Predicting inpatient mortality in pediatric traumatic brain injury: insights from a national database.

PURPOSE: The purpose of this study was to determine factors significantly associated with mortality and length of stay (LOS) in admissions to the pediatric intensive care unit (PICU) for traumatic brain injury (TBI). METHODS: A cross-sectional, retrospective cohort study that identified PICU admissions with TBI from forty-nine hospitals in the USA using the Pediatric Health Information System database from 2016 to 2021. Univariable analyses comparing those who did and did not experience mortality were performed. The following regression analyses were conducted: logistic regression with mortality as dependent variable; linear regression with LOS as the dependent variable; logistic regression with mortality as the dependent variable but only included patients with cerebral edema; and linear regression with LOS as the dependent variable but only included patients who survived. From the regression analysis for mortality in all TBI patients was utilized to develop a mortality risk score. RESULTS: A total of 3041 admissions were included. Those with inpatient mortality (18.5%) tended to be significantly younger (54 vs. 92 months, p < 0.01), have < 9 pediatric Glasgow Coma Scale on admission (100% vs. 52.9%, p < 0.01) and more likely to experience acute renal, hepatic and respiratory failure, acidosis, central diabetes insipidus, hyperkalemia, and hypocalcemia. Regression analysis identified that pediatric Glasgow Coma Scale, alkalosis and cardiac arrest significantly increased risks of mortality. The TBI mortality risk score had an area under the curve of 0.89 to identify those with mortality; a score of 6 ≤ was associated with 88% mortality. CONCLUSION: Patients admitted to the PICU with TBI have 18.5% risk of inpatient mortality with most occurring the first 48 h and these are characterized with greater multisystem organ dysfunction, received medical and mechanical support. TBI mortality risk score suggested is a practical tool to identify patients with an increase likelihood to die.

Randomized Controlled Trials in Pediatric Cardiology: A Power Struggle?

Sample size and statistical power are often limited in pediatric cardiology studies due to the relative infrequency of specific congenital malformations of the heart and specific circulatory physiologies. The primary aim of this study was to determine what proportion of pediatric cardiology randomized controlled trials achieve an 80% statistical power. Secondary aims included characterizing reporting habits in these studies. A systematic review was performed to identify pertinent pediatric cardiology randomized controlled trials. The following data were collected: publication year, journal, if "power" or "sample size" were mentioned if a discrete, primary endpoint was identified. Power analyses were conducted to assess if the sample size was adequate to demonstrate results at 80% power with a p-value of less than 0.05. A total of 83 pediatric cardiology randomized controlled trials were included. Of these studies, 48% mentioned "power" or "sample size" in the methods, 49% mentioned either in the results, 12% mentioned either in the discussion, and 66% mentioned either at any point in the manuscript. 63% defined a discrete, primary endpoint. 38 studies (45%) had an adequate sample size to demonstrate differences with 80% power at a p-value of less than 0.05. A majority of these are not powered to reach the conventionally accepted 80% power target. Adequately powered studies were found to be more likely to report "power" or "sample size" and have a discrete, primary endpoint.