ORIGINAL ARTICLE Development and Validation of the New International Classification for Scapula Fractures Edward Harvey, MD,* Laurent Audigé, MD,† Dolfi Herscovici, Jr, DO,‡ Julie Agel, MA,§ Jan E. Madsen, MD,¶ Reto Babst, MD,k Sean Nork, MD,** and Jim Kellam, MD†† Objectives: Multiple scapula classification systems exist in the literature and were developed using a consensus approach with one or several experts agreeing on a classification without stringent validation. None have gained widespread acceptance. A decision was made by the OTA classification committee and the AO Classification Advisory Group to collaborate on the development of a new validated classification system capable of addressing the limitations of the existing systems. Methods: A feedback validation process through 4 iterations of revised classifications on radiographs and computed tomography (CT) scans was used. Statistical analyses calculated the proportion of agreement among surgeons and kappa statistics for the assessment of coding reliability. Estimates of classification accuracy were obtained using latent class modeling. Results: Fractures of the scapular neck are rare injuries and were difficult to define and diagnose with kappa values ranging from 0.28 to 0.40. Although fossa fractures could be identified on plain radiographs, specific fracture patterns could only be classified with CT scans. The new classification divides the scapula into 3 segments: fossa, body, and processes. The validation has shown that the classification can be reliable using plain radiographs (kappa 0.66), increasing to kappa of 0.78 when CT scans were added. Conclusions: This basic coding system allows clinicians to describe and classify scapula fractures with a reasonable degree of reliability. This validated classification that has resulted from this process has been accepted by a disparate group of orthopaedic traumatologists as a better option for clinical communication and research documentation. Key Words: classification, scapula fractures (J Orthop Trauma 2012;0:1–6) INTRODUCTION Classification systems are designed to allow clinicians to discuss fractures using a similar language. They are initially meant to be descriptive of fractures and not prognostic of treatment or outcome. Only after the classification has been validated as to reliability, face value, and reproducibility may it then be tested to determine its prognostic value. The most commonly used scapula fracture classification was developed by the OTA Classification Committee in 1996 with a revision a decade later. 1 This and other fracture classifications for the scapula are based on anatomic 2 or subanatomic 3,4 relation- ships. They were all developed using a consensus approach with one or several experts agreeing on a classification without stringent validation. None have gained widespread acceptance. In light of this, a decision was made by the OTA classification committee and the AO Classification Advisory Group (CAG) to collaborate in the development of a new validated system capable of addressing the limitations of the existing systems. The CAG also developed a validation process to assure that any classification meets minimal standards of reliability and validity. 5 The principles of the validation process applied in the development of this new scapula classification system were ease of application by any clinician and a system designed to allow for a reliable and accurate classification of fracture patterns independent of any associated extrascapular (clavicular or ligamentous) injuries. The purpose of this article is to outline this new classification system for fractures of the scapula and to describe its development and validation. 5 MATERIALS AND METHODS Iterative Consensus Review Meeting and Evaluation A study group (scapula classification group) consisting of a methodologist-statistician, a medical coordinator, and 6 experienced orthopaedic traumatologists with upper extremity experience was formed by the AO CAG and the OTA Classification and Outcomes Committee. This group met to review the existing systems, identify their limitations, and then draft a proposal for a new classification. The particular Accepted for publication September 16, 2011. From the *Montreal General Hospital, Montreal, Canada; †AO Clinical Investigation and Documentation, Dübendorf, Switzerland; ‡Florida Ortho- paedic Institute, Tampa, Florida; §Orthopaedic and Sports Medicine, Har- borview Medical Center, Seattle, WA; ¶Institute for Surgical Research, Rikshospitalet The National Hospital, University of Oslo, Oslo, Norway; kDepartment of Surgery, Cantonal Hospital, Lucerne, Switzerland; **Orthopaedic and Sports Medicine, Harborview Medical Center, Seattle, WA; and ††Carolinas Medical Center, Department of Orthopaedic Surgery, Charlotte, NC. Funding was received from the AO Foundation. The authors declare no conflict of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions this article on the journal’s Web site (www.jorthotrauma.com). Reprints: Julie Agel, MA, Research Coordinator, Orthopaedic and Sports Medicine, Harborview Medical Center, Box 359798, 325 Ninth Avenue, Seattle, WA 98104 (e-mail: bagel@u.washington.edu). Copyright © 2012 by Lippincott Williams & Wilkins J Orthop Trauma  Volume 0, Number 0, Month 2012 www.jorthotrauma.com | 1