Interobserver Reliability of Classification Systems to Rate the Quality of Femoral Neck Fracture Reduction

Abstract

We compared the interrater reliability of five classification systems and surgeons' global impressions of the quality of femoral neck fracture reduction. Six orthopedic trauma surgeons, six orthopedic nontrauma surgeons, and six orthopedic residents from three sites each rated 50 radiographs of postoperative femoral neck fractures fixated with cannulated screws or a sliding hip screw, using their overall impression, the Garden Index, Lowell's criteria, Lindquist and Tornkvist's criteria, the Western Infirmary Glasgow (WIG) angle, and standards established by a working group of orthopedic trauma surgeons. Reliability estimates for the Garden Index, Lowell's criteria, and the working group standards all fell within the range of moderate agreement [intraclass correlation coefficient (ICC) range 0.41-0.48], with no instrument achieving higher reliability than reviewers' overall impressions (ICC 0.49, 95% CI 0.39-0.61). Reviewers reached only fair agreement using Lindquist and Tornkvist's criteria and the WIG angle (kappa = 0.27 and 0.39, respectively). Trauma surgeons consistently achieved higher agreement than did nontrauma surgeons and trainees. Future studies using quality of reduction as an outcome measure or exploring the prognostic importance of reduction quality should measure this with trauma surgeons' overall impression, rather than with less experienced assessors or using one of the alternative instruments.