Evaluating presence/absence of target microbes in microbiological tests

A typical qualitative microbiological method performance (collaborative) st udy gathers a data set of responses about a test for the presence or absenc e of a target microbe. We developed 2 models that estimate false-positive a nd false-negative rates. One model assumes a constant probability that the tests will indicate the target microbe is present for any positive concentr ation in the test portion. The other model assumes that this probability fo llows a logistic curve. Test results from several method performance studie s illustrate these estimates.