Folks prescribed Metformin in medicare data would more than likely be taking it as treatment for diabetes. That's a fundamentally different population than hea

Folks prescribed Metformin in medicare data would more than likely be taking it as treatment for diabetes. That's a fundamentally different population than healthy people taking it for longevity. It also limits you to outcomes measured in claims data - one of the primary outcomes in this study is related to gene expression, which would not be found in claims data. There's definitely preliminary work that can be done by looking at claims data, but that work just sets the stage for actual randomized clinical trials.