Reliability of a novel classification system for thoracolumbar injuries: the Thoracolumbar Injury Severity Score

Abstract

Study design: Prospective study of 5 spine surgeons rating 71 clinical cases of thoracolumbar spinal injuries using the Thoracolumbar Injury Severity Score (TLISS) and then re-rating the cases in a different order 1 month later.

Objective: To determine the reliability of the TLISS system.

Summary of background data: The TLISS is a recently introduced classification system for thoracolumbar spinal column injures designed to simplify injury classification and facilitate treatment decision making. Before being widely adopted, the reliability of the TLISS must be studied.

Methods: A total of 71 cases of thoracolumbar spinal trauma were distributed on CD-ROM to 5 attending spine surgeons, including clinical/radiographic data, details of the TLISS, and a scoring sheet in which cases would be scored using the system. The surgeons were later assigned the task with the cases reordered. Intraobserver and interobserver reliability was calculated for TLISS components, total score, and surgeon's treatment decision using the Cohen unweighted kappa coefficients and Spearman rank-order correlation.

Results: Interrater reliability assessed by generalized kappa coefficients was 0.33 +/- 0.03 for injury mechanism, 0.91 +/- 0.02 for neurologic status, 0.35 +/- 0.03 for posterior ligamentous complex status, 0.29 +/- 0.02 for TLISS total, and 0.52 +/- 0.03 for treatment recommendation. Respective results using the Spearman correlation were 0.35 +/- 0.04, 0.94 +/- 0.01, 0.48 +/- 0.04, 0.65 +/- 0.03, and 0.51 +/- 0.04. Surgeons agreed with the TLISS recommendation 96.4% of the time. Intrarater kappa coefficients were 0.57 +/- 0.04 for injury mechanism, 0.93 +/- 0.02 for neurologic status, 0.48 +/- 0.04 for posterior ligamentous complex status, 0.46 +/- 0.03 for TLISS total, and 0.62 +/- 0.04 for treatment recommendation. Respective results using the Spearman correlation were 0.70 +/- 0.04, 0.95 +/- 0.02, 0.59 +/- 0.05, 0.77 +/- 0.04, and 0.59 +/- 0.05.

Conclusions: The TLISS has good reliability and compares favorably to other contemporary thoracolumbar fracture classification systems.