Pharmaceutical Applications of a Multi-Stage Group Testing Method

Abstract

An important problem in pharmaceutical research is whether individualtesting of components should be made, or alternatively, if thecomponents should be tested in groups. It is important that theexperiment is economically viable since, for multi-stage procedures, thecost of additional stages must be taken into consideration along withthe cost of testing the mixtures of components. Optimum group sizes arecalculated for two-stage and three-stage members of Li's family ofalgorithms and for row-and-column procedures, along with the minimumnumber of tests required to determine all of the active components.Finally, comparisons are made between the costs of the one-, two- andthree-stage procedures using two different cost functions for the costof testing mixtures of components.