Depends on the results. A blinded study which has a strong enough effect can definitely be solid indicator. > Without diet treatment none of 12 participants in

Depends on the results. A blinded study which has a strong enough effect can definitely be solid indicator. > Without diet treatment none of 12 participants in the control group had remission, but after they had also received the diet program, 10 out of a total of 23 participants (43%) were free from diabetes, without need for medication. That's very unlikely to be due to chance - a difference of 0/12 vs 10/23. Of course, this doesn't protect us from failure modes where the study was set up badly or someone just plain cheated, which is why we do want more than one study, but as far as predictive power goes, this is solid. And fortunately this is already confirmatory research, not exploratory - we've either known or strongly suspected this for quite a while. Now we're just in the stage where it goes mainstream. And speaking of what we know/strongly suspect, adding muscle building into the mix also helps. Living with 800 calories per day is not sustainable, but losing weight once, then building (a lot of) muscle then maintaining a reasonable body fat is hard, but doable. Lately I wonder if 3 months on semaglutide, low dose steroids, protein and gym is actually an intervention well worth its risks, at least for some categories. Replacing 10kg of fat with 10kg of muscle is much less likely to yoyo than just losing weight.