This is pretty astonishing research, not so much for the results, as for the methodology. Doing research on what people eat is VERY hard, it's very hard to get accurate data. Pairing people and having them eat identically is brilliant. I hope it's also scalable for longer term studies. By pairing people they are likely to give much more accurate reports about what they actually eat, and it allows direct comparisons.
It is very interesting. I wonder, however, if it's enough to isolate all relevant variables (this is peanut gallery speculation, I am not a scientist of any stripe)? The narrator explains that she is "an accident & emergency doctor [...]. Rushing around on my feet all day with strange working patterns". Never mind that the stress and adrenaline of such work must affect her body in some ways, she probably also moves quite a bit more than the average person. We're not told anything about her counterpart than her name, and that she's the same age and gender.