Consider whether affiliated people on Twitter are the sort of people you'd feel comfortable being able to anonymously browse your medical history without any log of the action because it was classified as research on a data set. Researchers are people like anyone else, except they're just as awful, with more access, fewer adult incentives and boundaries, and almost no personal accountability for privacy. If you have doubts about whether you would trust U.S. college IT people to run facebook, pornhub, reddit, or tinder on your behalf, I recommend paying attention to what this new national data guardian has to say.
The risk I think she will need to manage is the pressure universities apply to health agencies for data. There is a perfect storm brewing where privacy for outsiders isn't a big cultural norm in some reserach communities, but using data for social political purposes is. Reserach ethics boards are nice, but they're like the police investigating themseves.
The question isn't so much whether we could cure cancer with just this bit more data without oversight, and more whether the people demanding it are above using a perceived crisis to squeeze the toothpaste out of the tube. She has a very interesting job ahead of her.