# Pooled Prevalence Calculator – Demonstration analyses

## Background

For these demonstration analyses (except where otherwise
stated), a hypothetical pooled testing strategy for the estimation of
prevalence of Hendra virus in fruit bats was used. The data was based on real
data for Hendra virus testing in the little red flying fox (*Pteropus scapulatus*)
in Queensland
during the period 1996 – 1999 (H. Field, pers com).
During this period, 162 samples were tested from little red flying foxes, with
22 samples positive, for an estimated prevalence of 13.6% (95% CI: 8.7 – 19.8%).
For these analyses, pool sizes and numbers of pools were calculated to provide
95% confidence of estimating a true prevalence of 14 % with a desired precision
of ± 5.5% for the various estimation methods, corresponding to the estimated
prevalence and confidence interval for the original (unpooled) data. The most
frequent result from simulation studies was then used to estimate the
prevalence and confidence interval for each scenario.

## Contents

- Prevalence estimation for fixed pool size and perfect test
- Prevalence estimation for fixed pool size and known test sensitivity and specificity
- Prevalence estimation for fixed pool size and uncertain test sensitivity and specificity
- Prevalence estimation for variable pool size and perfect test
- Prevalence estimation using Bayesian estimation and a Gibbs sampler
- Sample size for fixed pool size and perfect test
- Sample size for fixed pool size and known test sensitivity and specificity
- Sample size for fixed pool size and uncertain test sensitivity and specificity
- Simulation for fixed pool size and perfect test
- Simulation for fixed pool size and known test sensitivity and specificity
- Simulation for fixed pool size and uncertain test sensitivity and specificity
- Simulation for variable pool size and perfect test
- Pooled testing for demonstration of freedom
- Prevalence estimation using Bayesian estimation for unpooled testing and one test
- Prevalence estimation using Bayesian estimation for unpooled testing and two tests
- Parameterise Beta probability distributions from mode and 5/95 percentiles
- Summarise Beta probability distributions for specified alpha and beta parameters