Can PIM3 predict mortality adequately for patients admitted to Pediatric Intensive Care Unit in a resource limited setting? A single-center study

Background: In Indonesia the number of critically ill children by far exceeds Pediatric Intensive Care Unit (PICU) capacity. A quantitative scoring system that can be used by clinicians from the initial encounter might be helpful as an objective assessment method to predict patient outcomes, thus aiding clinical decision-making, patient stratification and prioritization. The aim of this study was to assess the performance of the Paediatric Index of Mortality-3 (PIM3) score to predict mortality of critically ill pediatric referral patients that had been accepted and admitted to the PICU of Dr Sardjito General Hospital Yogyakarta. Methods: This retrospective observational cohort study included patients from August 1st, 2022 -January 31st, 2023 who were referred and admitted to the PICU of Dr. Sardjito General hospital, Yogyakarta, Indonesia. The discriminative ability (to differentiate between survivors and non survivors) of PIM3 scoring system was analyzed with the area under the receiver-operating characteristic curve (AUC). Results: A total of 220 cases (median age 64 (1-215) months) were included. The overall mortality rate was 22.1% (49/220). Subjects’ PIM3 score ranged from 0.11 to 98.03. The Standardized Mortality Ratio (SMR) of PIM3 with 95% confidence interval (CI) was 2.23 (1.65–2.90). The AUC with 95%-CI was 0.84 (0.77–0.91, p < 0.001). The difference between observed and predicted mortality by PIM3 was statistically significant, indicating that actual mortality was higher than PIM3 predicted. Conclusion: Among critically ill children who were referred form peripheral hospitals and admitted to PICU at Dr Sardjito General Hospital, PIM3’s discrimination was good, but the calibration was poor, indicating that PIM3 was unfitted as mortality predictor in this study population.