Intraobserver and interobserver reliability of the Sub-axial injury classification evaluation system for sub-axial cervical spine injury

Abstract

Objective To evaluate the intraobserver and intreobserver reliability of the Sub-axial Injury Classification (SLIC) evaluation system for sub-axial cervical spine injury. Methods The preoperative imaging data (posteroanterior and lateral radiographs,CT,MRI) and clinical data which had not been premeasured were evaluated and classified using SLIC system in the 54 patients with sub-axial cervical spine injury who had been treated in our department between May 2008 and April 2010.The same data which had been presented in a different order were evaluated and classified again 4 weeks later in a same manner.The average percentages of intraobserver and interobserver agreements were calculatec.The reliability and reproducibility were quantified using kappa statistics and intraclass correlation coefficients (ICC). Results The SLIC system was found to have intermediate and high interobserver and intraobserver agreements respectively in evaluation of the morphology and neurological function of the injury.The interobserver reliabilities were 62.7％ (Kappa =0.502,ICC =0.581 ) and 69.7％ ( Kappa =0.616,ICC =0.867 ) respectively while the intraobserver reliabilities were71.8％ (Kappa =0.631,ICC =0.746) and 79.6％ (Kappa =0.735,ICC =0.871) respectively.The SLIC system had poor interobserver and intraobserver agreement in evaluation of diseo-ligamentous complex (DLC) injury.The interobserver reliability was 55,8％ ( Kappa =0.317,ICC =0.483 ) and the intraobserver reliability was 68.9％ ( Kappa =0.517,ICC =0.643).The global SLIC evaluation obtained an interobserver reliability of 30.7％ (Kappa=0.218,ICC=0.725) and an intraobserver reliability of 44.5％ (Kappa=0.394,ICC=0.812).For therapy selection,the interobserver reliability was 72.4％ ( Kappa =0.431,ICC =0.578 ) and the intraobserver reliability was 78.8％ ( Kappa =0.581,ICC =0.759). Conclusions The SLIC system has excellent interobserver and intraobserver agreements in evaluation of morphology and neurological function of the injury and in therapy selection as well.This system is not so good for evaluation of DLC injury because MRI imaging cannot make a definite judgment on ligament injury. Key words: Cervical spine; Wounds and injuries; Reliability and validity