Teddy Muisyo1, Erika O Bernardo2, Maraya Camazine3, Ryan Colvin4, Kimberly A Thomas3, Matthew A Borgman5, Philip C Spinella3. 1. Department of Pediatrics, Division of Critical Care Medicine, Washington University School of Medicine, 660 S. Euclid Avenue, St. Louis, MO, 63110, USA. Electronic address: tmuisyo@wustl.edu. 2. Department of Pediatrics, Section of Critical Care Medicine, Baylor College of Medicine, Houston, TX, 6621 Fannin Street, Houston, TX, 77030, USA. Electronic address: erika.bernardo@bcm.edu. 3. Department of Pediatrics, Division of Critical Care Medicine, Washington University School of Medicine, 660 S. Euclid Avenue, St. Louis, MO, 63110, USA. 4. Department of Pediatrics, Pediatric Computing Facility, Washington University School of Medicine, 660 S. Euclid Avenue, St. Louis, MO, 63110, USA. Electronic address: r.colvin@wustl.edu. 5. Department of Pediatrics, San Antonio Military Medical Center, 3551 Roger Brooke Dr, Fort Sam, Houston, TX, 78234, USA. Electronic address: matthew.a.borgman.mil@mail.mil.
Abstract
BACKGROUND: In trauma research, accurate estimates of mortality that can be rapidly calculated prior to enrollment are essential to ensure appropriate patient selection and adequate sample size. This study compares the accuracy of the BIG (Base Deficit, International normalized ratio and Glasgow Coma scale) score in predicting mortality in pediatric trauma patients to Pediatric Risk of Mortality III (PRISM III) score, Pediatric Index of Mortality 2 (PIM2) score and Pediatric Logistic Organ Dysfunction (PELOD) score. METHODS: Data were collected from Virtual Pediatric Systems (VPS, LLC) database for children between 2004 and 2015 from 149 PICUs. Logistic regression models were developed to evaluate mortality prediction. The Area under the Curve (AUC) of Receiver Operator Characteristic (ROC) curves were derived from these models and compared between scores. RESULTS: A total of 45,377 trauma patients were analyzed. The BIG score could only be calculated for 152 patients (0.33%). PRISM III, PIM2, and PELOD scores were calculated for 44,360, 45,377 and 14,768 patients respectively. The AUC of the BIG score was 0.94 compared to 0.96, 0.97 and 0.93 for the PRISM III, PIM2, and PELOD respectively. CONCLUSIONS: The BIG score is accurate in predicting mortality in pediatric trauma patients. LEVEL OF EVIDENCE: Level I prognosis.
BACKGROUND: In trauma research, accurate estimates of mortality that can be rapidly calculated prior to enrollment are essential to ensure appropriate patient selection and adequate sample size. This study compares the accuracy of the BIG (Base Deficit, International normalized ratio and Glasgow Coma scale) score in predicting mortality in pediatric traumapatients to Pediatric Risk of Mortality III (PRISM III) score, Pediatric Index of Mortality 2 (PIM2) score and Pediatric Logistic Organ Dysfunction (PELOD) score. METHODS: Data were collected from Virtual Pediatric Systems (VPS, LLC) database for children between 2004 and 2015 from 149 PICUs. Logistic regression models were developed to evaluate mortality prediction. The Area under the Curve (AUC) of Receiver Operator Characteristic (ROC) curves were derived from these models and compared between scores. RESULTS: A total of 45,377 traumapatients were analyzed. The BIG score could only be calculated for 152 patients (0.33%). PRISM III, PIM2, and PELOD scores were calculated for 44,360, 45,377 and 14,768 patients respectively. The AUC of the BIG score was 0.94 compared to 0.96, 0.97 and 0.93 for the PRISM III, PIM2, and PELOD respectively. CONCLUSIONS: The BIG score is accurate in predicting mortality in pediatric traumapatients. LEVEL OF EVIDENCE: Level I prognosis.