Prediction of Acquired Morbidity Using Illness Severity Indices in Pediatric Intensive Care Patients.

Abstract

To assess the ability of two illness severity scores, Pediatric Logistic Organ Dysfunction Score 2 and Pediatric Index of Mortality 3, in predicting PICU-acquired morbidity. Retrospective chart review conducted from April 2015 to March 2016. Single-center study in a multidisciplinary PICU in a tertiary pediatric hospital in Singapore. The study included all index admissions of patients 0-18 years old to the PICU during the study period. None. Three outcomes were assessed at hospital discharge: mortality, survival with new morbidity defined as an increase in the Functional Status Scale score of greater than or equal to 3 points from baseline, and survival without morbidity. Of 577 consecutive admissions, 95 were excluded: 82 readmissions, 10 patients greater than or equal to 18 years old, two patients with missing baseline data, and one transferred to another PICU. Of 482 patients, there were 37 hospital deaths (7.7%) and 39 (8.1%) with acquired new morbidity. Median admission Pediatric Logistic Organ Dysfunction Score 2 and Pediatric Index of Mortality 3 scores differed among the three outcome groups. In addition, differences were found in emergency admission and neurologic diagnosis rates, PICU mechanical ventilation usage rates, and PICU length of stay. The highest proportion of neurologic diagnoses was observed in the new morbidity group. The final model simultaneously predicted risks of mortality, survival with new morbidity and survival without morbidity using admission Pediatric Logistic Organ Dysfunction Score 2 score, admission type, neurologic diagnosis, and preexisting chronic disease. Pediatric Logistic Organ Dysfunction Score 2 was superior to Pediatric Index of Mortality 3 in predicting risks of mortality and new morbidity, as indicated by volume under surface values of 0.483 and 0.362, respectively. Risk of mortality, survival with new morbidity, and survival without morbidity can be predicted simultaneously using admission Pediatric Logistic Organ Dysfunction Score 2, admission type, admission diagnosis, and preexisting chronic disease. Future independent studies will be required to validate the proposed model before clinical implementation.