Considerations in the Evaluation of Surrogate Endpoints in Clinical Trials

摘要: We report on recommendations from a National Institutes of Health Workshop methods for evaluating the use surrogate endpoints in clinical trials, which was attended by experts biostatistics and trials broad array disease areas. Recent advances biosciences technology have increased ability to understand, measure, model biological mechanisms; appropriate application these research settings requires collaboration quantitative laboratory scientists. Biomarkers, new examples arise rapidly technologies, are used frequently such areas as early detection identification patients most likely benefit therapies. There is also scientific interest exploring whether, under what conditions, biomarkers may substitute phase III although workshop participants agreed that considerations apply primarily situations where using not feasible. Evaluating candidate exploratory phases drug development investigating confirmatory require establishment statistical inferential framework. As first step, reviewed degree can explain or predict effect treatments measured trials. They suggested approaches reflect only indirectly important processes causal path biomarker measurement errors concern. Participants emphasized need further models, whether they empirical nature attempt describe mechanisms mathematical terms. Of special were meta-analytic models combining information multiple studies involving interventions same condition. Recommendations included design conduct assemblage databases needed research. Finally, there strong recommendation training scientists biologic well modeling ensure will be an adequate workforce meet future needs.