Intraobserver and interobserver reliability of recategorized Neer classification in differentiating 2-part surgical neck fractures from multi-fragmented proximal humeral fractures in 116 patients
Self archived versionpublished version
MetadataShow full item record
CitationSumrein, BO. Mattila, VM. Lepola, V. Laitinen, MK. Launonen, AP. Paloneva, J. Jonsson, K. Wolf, O. Ström, P. Berg, H. Felländer-Tsai, L. Mechlenburg, I. Døssing, K. Østergaard, H. Rahnel, T. Märtson, A. (2018). Intraobserver and interobserver reliability of recategorized Neer classification in differentiating 2-part surgical neck fractures from multi-fragmented proximal humeral fractures in 116 patients. Journal of Shoulder and Elbow Surgery, 27 (10) , 1756-1761. 10.1016/j.jse.2018.03.024.
Optimal fracture classification should be simple and reproducible and should guide treatment. For proximal humeral fractures, the Neer classification is commonly used. However, intraobserver and interobserver reliability of the Neer classification has been shown to be poor. In clinical practice, it is essential to differentiate 2-part surgical neck fractures from multi-fragmented fractures. Thus, the aim of this study was to evaluate whether surgeons can differentiate 2-part surgical neck fractures from multi-fragmented fractures using plain radiographs and/or computed tomography (CT).
Three experienced upper limb specialists and trauma surgeons (B.O.S., A.P.L., and V.L.) independently reviewed and classified blinded plain radiographs and CT scans of 116 patients as showing 2-part surgical neck fractures or multi-fragmented fractures. Each imaging modality was reviewed and classified separately by each surgeon, after which each surgeon reviewed both modalities at the same time. This process was repeated by all surgeons after 24 weeks. Intraobserver and interobserver analyses were conducted using Cohen and Fleiss κ values, respectively.
The κ coefficient for interobserver reliability showed substantial correlation (0.61-0.73) and was as follows: 0.73 for radiographs alone, 0.61 for CT scans alone, and 0.72 for radiographs and CT scans viewed together. After 24 weeks, the process was repeated and intraobserver reliability was calculated.The κ coefficient for intraobserver reliability showed substantial correlation (0.62-0.75) and was as follows: 0.62 for radiographs alone, 0.64 for CT scans alone, and 0.75 for radiographs and CT scans viewed together.
Clinicians were able to differentiate 2-part surgical neck fractures from multi-fragmented fractures based on plain radiographs reliably.