Özet:

Objective: There is a need for a highly technological setting in pediatric intensive care units. In addition to advanced technology, mortality scoring systems are necessary to standardize the care given in these units. Pediatric Risk of Mortality (PRISM) and Pediatric Mortality Index (PIM) are the best known mortality prediction scores. The aim of this study was to determine and validate the scoring systems' effectiveness in order to choose the most suitable one and put it in regular practice in our unit. We included the three hundred eighty nine admissions to the PICU within one year in the study. Material and Methods: Mortality scores for PIM2 and PRISM were calculated prospectively. Discriminatory performance was assessed with the Receiver Operating Curve (ROC). Statistical analysis for calibration of the scoring systems was carried out using the Hosmer and Lemeshow goodness-of-fit test. P values were < 0.05 for all scoring systems. Standardized mortality rates (SMR) were also calculated.Results: All three scoring systems exhibited good discrimination with area under the curve values of 0.84, 0.86 and 0.89 for PIM2, PRISM12 and PRISM24 respectively. SMRs were less than 1.0 for all scoring systems. Conclusion: This result showed that mortality was lower than expected for our unit indicating that these scoring systems have good discriminatory performance, but their calibrations were poor for our unit

Anahtar Kelimeler: