The Correction for Bias in Prevalence Estimation with Screening Tests