Related Experiment Video
Updated: May 11, 2026
Murine Model of Controlled Cortical Impact for the Induction of Traumatic Brain Injury
Published on: August 16, 2019
Pediatric traumatic brain injury: precision risk assessment models and an online calculator for enhanced patient care
Foad Kazemi1, Elena Ghotbi2, Julian L Gendreau1
1Department of Neurosurgery, Johns Hopkins University School of Medicine, Baltimore, MD, United States.
Insights
This study developed a reliable risk stratification tool for pediatric traumatic brain injury (TBI) patients. The tool accurately predicts outcomes like extended length of stay and discharge type, aiding clinical decisions.
Area of Science:
- Pediatric neurosurgery
- Public health
- Clinical decision support
Background:
- Traumatic brain injury (TBI) presents a significant public health challenge.
- Accurate risk assessment for pediatric TBI is crucial for optimizing patient care and resource allocation.
Purpose of the Study:
- To develop and validate a reliable risk stratification tool for pediatric TBI patients.
- To support clinicians and multidisciplinary teams in making informed decisions.
Main Methods:
- Retrospective review of 2954 pediatric TBI cases (age ≤18) using electronic health records.
- Analysis incorporated demographics, Social Deprivation Index (SDI), and Injury Severity Scores (ISS).
- Multivariate regression with backward elimination, assessing model discrimination (AUC) and calibration (Spiegelhalter's z-test).
Main Results:
- Predictive models demonstrated good discrimination with AUCs ranging from 0.87 to 0.89 for extended LOS, non-routine discharge, ICU/OR transfer, and direct ED discharge.
- Models showed adequate fit, with Spiegelhalter's z-test p-values > 0.05.
- An open-access online calculator was developed based on these models.
Conclusions:
- The developed predictive models and online calculator enable precise, individualized risk assessments for pediatric TBI patients.
- These tools enhance neurosurgical decision-making and promote high-value care.
- Integration of readily accessible ED data facilitates improved patient management.
Background:
Traumatic brain injury (TBI) is a significant public health challenge demanding extensive medical resources. Accurate, individualized risk assessments for extended length of stay (LOS), non-routine discharge, ICU/OR transfers, and direct ED discharges are crucial for optimizing patient care, prompting the authors to develop a reliable risk stratification tool to support clinicians and multidisciplinary care teams.
Methods:
A retrospective review of electronic health records was conducted to identify pediatric TBI cases (age ≤18) using ICD-10 codes based on the modified CDC framework. Data on demographics, neighborhood socioeconomic disadvantage (assessed using the Social Deprivation Index [SDI]), and injury severity (assessed using Injury Severity Scores [ISS]) were collected. The backward elimination method was employed in the multivariate regression analysis to achieve the most parsimonious model. Model discrimination and calibration were assessed using the area under the receiver operating characteristic curve (AUC) and Spiegelhalter's z-test, respectively.
Results:
A total of 2954 TBI cases were identified with an average age of 7.05 years. Of these, 28.4 % had extended LOS, 8.3 % had a non-routine discharge, 23.4 % required ICU/OR transfer, and 52.3 % were discharged directly from the ED; respective predictive models achieved AUCs of 0.89, 0.87, 0.89, and 0.88, demonstrating good discrimination. All the referenced models had a Spiegelhalter z-test p-value greater than 0.05, indicating an adequate fit. All models were used to develop an open-access online calculator available at: https://jhpedsnsgy.shinyapps.io/JHPedsNSGY/.
Conclusions:
By integrating readily accessible data in the ED, these predictive models and the online calculator empower clinicians to deliver precise, individualized risk assessments, enhance neurosurgical decision-making, and improve high-value care for pediatric TBI patients.
