Thirty-nine randomised trials, 3,236 adults. Earlier eating won — on the numbers. But the certainty of evidence runs very low to moderate, and none of those trials tested a household dinner rule. They tested eating windows in adults who already had cardiometabolic risk. A real signal measured in the wrong people is still a real signal. It just isn’t your kitchen yet. Not medical advice. Source: Hamsho et al., 2026 — network meta-analysis, 39 RCTs.
What it says
Your grandmother’s seven o’clock dinner rule just got tested in thirty-nine trials.
Across three thousand adults, eating earlier lowered fasting blood glucose by three point six milligrams per decilitre.
Then I read the certainty grading: very low to moderate, which is the researchers’ own words.
And nobody tested the household rule — they tested eating windows in adults already at cardiometabolic risk.
So the honest version: a real signal, measured in the wrong people, and not medical advice.
That was the Evidence Desk, where the practice and the trials testing it get graded separately.
Sources
Hamsho et al., 2026 · 39 RCTs
Watch more on YouTube: @chittalab.
Narration and illustrations are AI-generated. Every claim is sourced above.
