Evaluation of interobserver agreement in Albertoni's classification for mallet finger☆

Objective To measure the reliability of Albertoni's classification for mallet finger. Methods Agreement study. Forty-three radiographs of patients with mallet finger were assessed by 19 responders (12 hand surgeons and seven residents). Injuries were classified by Albertoni's classification. For agreement comparison, lesions were grouped as: (A) tendon avulsion; (B) avulsion fracture; (C) fracture of the dorsal lip; and (D) physis injury-and subgroups (each group divided into two subgroups). Agreement was assessed by Fleiss's modification for kappa statistics. Results Agreement was excellent for Group A (k = 0.95 (0.93-0.97)) and remained good when separated into A1 and A2. Group B was moderate (k = 0.42 (0.39-0.44)) and poor when separated into B1 and B2. In the Group C, agreement was good (k = 0.72 (0.70-0.74)), but when separated into C1 and C2, it became moderate. Group D was always poor (k = 0.16 (0.14-0.19)). The general agreement was moderate, with (k = 0.57 (0.56-0.58)). Conclusion Albertoni's classification evaluated for interobserver agreement is considered a reproducible classification by the method used in the research.