As I understand their method, they used students, tested once using an oral glucose tolerance test (OGTT) to rule out diabetes, each test used 3 different students and the values of all 3 were averaged for that food.
The obvious problem is glucose control is inadequate to infer insulin production. Dr Joeseph Kraft found that 78% of people who pass an OGTT and have normal glucose are making a LOT of insulin in order to achieve that. The score each discrete food receives depends on which Kraft pattern it’s 3 students had.
I think the insulin index is mostly bullshit. But I also believe this is something we do need to measure, so it needs to be redone for each food tested with multiple subjects that exhibit each of the first 4 Kraft patterns. Now THAT would be interesting and actionable information.