Related Experiment Video
Updated: Apr 16, 2026
Minimally Invasive Treatment for Thoracolumbar Burst Fracture Using Sagittal Alignment Screws and A Trauma Reduction Device
Published on: November 8, 2024
Interobserver agreement in retrospective chart reviews for factors associated with cervical spine injuries in
Cody S Olsen1, Nathan Kuppermann, David M Jaffe
1The Division of Critical Care, Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, UT.
Insights
Trained reviewers and physicians showed moderate to excellent agreement on most factors predicting cervical spine injuries (CSIs) in children. This suggests retrospective chart data is reliable for identifying potential CSI predictors.
Area of Science:
- Pediatric Emergency Medicine
- Trauma Surgery
- Radiology
Background:
- Cervical spine injuries (CSIs) in children are a significant concern in emergency medicine.
- Accurate identification of predictive factors is crucial for timely diagnosis and management.
- Retrospective chart reviews are often used to study rare pediatric injuries, but data validity is a concern.
Purpose of the Study:
- To assess the interobserver agreement between trained chart reviewers and independent physician reviewers.
- To determine the reliability of retrospectively collected data for identifying predictors of pediatric CSIs.
- To validate data used in multicenter studies on pediatric cervical spine trauma.
Main Methods:
- A multicenter retrospective chart review of children (<16 years) with cervical spine radiography was conducted.
- Trained reviewers abstracted data, and independent physicians reviewed a sample of medical records.
- Interobserver agreement was quantified using percent agreement and weighted kappa (κ) statistics.
Main Results:
- Moderate to excellent agreement (κ > 0.4) was found for most candidate predictors, including altered mental status, neck tenderness, and loss of consciousness.
- High-risk injury mechanisms, such as diving and motor vehicle collisions, showed near-perfect agreement.
- Fair agreement was observed for clotheslining mechanisms and substantial face injuries.
Conclusions:
- Retrospectively assessed variables for predicting CSIs in pediatric blunt trauma patients demonstrate at least moderate interobserver agreement.
- The findings suggest that retrospective chart data is sufficiently valid for identifying potential predictors of pediatric cervical spine injuries.
- This supports the use of such data in research aimed at improving the diagnosis and management of pediatric CSIs.
Objectives:
The objective was to describe the interobserver agreement between trained chart reviewers and physician reviewers in a multicenter retrospective chart review study of children with cervical spine injuries (CSIs).
Methods:
Medical records of children younger than 16 years old with cervical spine radiography from 17 Pediatric Emergency Care Applied Research Network (PECARN) hospitals from years 2000 through 2004 were abstracted by trained reviewers for a study aimed to identify predictors of CSIs in children. Independent physician-reviewers abstracted patient history and clinical findings from a random sample of study patient medical records at each hospital. Interobserver agreement was assessed using percent agreement and the weighted kappa (κ) statistic, with lower 95% confidence intervals.
Results:
Moderate or better agreement (κ > 0.4) was achieved for most candidate CSI predictors, including altered mental status (κ = 0.87); focal neurologic findings (κ = 0.74); posterior midline neck tenderness (κ = 0.74); any neck tenderness (κ = 0.89); torticollis (κ = 0.79); complaint of neck pain (κ = 0.83); history of loss of consciousness (κ = 0.89); nonambulatory status (κ = 0.74); and substantial injuries to the head (κ = 0.50), torso/trunk (κ = 0.48), and extremities (κ = 0.59). High-risk mechanisms showed near-perfect agreement (diving, κ = 1.0; struck by car, κ = 0.93; other motorized vehicle crash, κ = 0.93; fall, κ = 0.92; high-risk motor vehicle collision, κ = 0.89; hanging, κ = 0.80). Fair agreement was found for clotheslining mechanisms (κ = 0.36) and substantial face injuries (κ = 0.40).
Conclusions:
Most retrospectively assessed variables thought to be predictive of CSIs in blunt trauma-injured children had at least moderate interobserver agreement, suggesting that these data are sufficiently valid for use in identifying potential predictors of CSI.
