IMPROVEMENT OF INTEROBSERVER REPRODUCIBILITY OF ADHESION SCORING SYSTEMS

Objective: To compare the interobserver reproducibility of two adhesio n scoring methods, a more comprehensive adhesion scoring method and th e American Fertility Society (AFS) adhesion scoring method. Design: El even endoscopic surgeons independently evaluated and scored 13 surgica l video recordings using both systems. Material and Methods: The stand ardized AFS adhesion scoring method and the more comprehensive adhesio n scoring method were utilized to assess abdominal adhesions. The more comprehensive adhesion scoring method scored 23 individual locations in the abdominal cavity for severity (0, none; 1, filmy, avascular; 2, vascular and/or dense; 3, cohesive) and extent of total area or lengt h (0, none; 1, less than or equal to 25%; 2, 26% to 50%; 3, >50%). Bec ause the best method of representing a composite adhesion score is unc lear, the more comprehensive adhesion scoring method employed two inde pendent methods of determining total score based on the severity and e xtent at each location, either adding severity plus extent or multiply ing severity times extent before summing all 23 locations. Results: Fo r each scoring method, a correlation coefficient was calculated for ea ch of the 55 pairs of surgeons. Significant correlations were identifi ed for all methods (AFS adhesion scoring method: 35 of 55, 64%; more c omprehensive adhesion scoring method: severity plus extent, 49 of 55, 89%; and more comprehensive adhesion scoring method: severity times ex tent, 53 of 55, 96%). When the more comprehensive adhesion scoring met hod is limited to 13 areas in the lower pelvis corresponding to the lo cations represented in the AFS adhesion scoring method, the proportion of significant correlation coefficients become 54 of 55 and 50 of 55 for severity plus extent and severity times extent, respectively. Conc lusion: Although the AFS adhesion scoring method generated significant agreement between pairs of surgeons, less than a third correlated at the 0.7 level. Use of the more comprehensive adhesion scoring method s pecifically demonstrating locations, severity, and extent of adhesions produced a marked increase in reproducibility between surgeon pairs i n scoring pelvic adhesions.