Study design: Prospective study of 3 spine surgeons, 3 spine fellows, 3 nonspine orthopedists, and 12 orthopedic residents classifying 97 thoracolumbar fractures using the Denis, Association for Osteosynthesis (AO), and Thoracolumbar Injury Severity Score (TLISS) systems and reclassifying them 3 months later.
Objective: To compare the reliability of the Denis, AO, and TLISS classification systems and evaluate the skills necessary for their use.
Summary of background data: The Denis and AO systems are the traditional methods of classification of thoracolumbar fractures. The purpose of this study was to evaluate a novel classification system, the TLISS and compare its reliability among observers as compared with the Denis and AO classifications.
Method: Ninety-seven sequential fractures from 1 surgeon's practice at a level 1 trauma center were collected. Twenty-one orthopedic physicians from two area level 1 trauma centers then completed the evaluation of the all fractures. Evaluator experience included staff, spine fellows, and residents. The interobserver and intraobserver reliability were determined.
Result: In the TLISS, subgroups of evaluators, showed variation in reliability as expected with the highest reliability occurring in the senior resident group and attending spine surgeon group. The lowest reliabilities were in the nonspine attending orthopedists and junior residents. In each group, the neurologic status was consistently the category with the highest interobserver and intraobserver reliability. In the Denis and AO classifications, the highest reliabilities were again in the senior residents and spine attendings. The lowest were again in the nonspine attendings and junior residents.
Conclusions: As a management tool, the TLISS seems to be an acceptably reliable system when compared with the Denis and AO systems. There is a base level of knowledge and familiarity necessary for the application of the system at reliable levels.