Pooled or group testing is a testing strategy where samples from a number of individuals are aggregated into a single sample (or pool), which is then tested for the disease or agent of interest. Pooled testing strategies have been proposed for identification of infected individuals (classification), for demonstration of herd or area infection status and for estimation of prevalence of infection.
The utilities provided on this web-site are designed to assist with the estimation of prevalence from the testing pof pooled samples. When used for this purpose, individual samples are aggregated into pools for testing. Individual-level prevalence is then estimated based on the number of individuals represented in each pool and the test result for that pool. Estimation methods are provided for both fixed and variable pool size and to adjust estimates for imperfect sensitivity and specificity of the test used.
Pooled testing has some significant advantages over individual testing. In particular, pooled testing provides:
- significant cost savings, particularly where the cost of collecting additional samples is small relative to the cost of the test or if prevalence is low;
- increased precision of estimates compared to individual testing where the same number of tests is undertaken, particularly if prevalence is less than about 30%;
- increased precision of estimates compared to individual testing where test sensitivity and/or specificity are less than 1; and
- reduced bias in estimates when assumed values for sensitivity and specificity are not equal to the true values.
Pooled testing strategies also have a number of disadvantages, including:
- logistical requirements and cost for pooling of samples, either in the laboratory or in the field;
- the requirement of some strategies for re-testing of samples from positive pools either as smaller pools or individually, increasing requirements for sample volumes, storage and handling;
- the possible effect of dilution on test performance, particularly test sensitivity cannot be ignored, and any appropriate for imperfect sensitivity should include consideration of any effect of dilution on test sensitivity;
- estimates may be biased to a variable degree, depending on prevalence, number of pools and pool size; and
- The methods depend on a number of major assumptions, and may result in biased estimates if the assumptions are violated. These assumptions include that:
- the outcome is assumed to follow a binomial distribution - clustering or overdispersion of the positive outcome can result in substantial bias in the resulting estimate;
- sample size is small compared to the population;
- the health status of each individual is independent of the status of others, both within and between pools;
- test sensitivity and specificity are 100% (for some methods);
- samples are assumed to be mixed homogeneously in the pools and any sub-samples taken for testing are equally representative of all of the individuals contributing to each pool; and
- all pools represent the same number of individuals (except for the variable pool-size method).
Additional utilities are also provided for the estimation of true prevalence based on testing of individual samples (unpooled) using either one or two tests of inperfect and uncertain sensitivity and specificity. These utilities use Bayesian methods to account for uncertainty about the true values of sensitivity and specificity for the test(s) used.