To definitively test the carbohydrate-insulin model, studies should use extreme isocaloric diets with very high and very low insulin responses.

“I think as an efficacy study that was testing a theory just a theory with again you start with a very theoretical response uh and then you build from there to say okay is this Theory applicable in retrospect I think it should have been more extreme in the carbohydrate and fat differences I think one should have been if the carbohydrate insulin model was being tested one should have been a very very very high insulin diet and a very very low insulin diet that were isocaloric”