Reliability of Hand Fellows’ Measurements and Classifications from Radiographs of Distal Radius Fractures

The inter-rater reliability of classification systems and the direct measurement of fracture displacement was evaluated for two hand fellows in 128 radiographs of distal radius fractures. The fracture classifications rated were the Older, Mayo, AO, McMurtry, Universal and Frykman systems. Measurements of displacement were radial tilt, radial shortening and dorsal angulation on pre- and postreduction films. Intraclass correlation coefficients (ICCs) and kappas, and their associated 95% confidence intervals were calculated. Inter-rater reliability for classification systems was poor, with the exception of the Older system (kappa = 0.73). Prereduction measurement of radial inclination, dorsal angulation and radial shortening had excellent reliability (ICC 0.77). Postreduction films exhibited lower reliability in the same measurements (ICC 0.76). Hand fellows reported inconsistent use of classification systems or radiographical measurements in clinical practice. Further training and/or an increased emphasis on direct measurements, rather than classifications, may be warranted.