n = 118,000 · Observational · 5 min read
Ultra-processed foods and cardiovascular risk: a new cohort
A large observational study, a real signal, and three reasons it can't prove causation on its own.
Read more
Health news
New research, read properly
We summarise the studies making headlines — with sample size, design, funding and limitations stated before the conclusion, not after it.
Health News
n = 25,871 · RCT · 6 min read
Vitamin D supplementation misses another endpoint
A well-powered randomised trial finds no fracture benefit in replete adults. Who might still benefit.
Read morePolicy analysis · 8 min read
Sugar taxes, four years on
Reformulation has done more than consumer behaviour change. The natural experiments now agree.
Read more4 min read
The coffee study everyone misread
A hazard ratio of 0.96 is not a health revolution, whatever the headline said.
Read more9 min read · Reviewed by a clinician
GLP-1 drugs and diet quality
Appetite falls, but so can protein and fibre intake. What clinicians are now recommending alongside.
Read more6 min read · Corrections log
Three nutrition claims retired this quarter
Including one we reported in 2024 — and what the newer evidence says instead.
Read more
Start here
How to read a health headline
Most misleading health news isn't invented. It's a small study described as if it were a large one.
- Check whether it was humans, mice or cells.
- Look for the absolute risk, not just the relative one.
- Observational studies show links, not causes.
- Ask who funded it — and who ran the press release.
- One study rarely overturns a body of evidence.
“You're the only outlet I've seen lead with the sample size. It changed how I read everything else.”
— Jonas M., reader
How we review. Articles in Health News are checked against primary research and re-read whenever significant new evidence appears. See our editorial policy.