Thomsen N O, Jensen C M, Skovgaard N, Pedersen M S, Pallesen P, Soe-Nielsen N H, Rosenklint A
Department of Orthopaedic Surgery and Diagnostic Radiology, Gentofte Hospital, University of Copenhagen, Hellerup, Denmark.
Int Orthop. 1996;20(5):326-9. doi: 10.1007/s002640050087.
By means of Kappa statistics, we calculated the inter- and intra-observer variation in the classification of fractures of the neck of the femur according to Garden's system. Radiographs of 96 consecutive patients were assessed independently by six observers who agreed on classification for only 14 fractures (15%). The level of agreement was poor for the overall classification (Kappa = 0.39). When reducing Garden's system into non-displaced (Stage I and II) and displaced fractures (Stage III and IV) the level of agreement became acceptable (Kappa = 0.68). However, problems remain in distinguishing Stage II and Stage III fractures, and further improvements in the classification system are necessary.
通过Kappa统计方法，我们根据Garden分型系统计算了股骨颈骨折分类中观察者间和观察者内的变异。96例连续患者的X线片由6名观察者独立评估，他们仅对14例骨折（15%）的分类达成一致。总体分类的一致性水平较差（Kappa = 0.39）。当将Garden分型系统简化为无移位骨折（I期和II期）和移位骨折（III期和IV期）时，一致性水平变得可以接受（Kappa = 0.68）。然而，区分II期和III期骨折仍存在问题，因此有必要对分类系统进行进一步改进。
