Moderate Interrater Reliability in the Diagnosis of Scaphoid Waist Fractures.

Abstract

Background Conventional radiographs have been shown to yield unreliable results in classifying scaphoid fractures. Computed tomography (CT) has been claimed to be the tool of choice in determining the treatment as well as fracture displacement. Purpose The purpose of the study was to examine the interrater reliability and intrarater reproducibility in the decision-making of the treatment of scaphoid waist fractures. Patients and Methods Fifty-one CT scans of scaphoid waist fractures were utilized. Seven orthopaedic surgeons with a particular interest in hand surgery independently scrutinized the scans classifying each in undisplaced, < 2 mm displaced, or > 2 mm displaced, and suggested a treatment of immobilization in cast or screw fixation. The Fleiss' and Cohen's kappa values using SPSS (Statistical Package for Social Science) version 24 were calculated and interpreted according to Landis and Koch. Results The kappa value representing interrater reliability when choosing between operative or nonoperative treatment was 0.58. Interrater reliability of the distinction between < 2 mm displaced or > 2 mm displaced fractures was 0.61. On average 79.5% of the fractures were suggested treated nonoperatively and 20.5% operatively. Overall, intrarater reproducibility was 0.75 when classifying between < 2 mm displaced or > 2 mm displaced fractures. When choosing between operative or nonoperative treatment, intrarater reproducibility was 0.69. Conclusion Moderate interrater reliability was found when choosing between nonoperative and operative treatment. The use of CT showed substantial reliability in the distinction between < 2 mm displaced and > 2 mm displaced fractures. Intrarater reproducibility was substantial when classifying between < 2 mm displaced and > 2 mm displaced fractures as well as when choosing between operative or nonoperative treatment. Level of Evidence This is a Level III study.