Interobserver Agreement in Pediatric Cervical Spine Injury Assessment Between Prehospital and Emergency Department Providers

Background: Investigators have derived cervical spine injury (CSI) decision support tools from physician observations. There is a need to demonstrate that prehospital emergency medical services (EMS) providers can use these tools to appropriately determine the need for spinal motion restrictions and make field disposition decisions.

Objectives: The objective was to determine the interobserver agreement between EMS and emergency department (ED) providers for CSI risk assessment variables and overall gestalt for CSI in children after blunt trauma.

Methods: This was a planned, substudy of a four-site, prospective cohort of children < 18 years transported by EMS to pediatric EDs for evaluation of CSI after blunt trauma. Inclusion criteria were trauma team activation and/or EMS-initiated spinal motion restriction. Exclusion criteria were penetrating trauma, transfer to another facility for definitive care, state custody, or substantial language barrier. For each eligible child, the transporting EMS provider and treating ED provider independently recorded their clinical assessment for CSI. This included mechanism of injury and patient history and physical examination findings. We assessed each paired variable for interobserver agreement between EMS and ED provider using kappa (κ) analysis. We considered variables with κ lower confidence interval values ≥0.4 to have moderate or better agreement.

Results: We obtained 1,372 paired observations for 29 variables. After finding prevalence and observer bias were adjusted for, all variables achieved moderate to better agreement including eight variables previously shown to be independently associated with CSI in children: diving mechanism, high-risk motor vehicle collision, altered mental status, focal neurologic findings, neck pain, torticollis, substantial torso injury, and predisposing medical condition. EMS and ED providers, however, showed less than moderate agreement for their overall gestalt for CSI in children. Of note, both EMS and ED providers did not assess for neck pain, inability to move the neck, and/or cervical spine tenderness in more than 10% of study patients.

Conclusions: Emergency medical services and ED providers achieved at least moderate agreement in the assessment of CSI risk factors in children after blunt trauma. However, EMS and ED providers did not achieve moderate agreement on gestalt for CSI and some risk factors went unassessed by providers. These findings support the development of a pediatric CSI risk assessment tool for EMS and ED providers to reduce interventions for those children at very low risk for CSIs while still identifying all children with injury.