Christensen F, Søballe K, Ejsted R, Luxhøj T
J Bone Joint Surg Br. 1986 Aug;68(4):614-5. doi: 10.1302/0301-620X.68B4.3733840.
The reliability of the Catterall grouping of Perthes' disease was examined by determining the agreement between pairs of observers using weighted kappa statistics. Anteroposterior and lateral radiographs of 100 hip joints were grouped independently by four experienced observers. There was a low, and in our opinion, unacceptable degree of inter-observer agreement even when Groups 2 and 3 were combined.
通过使用加权kappa统计量来确定观察者之间的一致性，对佩特兹病的卡特拉尔分组的可靠性进行了检验。100个髋关节的前后位和侧位X线片由四位经验丰富的观察者独立分组。即使将2组和3组合并，观察者间的一致性程度仍然很低，而且在我们看来是不可接受的。
