Interobserver variability of injury severity assessment in polytrauma patients: does the anatomical region play a role?

Background: The Abbreviated Injury Scale (AIS) and Injury Severity Score (ISS) are widely used to assess trauma patients. In this study, the interobserver variability of the injury severity assessment for severely injured patients was analyzed based on different injured anatomical regions, and the various demographic backgrounds of the observers.
Methods: A standardized questionnaire was presented to surgical experts and participants of clinical polytrauma courses. It contained medical information and initial X-rays/CT-scans of 10 cases of severely injured patients. Participants estimated the severity of each injury based on the AIS. Interobserver variability for the AIS, ISS, and New Injury Severity Score (NISS) was calculated by employing the statistical method of Krippendorff's α coefficient.
Results: Overall, 54 participants were included. The major contributing medical specialties were orthopedic trauma surgery (N = 36, 67%) and general surgery (N = 13, 24%). The measured interobserver variability in the assessment of the overall injury severity was high (α _ISS : 0.33 / α _NISS : 0.23). Moreover, there were differences in the interobserver variability of the maximum AIS (MAIS) depending on the anatomical region: α _{head and neck} : 0.06, α _thorax : 0.45, α _abdomen : 0.27 and α _extremities : 0.55.
Conclusions: Interobserver agreement concerning injury severity assessment appears to be low among clinicians. We also noted marked differences in variability according to injury anatomy. The study shows that the assessment of injury severity is also highly variable between experts in the field. This implies the need for appropriate education to improve the accuracy of trauma evaluation in the respective trauma registries.