This is an interesting proposal and using empirical, data driven approaches to study heterogeneity in the response to therapy is a very important area of investigation. However, the clinical trials are limited to sitagliptin and canagliflozin so you will only have these effects. In addition, how much difference was there in HbA1c? Was there substantial enough variation to do an analysis like the one the author’s propose.
Recommendation for this data request: Not Approve