STATISTICAL APPROACHES TO PHARMACODYNAMIC MODELING - MOTIVATIONS, METHODS, AND MISPERCEPTIONS

We have attempted to outline the fundamental statistical aspects of ph armacodynamic modeling. Unexpected yet substantial variability in effe ct in a group of similarly treated patients is the key motivation for pharmacodynamic investigations. Pharmacokinetic and/or pharmacodynamic factors may influence this variability. Residual variability in effec t that persists after accounting for drug exposure indicates that furt her statistical modeling with pharmacodynamic factors is warranted. Fa ctors that significantly predict interpatient variability in effect ma y then be employed to individualize the drug dose. In this paper we ha ve emphasized the need to understand the properties of the effect meas ure and explanatory variables in terms of scale, distribution, and sta tistical relationship. The assumptions that underlie many types of sta tistical models have been discussed. The role of residual analysis has been stressed as a useful method to verify assumptions. We have descr ibed transformations and alternative regression methods that are emplo yed when these assumptions are found to be in violation. Sequential se lection procedures for the construction of multivariate models have be en presented. The importance of assessing model performance has been u nderscored, most notably in terms of bias and precision. In summary, p harmacodynamic analyses are now commonly performed and reported in the oncologic literature. The content and format of these analyses has be en variable. The goals of such analyses are to identify and describe p harmacodynamic relationships and, in many cases, to propose a statisti cal model. However, the appropriateness and performance of the propose d model are often difficult to judge. Table 1 displays suggestions (in a checklist format) for structuring the presentation of pharmacodynam ic analyses, which reflect the topics reviewed in this paper.