Reliability and inter-observer reproducibility of GATA classification for spinal tuberculosis

English Keywords:GATA classification for spinal tuberculosis  Reliability  Reproducibility

English Abstract:

  【Abstract】 Objective：To determine the reliability and reproducibility of GATA classification for spinal tuberculosis.Method：A total of 100 cases with spinal tuberculosis were distributed to 4 attending spine surgeons. All surgeons received files including clinical/radiographic data，and an evaluating sheet in terms of GATA classification system.The results were determined by calculating the average percentage of intra-observer and inter-observer agreement to induce the last classifications and the indicators.Reliability and reproducibility were quantified using Kappa coefficients and Spearman rank-order correlation.Moderate consistency was defined as a Kappa value greater than 0.5，moderate positive correlation was defined as Spearman correlation greater than 0.4.Result：Inter-reliability assessed by mean percentage and Kappa coefficiency was 61.92%，0.43 respectively for vertebral collapse，91.83%，0.519 respectively for abscess formation，63.3%，0.317 respectively for instability，60.08%，0.381 respectively for kyphosis，78.25%，0.536 respectively for sagital index，82%，0.531 respectively for disc degeneration，52.33%，0.343 respectively for classification standard.Respective results using Spearman correlation were 0.392，0.526，0.392，0.503，0.569，0.582，and 0.576.Intra-rater mean percentage and Kappa coefficiency were 72.75%，0.574 for vertebral collapse，91%，0.453 for abscess formation，85.25%，0.627 for instability，70.75%，0.523 for kyphosis，88.75%，0.724 for sagital index，81.75%，0.584 for disc degeneration，and 73%，0.591 for classification standard.Respective results using Spearman correlation were 0.748，0.473，0.632，0.66，0.730，0.613，and 0.762.Conclusion：The GATA classification for spinal tuberculosis has good inter-observer reproducibility and low intra-observer reliability.