Objective: The aim of this study is to select a better one from five EIA kits made in China for the serological detection of anti-HBe in a national hepatitis B survey.
Methods: Top 5 kinds of kits on the best seller list of China-made anti-HBe EIA detection were selected and each was detected the anti-HBe panel (66 positive and 88 negative) 5 times. The co/s ratio of each test was recorded and transformed by log(co/s+1). ICC and CV, the reliability indices and AUC, pAUC and Se(PR=e), the validity indices were calculated and compared between every two kits.
Results: The indices of ICC and CV of the 5 kinds of kits arranged in the same order of A, E, C, D and B, from the best to the worst. For ICC, there were significantly different between every two of them (Bootstrap method, P < 0.05), except for between kit C and E. The magnitude ofAUCs were from 0.969 to 0.996; the orders of pAUC and Se(FPR=e) indices was C, E, D, A, B, and C, D, E, B, A respectively; both the indices of kit C were significantly higher than others (Bootstrap method, P < 0.05).
Conclusion: All the 5 kinds of kits are excellent and kit C is the best.