Abstract:
|
Project Data Sphere is a public database of oncology studies with patient level data. Here, we present two examples of how such data can be used to augment internal data and produce more precise estimates of treatment effects. The first example is a Phase II oncology study with no control group. In this case, the patient level data allows us to determine the relationship between covariates and adjust for differences in the study populations. In the second example, we show how a prognostic model generated from the external data can be used to stratify patients in an internal Phase III study. We also discuss how this work differs from traditional meta-analysis.
|