新AOSpine胸腰椎损伤分类的可信度与可重复性分析

目的评估新AOSpine胸腰椎损伤分类的可信度与可重复性。方法对45例胸腰椎损伤患者进行回顾性分析。由4名脊柱外科医生对患者医的影像学资料（腰椎正侧位X线片及CT扫描）根据新AOSpine胸腰椎损伤分类方法进行分型。1个月后重新打乱顺序由该4名医生再次进行分型，收集分型结果，分别作可信度和可重复性分析，计算Kappa值检验其一致性。结果4名观察者均采用新AOSpine胸腰椎损伤分类方法进行分型，结果的一致性为68.9%（60.0%～75.6%），可信度kappa值为0.644（0.543～0.720），属基本可信。同1名观察者先后两次分型结果的可重复性为76.7%（73.3%～80.0%），kappa值为0.732（0.691～0.771），属基本可信。结论新AOSpine胸腰椎损伤分类方法具有较高的可信度与可重复性，然而其临床意义还亟需大样本长期随访研究。

Objective To evaluate the interobserver reliability and intraobserver reproducibility of the new AOSpine Thoracolumbar Spine Injury Classification System. Methods 45 patients with thoracolumbar injury were identified and reviewed. Anteroposterior and lateral radiographs and computed tomography scans were taken in all patients. The classification was based on the new AOSpine Thoracolumbar Spine Injury Classification System by four spine surgeons independently. The procedure was repeated with an interval of one month. The interobser verand intraobserver agreement of the new AOSpine Thoracolumbar Spine Injury Classification System were determined by the Kappa coefficient(κ).Results All patients were classified by four observers with the new AOSpine Thoracolumbar Spine Injury Classification System. The interobserver reliability was substantial, with 68.9% full interobserver agreement (60.0%-75.6%), with a Kappa value of 0.644 (0.543-0.720). The intraobserver reproducibility was also substantial with were 76.7%(73.3%-80.0%) full intraobserver reproducibility, with a Kappa value of 0.732 (0.691-0.771).Conclusion The new AOSpine Thoracolumbar Spine Injury Classification System has substantial interobserver reliability and intraobserver reproducibility. However its clinical value still needs be evaluated further by a large number of patients with long follow-up.