Statistical methods and applications of bioassay.

Abstract

A bioassay experiment is designed to estimate the relative potency of a test compound (T) to a standard compound (S). The relative potency of T to S is defined as the dose of T that produces the same biological response as does a unit dose of S (1). Traditionally, relative potency has been estimated for a test compound that is a dilution of the standard. Currently, two different compounds that have parallel dose-response curves over some range of doses are rountinely assayed. This case is known as a comparative assay. A bioassay experiment may be either quantal or quantitative, direct or indirect. If the response measure is binary, the assay is said to be quantal (2). Otherwise, it is quantitative. In a direct assay the threshold dose required for response is determined for each experimental unit. Thus, the observed data are dose units. In an indirect assay the experimental unit receives one or more specified doses of the preparation and the observed data may be either quantal or quantitative responses. Depending on the experimental design, several dose levels of T and S are given to the same or different experimental units. The former experiment is called a crossover, and the latter, a paraUel group or completely randomized design. In this chap ter we discuss only indirect quantita tive assays. We first review the elementary theory developed over the past forty years for the univariate