|
Abstract:
|
A variety of approaches have been proposed to provide formal and informal validation of proposed surrogate markers. To achieve true clinical impact, the validation must convince statistical, regulatory, and clinical communities. In this talk I argue, through an actual data example, that the best approach is not a single method but a multi-faceted exploration. Effective validation will use at least two types of approaches (1) those that directly appeal to clinicians but with less statistical foundation (focusing on graphics and simple data summaries), and (2) those arising from statistical modeling (such as multivariate models), but which are more difficult to interpret clinically. Both methods are maximally powerful when using the meta-analytical approach.
|
