The overall goal of this research is to develop interactive software for performing sophisticated heterogeneity analysis of clinical trials involving multiple centers and/or multiple strata. In particular, the software will enable the user to explore how different strategies for combining treatment effects across strata (e.g. center, sex, race, disease severity, etc.) affect the resulting estimate of the overall treatment effect. By providing tools for systematically examining the magnitude and precision of multiple estimates of the overall treatment effect (based on the different combining strategies), the software will support the researcher in assessing the impact of any observed heterogeneity with regard to overall inference. This can also be viewed as a data-mining tool, enabling the researcher to interactively search for different possible structures in the data set. The ultimate benefit of successful development and deployment of this technology is improvement in health care by enabling researchers to better understand the nature of heterogeneity that may be encountered in clinical trials, thereby increasing the understanding of the overall safety and efficacy of drug products. PROPOSED COMMERCAIL APPLICATION: NOT AVAILABLE
