Comparison of mediational selected strategies and sequential designs for preventive trials: Comments on a proposal by pillow et al.

Mediational selection designs can increase the chance of finding an intervention to be statistically better than a control when compared to interventions applied to a selected subpopulation such as all children of divorce. Selection based on poor mediational variable scores creates a high expectation that many subjects will benefit because it matches a population to a given intervention strategy. It does not permit examination of the impact on other subjects nor the possibility of interventions that may be tuned to the specific needs of different subpopulations. A sequence of trials each built on the results of previous ones has the potential for extending the results obtained from a mediational selection design.