Evaluation of surrogate endpoints in randomized experiments with mixed discrete and continuous outcomes

A statistical definition of surrogate endpoints as well as validation crite ria was first presented by Prentice. Freedman et al. supplemented these cri teria with the so-called proportion explained. Buyse and Molenberghs pointe d to inadequacies of these criteria and suggested a new definition of surro gacy based on (i) the relative effect linking the overall effect of treatme nt on both endpoints and (ii) an individual-level measure of agreement betw een both endpoints. Using data from a randomized trial, they showed how a p otential surrogate endpoint can be studied using a joint model for the surr ogate and the true endpoint. Whereas Buyse and Molenberghs restricted thems elves to the fairly simple cases of jointly normal and jointly binary outco mes, we treat the situation where the surrogate is binary and the true endp oint is continuous, or vice versa. In addition, we consider the case of ord inal endpoints. Further, Buyse et al. extended the approach of Buyse and Mo lenberghs to a meta-analytic context. We will adopt a similar approach for responses of a mixed data type. Copyright (C) 2001 John Wiley & Sons, Ltd.