Observer reliability in ankle radiographic measurements

We analyzed 50 sets of ankle radiographs to determine the interobserver and intraobserver reliability when obtaining common linear and angular measurements. The radiographs were divided into two groups: one group included 25 normal ankles, and the second group included 25 fractured ankles. Each set of radiographs was evaluated independently by four different observers on two separate occasions under controlled conditions. Six radiographic parameters were measured on all 50 sets of films: syndesmosis A, syndesmosis B, syndesmosis C, the medial clear space, and the talocrural and bimalleolar angles. On the 25 sets of fracture films, four additional measurements of fracture displacement were included: displacement of the medial malleolus (mortise), displacement of the lateral malleolus (AP and lateral), and displacement of the posterior malleolus. Reliability was evaluated with an analysis of variance intraclass correlation coefficient. Among the examiners, 9 of the 10 parameters could be measured reliably. Intraobserver reliability was found to increase with the experience of the examiner.