Analysis of interobserver and intraobserver variability in the diagnosis and treatment of SLAP tears using the Snyder classification

Background: Superior labral anterior posterior lesions are a relatively rare entity, and classification as a basis for selection of treatment has remained a point of controversy. Hypothesis: There will be substantial interobserver and intraoloserver variability in the diagnosis and treatment of superior labral anterior posterior tears by experienced arthroscopic specialists. Study Design: Cohort study (diagnosis), Level of evidence, 2. Methods: Compact discs containing 22 video vignettes of approximately 15 seconds duration were sent to the membership of the Arthroscopy Association of North America, American Shoulder and Elbow Society, and AOSSM. Each surgeon was asked to review the vignettes, classify the superior labral anterior posterior lesion type, and provide a treatment recommendation for each vignette. Seventy-three expert surgeons responded to the solicitation with a completed analysis. The same CD-ROM was re-sent to each of these 73 surgeons at a minimum of 12 months after the first viewing to obtain data on intraobserver reliability. Seventeen of the 73 surgeons returned this second CD-ROM with a complete analysis. Demographic data were also obtained from each surgeon. Multivariable logistic regression analysis was used to analyze the data, and 95% confidence intervals were established for each superior labral anterior posterior type (I-IV) with regard to diagnosis and treatment decision. Results: The 22 vignettes analyzed by 73 surgeons resulted in 1606 responses. Several significant trends were noticed with regard to diagnosis and treatment from the responses: (1) surgeons had difficulty distinguishing type III lesions from type IV lesions, (2) the treatment of type III lesions is much more variable than that of any other subtype, and (3) surgeons had difficulty distinguishing normal shoulders from type 11 superior labral anterior posterior tears. No relationship was identified between correct treatment decisions based on diagnosis and any of the demographic factors analyzed. Our analysis of intracbserver variability showed only moderate agreement. The analysis of interobserver variability improved significantly when the diagnoses were analyzed based on treatment decision. Conclusions: There is substantial interobserver and intraobserver variability among experienced shoulder arthroscopic specialists with regard to diagnosis and treatment of superior labral anterior posterior tears. Intraobserver agreement using the Snyder classification indicated only moderate agreement. Analysis of interobserver agreement based on treatment decisions results in superior concordance among experienced surgeons for the diagnosis of superior labral anterior posterior lesions.