P. Aggarwal1, R. Kim2, R. Kuhia1, Z. Zhao3, S.M. Fiore4, D. Chesler4, M. Egnor4, A. Yurovsky5, H. Hsieh6 1Stony Brook University Medical Center, Renaissance School Of Medicine, Stony Brook, NY, USA 2University Of California – San Francisco, Department Of Pediatrics, San Francisco, CA, USA 3University Of Washington, Department Of Neurologic Surgery, Seattle, WA, USA 4Stony Brook University Medical Center, Department Of Neurosurgery, Stony Brook, NY, USA 5Stony Brook University Medical Center, Department Of Biomedical Informatics, Stony Brook, NY, USA 6Stony Brook University Medical Center, Department Of Surgery, Stony Brook, NY, USA
Introduction: Traumatic brain injury (TBI) is a leading cause of preventable morbidity and mortality in pediatrics. Pediatric TBI is known to occur in a bimodal distribution with respect to age, but how this distribution is accounted for from an analytical perspective varies considerably. This research aims to provide data-driven age categories for analysis of TBI in children.
Methods: A retrospective chart review was conducted on all pediatric TBI patients admitted to a Level I Pediatric Trauma Center between January 2015-June 2022. Data pertaining to demographics, mechanism of injury, clinical characteristics, and injury outcomes were collected. For patients under two years old, age was collected in months; for patients over two years old, age in years was reported. Breakpoints were estimated to calculate slope inflection points in age versus cumulative TBI frequency. As proof of concept, the predicted age groups were then used to compare patient demographics, injury mechanisms, and outcomes. Data was analyzed using chi-squared tests and fisher’s exact testing.
Results: In a sample of 596 patients, we redemonstrate the known bimodal distribution of age with respect to TBI. We highlight three breakpoints in pediatric TBI by age: X1 = 0.82 ± 0.08 years, X2 = 3.45 ± 1.25 years, X3 = 11.59 ± 0.45 years, with m0 = 0.255 ± 0.028, m1 = 0.046 ± 0.010, m2 = 0.026 ± 0.003, m3 = 0.076 ± 0.005. The groups of infant:[0,1), toddler:[1,4), child:[4,12), and adolescent[12,18) were used for bivariate analyses. Patient racial and ethnic background (p = 0.013), parent marital status (p = 0.013), rates of motor-vehicular accidents (p < 0.001), falls (p < 0.001), bicycle or motorcycle or rollerblade injuries (p < 0.001), sport-injuries (p < 0.001), other or unknown mechanisms of injury (p < 0.001), and discharge disposition (p = 0.008) varied by patient age group. Rates of violence (p = 0.053) and Glasgow Coma Score (p = 0.804) did not vary significantly by patient age group.
Conclusion: We identify four distinct age groups in pediatric TBI. For practicality in analysis, we suggest using the age ranges of infants under 1 year old, toddlers between 1-3 years old, children between 4-11 years old, and adolescents over 12 years old. While further research should be conducted to confirm the validity of this model at state and national levels, we provide informed benchmarks for TBI analyses as demonstrated by proof-of-concept analyses that highlight differences in the racial/ethnic makeup of patients, mechanisms of injury, and injury outcomes.