Reliability and repeatability of a modified thoracolumbar spine injury classification scoring system

On the basis of the Thoracolumbar Injury Classification and Severity Score (TLICS), an modified TLICS classification system was presented, its reliability and repeatability were assessed, and the factors influencing classification consistency were examined.

Five spinal surgeons were chosen at random. The clinical data of 120 patients with thoracolumbar fractures admitted to the Department of Spine Surgery, Ningbo Sixth Hospital from December 2019 to June 2021 were categorized using the modified TLICS system. After 6 weeks, disrupt the order of data again. Using unweighted Cohen's kappa coefficients, the consistency of the modified TLICS system was assessed in five aspects: neurofunctional status, disc injury status, fracture morphology, posterior ligament complex (PLC) integrity, and treatment plan.

In terms of reliability, the average kappa values for the subclasses of the modified TLICS system (neurofunctional status and disc injury status) were 0.920 and 0.815, respectively, reaching the category of complete confidence. Fracture morphology and treatment plan had average kappa values of 0.670 and 0.660, respectively, which were basically reliable. The average kappa value of PLC integrity was 0.453, which belonged to the category of moderate confidence. The average kappa coefficients of each subcategory (neurological status, disc injury status) had excellent consistency, and the kappa values were 0.936 and 0.879, respectively, which belonged to the completely credible category. The kappa values of fracture morphology and treatment plan repeatability were 0.772 and 0.749, respectively, reaching the basic credibility category. PLC integrity repeatability kappa value is low, 0.561, to moderate credibility category.

The modified TLICS system is intuitive and straightforward to understand. The examination of thoracolumbar fracture injuries is more exhaustive and precise, with excellent reliability and repeatability. The examination of neurological status and disc injury status is quite reliable and consistent. The consistency of fracture morphology is slightly poor, which is basically credible; the PLC integrity consistency is poor, reaching a reliability level of moderate, which may be associated with the subjectivity of clinical evaluation of PLC.