This study aimed to assess the validity of PRISM III score in predicting the quality of clinical outcomes in critically ill children admitted in the PICU and to find the correlation of PRISM III scores with duration of PICU stay.Prospective study.Setting: Pediatric Intensive Care Unit of a Tertiary level medical College.Methods: 315 children aged 1month to 15 years admitted in PICU from December 2012 to December 2013 were enrolled in the study.History, physical examination, provisional clinical diagnosis and PRISM III scoring at admission were performed and recorded .PRISM III scoring scale was applied for all patients in PICU in first two hours of PICU admission.Patient followed up and assessed at 6, 12, 24 hours by using clinical parameters of PRISM III.Relevant lab investigations were repeated in indicated patients .Statistical tests like chi square test, Paired and unpaired T tests through SPSS 17 software from the data on Microsoft office excel 2007 were applied.Results: Total 315 children satisfied the study criteria were enrolled in the study.Out of 315 children 193 (61%) were boys .Age and sex(demographic profile) did not influence the outcome significantly.Observed mortality rate was 3.2%.The area under the ROC curve for PRISM III scores at admission was 0.951, at 6 hours it was 0.965, at 12 hours it was 0.998 and at 24 hours it was 1.000 i.e. the maximum area under the ROC curve.At 24 hours for a score with a cut off of 9, the false positivity was zero with 100% sensitivity.According to the results of Hosmer Lemeshow goodness of fit test, the model of PRISM III designed in study has showed to be well fit.In our study the survival rate for a score between 0 to 9 was 100%, between 10 to 19 it was 0% and between 20 to 29 it was 0%, showing that if PRISM III score increases mortality chance increases.Pearson correlation showed a positive correlation of PRISM III score with PICU stay .PRISM III exhibited good discrimination and calibration in our study.Conclusion: PRISM III predicted the outcome in the PICU of our hospital and had good discrimination and calibration.According to Pearsons correlation there was definitely a positive correlation of length of stay with PRISM III scoring.
K et al. (Mon,) studied this question.