The Intrarater and Interrater Reliability of the OMT Classification Among Physicians With a Different Background

Background: The Oberg-Manske-Tonkin (OMT) classification established excellent reliability scores in several validation studies. However, one study published in 2022 found much lower scores in a subanalysis of their sample when very simple anomalies were excluded. Our study assessed the reliability of the OMT among physicians with a different background, all involved in congenital hand anomaly care, and analyzed codes with less agreement. Time required for classification was recorded to give an indication on its usability. Methods: One hundred digital cases were classified twice with a minimal 1-month time interval, with the use of the 2020 version of the OMT. Two pediatric hand surgeons, 2 rehabilitation specialists, and 2 plastic surgery residents participated in this reliability analysis. The use of multiple codes was allowed. The intra- and interrater reliability was assessed for all 15 possible rater couples by calculating percentage of agreement. Cohen's kappa was calculated along with a 95% confidence interval. For the analysis of individual codes with less agreement, we calculated positive agreement with the use of a summed agreement table. Time necessary for classification was documented in seconds. Results: The inter- and intrarater agreement was moderate with a mean Cohen's kappa of 0.45 and 0.60 retrospectively. On average, 39 seconds per case were necessary for the first and 24 seconds for the second rating. Background did not influence the level of agreement. Lowest agreement levels (ie, lowest positive agreement) were observed with all the arthrogryposis multiplex congenita subgroups, the "other"subgroups of isolated congenital contractures, syndromic syndactyly, and synpolydactyly. Codes commonly used interchangeably were symbrachydactyly and transverse deficiency and the distinction between these anomalies of only the hand or the entire upper limb; symbrachydactyly and brachydactyly; and camptodactyly and distal arthrogryposis. Conclusions: Our study showed a moderate reliability, emphasizing the complexity of this heterogeneous patient population. Despite its imperfections, the OMT remains the best and most versatile classification tool at hand. Its main purpose may lie in contributing to a universal language for research.