The programs available on this site provide a suite of utilities for estimating prevalence and to assist in the design of sampling and pooling strategies for the estimation of disease prevalence from the testing of pooled (or individual) samples. The various programs have been implemented in the statistical software environment "R", with your web browser used to pass input values to the program and to display the output of the resulting analysis.
To use the programs, enter the desired input values in each text box on the input screen and click on submit. Example values are already displayed in the input boxes, but can be over-written with your own values. Alternatively, you can use the default values to experiment and see how the program works. All input values are checked before processing, to ensure that they are valid and within the ranges specified in the accompanying description. An informative error message is displayed if invalid input values are entered, and the progam will not run until these values are corrected. Input values for parameters that can be represented as percentages, proportions or probabilities (prevalence, sensitivity, specificity, confidence limits) must be entered as proportions (decimal numbers between 0 and 1). Similarly, output results for these parameters wil also be expressed as proportions.
Error checking is limited to the numerical validity of input values, for example by checking valid ranges or that counts are input as integers. It is therefore possible to enter inappropriate or unlikely values which could result in non-sensical output. It is the user's responsibility to ensure that input values are appropriate and that results are meaningful.
Some of the utilities included in this web-site use simulation to estimate parameter values or to evaluate proposed testing strategies. These simulations require multiple iterations (runs) of the model to produce the required result. For the Bayesian analysis, many iterations are required to allow the model to converge on the true parameter values, and additional iterations are then required for inference about the value. In most cases a minimum of 10,000 iterations is recommended (in some cases 20,000 - 50,000 may be better), with 2,000 - 5,000 iterations discarded to allow for convergence of the model. For other programs, 5,000 - 10,000 iterations is usually sufficient. Because of the large number of iterations required, some of these simulations make take several minutes (or longer) to complete.
Output from each program is returned to your web-browser in a standard format. This starts with a brief description of the analysis/method used, followed by a summary of input values and finally a summary table of results. For most programs, graphical representations and text files of detailed results are available for most analyses by clicking on the appropriate icon in the results table. Text files of results can be either opened directly in MS Excel or saved on your PC in a tab-delimited format.
A summary description and brief help is provided on the input page for each program, with a more detailed description provided for all the programs in this user guide.
|3||Bayesian vs Frequentist methods|
|4||Fixed pool size and perfect tests|
|5||Fixed pool size and known Se & Sp|
|6||Fixed pool size and uncertain Se & Sp|
|7||Variable pool size and perfect tests|
|8||Pooled prevalence using a Gibbs sampler|
|9||True prevalence using one test|
|10||Estimated true prevalence using two tests with a Gibbs sampler|
|11||Estimation of parameters for prior Beta distributions|
|12||Sample size for fixed pool size and perfect test|
|13||Sample size for fixed pool size and known test sensitivity and specificity|
|14||Sample size for fixed pool size and uncertain test sensitivity and specificity|
|15||Simulate sampling for fixed pool size|
|16||Simulate sampling for variable pool sizes|
|18||Pooled prevalence estimates are biased!|