Development of decision support system for antibody identification

Choi, Kyung Hwan; Han, Kyou Sup; Han, Bok Yeon; Suh, Jin Tae; Kim, Suhng Gwon; Cho, Han Ik

Abstract

BACKGROUND: Determination of antibody specificity using antigram spread sheet requires experience and knowledge on in vitro characteristics of red cell antibodies, time-consuming, and still subjective to human error. A computer-based antibody identification system was developed to overcome these disadvantages.
METHODS
Decision support system program for antibody identification was designed using Visual Basic 5.0 for Dade Data-cyte Plus. This system integrates the reaction patterns of saline, 37degrees C albumin, antiglobulin, 4degrees C saline enzyme treated and user-defined phases and lists the antibodies according to the probability. 115 irregular antibodies previously confirmed by standard manual method reanalyzed with this program.
RESULTS
In 111 of 115 cases (96.5%), this system produced the same results with the manual identification. In two cases, of not matched 4 cases the computer program suggested additional antibodies and in one case, the computer program detected previous human error. In the other case, antibody identification was possible only after further tests including selective adsorption of multiple antibodies.
CONCLUSION
The decision support system was rapid and easy and showed good concordance rate when compared with manual antibody identificaion results. In addition, human error could be reduced. Decision support system for antibody identification could be used in small blood banks by less experienced staffs.