Clinical epidemiologic evaluation of BSE screening tests in Japan.

Abstract

To address the public concerns provoked by the first incidence of bovine spongiform encephalopathy (BSE) in Japan, the BSE screening tests in Japan are evaluated in use of modeling analysis in evidence-based diagnosis. Under the assumptions based on epidemiological statistics such as the annual number of screened cattle with 1,227,385, the annual incidence of BSE infection with four, and the sensitivity of 99.0% for both primary and secondary tests, it was estimated that, the current threshold of cut-off for the BSE positive would have 0.119 false negatives per year. The decrease of the sensitivity of ELISA down to 90.0% resulted in the increase up to 0.792 false negatives per year. Even with the 90.0% sensitivity, shifting the cut-off point from the current level to the best one remarkably reduced the false negatives per year down to 0.0004. Regarding false positives, with 99.7% specificity for both ELISA and the confirmatory tests revealed the risk of 0.03 false positives per year, while the cut-off shifting that can best minimize false negatives largely increased the false positives up to 11,013. Although it is confirmed the possibility of false negatives is very low, the current method of screening can be further improved by shifting the decision level of cut-off to define the BSE "positive". Such an improvement, however, raises an issue of trade-off : the less false negatives, the more false positives. We believe our approach can help the public perception of an optimum decision-making for BSE screening, considering the trade-off.