Scapula fractures: Interobserver reliability of classification and treatment

Objectives: There is substantial variation in the classification and management of scapula fractures. The first purpose of this study was to analyze the interobserver reliability of the OTA/AO classification and the New International Classification for Scapula Fractures. The second purpose was to assess the proportion of agreement among orthopaedic surgeons on operative or nonoperative treatment. Design: Web-based reliability study. Setting: Independent orthopaedic surgeons from several countries were invited to classify scapular fractures in an online survey. Participants: One hundred three orthopaedic surgeons evaluated 35 movies of three-dimensional computerized tomography reconstruction of selected scapular fractures, representing a full spectrum of fracture patterns. Main Outcome Measurements: Fleiss kappa (κ) was used to assess the reliability of agreement between the surgeons. Results: The overall agreement on the OTA/AO classification was moderate for the types (A, B, and C, κ = 0.54) with a 71% proportion of rater agreement (PA) and for the 9 groups (A1 to C3, κ = 0.47) with a 57% PA. For the New International Classification, the agreement about the intraarticular extension of the fracture (Fossa (F), κ = 0.79) was substantial and the agreement about a fractured body (Body (B), κ = 0.57) or process was moderate (Process (P), κ = 0.53); however, PAs were more than 81%. The agreement on the treatment recommendation was moderate (κ = 0.57) with a 73% PA. Conclusions: The New International Classification was more reliable. Body and process fractures generated more disagreement than intraarticular fractures and need further clear definitions.