICUs) of developed countries with very few reports from India. Objectives: This study was planned to evaluate the usefulness of the PRISM score and to correlate it with the mortality in patients admitted to PICU of a northern Indian PICU. Material and Methods: A prospective, observational study was conducted between January 2012 and June 2013, during which a total of 150 consecutive cases were enrolled for the study. PRISM score was calculated within 24 h of their admission and receiver operating characteristic (ROC) curve was used to establishing the validity of the PRISM score for predicting the mortality. Results: Overall mortality was 12.5%, and the PRISM score for survivors versus non-survivors was 7.5878 ± 5.032 versus 20.63 ± 3.41, respectively. No difference was seen between the observed and expected value of mortality calculated from PRISM score (Z = 0.467-1.521, p = 0.291–0.64), thereby establishing the validity of the PRISM score in predicting the mortality in our PICU (ROC curve - area under the curve 0.934). The observed and expected values of the mortality were more comparable for lower PRISM scores (0-15) and a PRISM score of 13.5 has the highest sensitivity and specificity. Conclusion: We concluded that PRISM score is a valid parameter to predict the mortality among the ICU patients in Indian set ups. Key words: Mortality, Pediatric intensive care unit, Pediatric risk of mortality score, Receiver operating curve