Biostatistics / Study Design
We will work with NIH staff and the demonstration project
teams to create initial guidance and technical documents regarding study design
and biostatistical issues.
Initial guidance and best practice documents will also address additional gaps
and limitations that will be unveiled as the UH2 grants are planned. Understanding
which methods can be directly applied in a distributed data setting, which can
potentially be modified for the distributed data setting, and which will
require individual level data will be among the critical issues this core addresses.
As the UH2 and UH3 projects are funded we will identify new issues and add
additional content and case studies. We will ally with NIH collaborators, UH2
and UH3 investigators, and academic institutions to gather additional input
into key methodological issues. Areas in need of methods development will
become apparent, and we will work to address these methodology challenges with our
collaborators. One method that will be developed as part of the Collaboratory
work is an approach to handling informative missing follow-up data when using
the EHR as the basis for follow-up data collection. Specifically, those that
are less healthy with more chronic conditions have more visits per year. If an
intervention is effective in improving general health than those that received
the intervention are more likely to have missing follow-up data compared to
those that did not receive the intervention. Ignoring this missing data issue
can lead to biased results (e.g. for this example one may conclude that the
effective intervention is not effective).