Frontiers in Public Health (Oct 2013)
Improved Confidence Intervals of a Small Probability From Pooled Testing With Misclassification
Abstract
This article concerns construction of confidence intervals of theprevalence of a rare disease using Dorfman's pooled testingprocedure when the disease status is classified with an imperfectbiomarker. Such an interval can be derived by converting aconfidence interval of the probability that a group is testedpositive. The Wald confidence intervals based on normalapproximation are shown to be inefficient in terms of coverageprobability, even for relatively large number of pools. A fewalternatives are proposed and their performance is investigated interms of coverage probability and length of intervals.
Keywords