Review: Clinical Trial Methodology: Surrogate endpoints: when should they be used?

Abstract

Surrogate end point evaluation has received a lot of attention for approximately a quarter of a century. Throughout this time, the topic has been surrounded with both hopes and perils. The history of the evaluation of surrogate end point evaluation is described, from Prentice’s definition, via single-trial evaluation to meta-analytic approaches. While based on technical developments and advances in statistical methodology, the treatment of the topic here is largely nontechnical. A perspective is given as to what might be a sensible way forward. It is clear that statistical evaluation can offer a contribution to the discussion as to whether a surrogate will be adopted and in what form, but can never be seen as the sole decision maker. While there is room for the adoption of surrogates, it is very important that pitfalls and drawbacks be kept in mind at all times.