Teachability and reliability of a new classification system for lower cervical spinal injuries.

Abstract

Reliability and validation study. The objective of this study is to evaluate a new lower cervical spine injury classification system and assess its reliability, teachability, and clinical applications. The recently proposed Cervical Spine Injury Severity Score (CSISS) morphologically describes lower cervical spine injuries and grades them on a score of 1 to 20 depending on the integrity of the 4 columns that make up the cervical spine. Early data suggests that this classification system is both reliable and reproducible. Reliability data from additional institutions and data exploring teachability of this classification system is not available. Fifteen subjects (12 residents and 3 attendings trained in the management of spinal trauma) reviewed radiographs and CT scans of 50 patients and scored them according to the CSISS. Six residents scored the patients 1 month before an instructional lecture given by the senior author and then again immediately following the lecture to assess teachability of the new classification system. All subjects then reviewed the films a final time 1 month later to assess both intraobserver and interobserver reliability. The patients' scores were also analyzed in conjunction with their clinical treatment. Interobserver reliability overall was excellent (0.975) with junior residents performing similarly to those with more extensive training. Intraobserver reliability was also excellent overall (0.983). Teachability scores improved in the ability to score all 4 columns. Furthermore, this classification system was a fair overall predictor of surgical candidates as a score of 7 predicted 19 out of 26 surgical patients (76% sensitivity, 100% specificity). The CSISS is a useful new adjunct in the treatment and classification of lower cervical spine injuries. The system is reliable, reproducible, and teachable. It is clinically useful for all levels of orthopedic training and experience.