Determining optimal screening policies using decision trees and spreadsheets.

Abstract

The determination of the proportion of population to be screened, so that overall costs are minimized, is treated. A decision tree approach is used, based on knowing the probabilities and costs of all possible outcomes. This involves specifying models that estimate the prevalences in the screened and non-screened groups for various screening levels. The generalized implementation on a microcomputer spreadsheet that is presented permits sensitivity and optimality to be explored for different parameter conditions. A numerical example illustrates the results.