Intraobserver and Interobserver Agreement in the Measurement of Displaced Humeral Medial Epicondyle Fractures in Children

Author(s): Pappas N, Lawrence JT, Donegan D, Ganley T, Flynn JM

Source: J Bone Joint Surg 92A: 322-7, 2010.


There is a general consensus in the orthopedic literature that fractures of the medial epicondyle in the pediatric population should be treated non-operatively if displaced less than 2mm and treated operatively if displaced greater than 5mm. From these numbers one can infer that the measurements of displacement should be precise and reproducible to ensure proper treatment. Accordingly, the authors undertook this study with the notion of testing the accuracy by which the amount of displacement can be measured. They sought to examine the intra- and inter- observer agreement of displacement amongst orthopedic surgeons in various levels of their career.

Utilizing digital radiography measurements, 38 patients with 39 fractures were examined for displacement on frontal radiographs. The correlation coefficient and the percentage of ratings that differed by 2mm were recorded. The authors felt that disagreement of less than 10% (utilizing >2mm as the point of difference) was acceptable.

Results showed that intra- and inter-observer agreement amongst the junior residents was not acceptable. Amongst the junior attendings and fellows, the correlation coefficient was higher in each case, with the fellow and the junior attending disagreeing most at 10% of the time. The overall interobserver correlation coefficient was .80, with 54% of the reviewers disagreeing with each other’s measurements (range 23-86%).

The authors conclude that there exists substantial intra- and inter-observer disagreement over the measuring of displacement of medial epicondyle fractures by plain radiographs. Attendings and fellows have less disagreement, but still enough to call into question the decision to treat the fracture operatively versus non-operatively. The authors did agree that anteroposterior radiographs should be the standard radiographic view utilized to measure displacement and that the amount of fracture gap should be standardized. It should reflect the distance of the most widely displaced portion of the fracture fragment and the bone from which it displaced.  This study highlights the limitations of measurements of three-dimensional distances using unilplanar radiographs.