715 - Use of MRI as an Initial Imaging Modality for Pediatric Head Trauma

Publication Number: 3692.715

Chanelle S. Boadu, Northeastern University/Boston Childrens Hospital, ROCKVILLE, MD, United States; Thea A. Carr, Boston Children's Hospital, North Kingstown, RI, United States; Yachne F. Elysse, Northeastern University/ Boston Children's Hospital, Brockton, MA, United States; Lucia V.. Gizzi, Boston Children's Hospital - Northeastern University, Plymouth, MA, United States; Imma N. Pierre, Northeastern University/Boston Children's Hospital, Randolph, MA, United States; Catherine Coughlin, Boston Children's Hospital, Boston, MA, United States; Michael Monuteaux, Boston Children's Hospital, Boston, MA, United States; Todd W. Lyons, Boston Children's Hospital, Wayland, MA, United States

Background: Magnetic resonance imaging (MRI) is an advanced neuroimaging modality without the radiation risk associated with computed tomography (CT). Little is known about how MRI is being used as the initial imaging modality for head-injured children.

Objective: We sought to investigate trends over time, hospital-level variability, and patient-level predictors of the MRI use as the initial neuroimaging modality for children presenting to the emergency department (ED) with head injuries.

Design/Methods: We performed a retrospective, cross-sectional study of 48 EDs of participating in the Pediatric Health Information Systems (PHIS) database from 2016-2025. We included ED visits for children < 18 years-old with a primary diagnosis of head injury [International Classification of Disease Code Version 10 (ICD-10)]. We excluded transfers, deaths, and visits where the initial imaging modality could not be determined. Our primary outcome was performance of neuroimaging on hospital day 0 or 1. Our secondary outcomes were hospitalization and diagnosis of traumatic brain injury (TBI). We measured predictors of MRI as the initial neuroimaging modality using a multivariate logistic regression model clustered by center.

Results: Over the 10-year study period, 1,677,623 ED visits for head injury met our inclusion and exclusion criteria. Median age was 4.7 [interquartile range (IQR) 2.1, 9.6] years and 40.2% were females. Neuroimaging was performed in 268,880/1,677,623 (16.0%) of all ED visits for head injuries. Over the 10-year study period MRI use as the initial imaging modality increased 239% from 0.18% (2016) to 0.43% (2025) of all ED visits for head injuries. Among ED visits where neuroimaging was performed, MRI was the initial imaging modality for 5,186/268,880 (1.9%). There was profound variability in the use of MRI (Figure) as an initial neuroimaging modality across hospitals. The following factors were associated with MRI as the initial imaging modality: age, sex, race/ethnicity, and the presence of complex chronic conditions (Table). Compared to children with CT as their initial neuroimaging modality, those with MRI were more likely to be admitted to the hospital (CT=16.9%, MRI=48.2, p < 0.05) and be diagnosed with a TBI [CT =6.1%, MRI 22.6%, p < 0.05 ].

Conclusion(s): While only a small percentage of head injured children have an MRI as their initial neuroimaging modality, MRI use increased significantly over the 10-year study period with significant variability between hospitals. We identified patient-level factors associated with MRI use. These data can help inform efforts to optimize MRI use for head-injured children.

Figure: Hospital level variability in the use of computed tomography (CT) and magnetic resonance imaging (MRI) as the initial neuroimaging modality for children presenting to the emergency department with a head injuryFigure: Hospital level variability in the use of computed tomography (CT) and magnetic resonance imaging (MRI) as the initial neuroimaging modality for children presenting to the emergency department with a head injury
• Abbreviations: computed tomography (CT), magnetic resonance imaging (MRI)
• Hospital numbers were randomly based on increasing rates of CT performance
• Denominator for each outcome is total number of head injury visits at that hospital

Table: Multivariable logistic regression model predicting MRI as the initial imaging modality for children with head injury treated in the ED of children's hospitals
n=1,670,996 used in multivariable analysis; n=6627 (0.6%) of full sample (n=1,677,623) dropped due to missing data