MissingDataMatters

 


Missing outcome data are a widespread problem in randomized trials. The analysis of studies with missing data require untestable assumptions. As a result, it is important that one evaluate the robustness of the trial results to such assumptions (i.e., sensitivity analysis). We have been funded by FDA, PCORI and NIH to develop methods and software tools that will allow users to conduct global sensitivity analysis of their studies. Global senstivity analysis is like "stress testing" in reliability engineering, where a product is systematically subjected to increasingly exaggerated forces (i.e., assumptions) in order to determine its breaking point (i.e., non-significant result).  If the breaking point occurs at forces that are judged to be extreme, then the results are judged to be robust; otherwise, the results are judged to be fragile. We will disseminate our sensitivity analysis methods and software through this website.  If you would like to become a user of the software and participate in the online discussion forum, please request an account.