Neer's classification system: a critical appraisal.
The purpose of this study was to examine the interobserver agreement using Neer's classification system for fractures of the proximal humerus. A preliminary series of 28 fractures of the proximal humerus admitted to the hospital over a 5-year retrospective period was independently assessed by two radiologists and two orthopedic surgeons. The degree of agreement between paired observations was determined using a kappa statistic. The observations were assessed according to the number of fracture segments, as well as site of fracture segment. Most of the patients assessed were multisegment tuberosity fractures (n = 18). Results showed that the overall agreement between pairs of observers was 65% (average kappa = 0.45). It was also shown that, within the tuberosity group, the percentage agreement was 50%, with an average kappa = 0.35, indicating only fair interobserver agreement. Clinical implications of these findings are such that, as major surgical decisions are made dependent on Neer's classification, closer evaluation techniques--such as the computerized tomography scan and magnetic resonance imaging--should be used to classify the fracture segments accurately. This is particularly true with the fractured tuberosity subgroup.