Analysis of interobserver reliability for radiographic staging of coxarthrosis and indexes of acetabular dysplasia: A preliminary study

Background: We are planning a multicenter survey on coxarthrosis and acetabular dysplasia in Japan. To collect reliable data, we performed a preliminary study to elucidate the observer agreement on assessment items. Methods: We collected radiographs of hip joints in eight patients with various findings of coxarthrosis. Twelve registered orthopedic specialists evaluated them regarding the roentgenographic stage of coxarthrosis and five indexes of acetabular dysplasia (acetabular angle, center-edge angle, acetabular roof obliquity, acetabular head quotient, approximate acetabular quotient). To assess observer agreement, we calculated the value of the kappa statistic for stages and the coefficient of variation for the indexes. The same 12 specialists then assessed the coxarthritis stage on the same radiographs 1 month after the first evaluation based on our own descriptions of the roentgenographic stages. Results: For the first evaluation of the roentgenographic stage, the value of the kappa statistic was 0.448; and for the second evaluation it was 0.600. The results of the coefficient of variation for the indexes of acetabular dysplasia, ranked in ascending order, were as follows: acetabular angle, acetabular head quotient, acetabular roof obliquity, center-edge angle, approximate acetabular quotient. Conclusions: For the upcoming multicenter survey, clear descriptions of the stages of coxarthrosis and selection of appropriate indexes can be helpful for collecting dependable results.