This is patch to fix a bug affecting PoolPrev. The bug affected the maximum likelihood estimates (MLE) and likelihood ratio confidence intervals (LR-CIs) of prevalence when the default Jeffrey's prior was being used. The bug would usually make the MLE and LR-CIs much closer to the Bayesian estimates than they should have been. As both sets of estimates are valid, the results will still have been approximately correct.
This patch also includes an option, replicate.poolscreen (default to FALSE), for PoolPrev. This options changes the way the likelihood ratio confidence intervals are calculated. With replicate.poolscreen = TRUE PoolPrev will more closely reproduce the results produced by Poolscreen. We believe that our implementation of these intervals is more correct so would recommend that users continue to use the default (replicate.poolscreen = FALSE), but this option may be helpful for those who are trying to compare results across the two programs.