Application of Geometric Mean Criteria for Anti- Blood Group Antibody Proficiency Testing in the Korean External Quality Assurance Program

Abstract

Background: The geometric mean criteria (GC) is an alternative for assessing proficiency testing (PT) acceptability. A recent study applied GC to the PT results provided by the College of American Pathologists. We assessed the feasibility of using GC for anti-blood group antibody titration testing (ABT) in the Korean PT program. Methods: The results of the ABT performed in 2019 were reviewed using GC. GC was calculated as geometric mean (GM)±multiples of geometric standard deviation (GSD). The number of acceptable results obtained using GC was compared to that of the conventional mode criteria (MC, mode±1 second). Only the results with 30 or more peer group responses were included in the analyses. Results: A total of 27 PT results (anti-A: 13, anti-B: 14) were analyzed. The acceptable proportions from MC were 82.9%?100.0% for anti-A and 76.2%? 100.0% for anti-B. The GC criteria yielded acceptable results of 46.9%?97.6% (1 GSD), 88.6%?100.0% (2 GSD), and 97.3%?100.0% (3 GSD) for anti-A. For anti-B, 1 GSD, 2 GSD, and 3 GSD criteria resulted in 44.7%?90.6%, 90.6%? 100.0%, and 97.4%?100.0%, respectively. In general, acceptable results using MC were found to be distributed between 1 GSD and 2 GSD. Conclusions: The GC can be used as an alternative assessment criterion with a more robust statistical rationale. While conventional MC struggles with representing the central tendency of data, GC provide better visualization of the central tendency.