How we check health claims
sciencebased.news covers the health and longevity claims people hear on podcasts, social media and in the news. For each one we find the studies behind it and give it a grade.
What we check
We look for research that could make a real difference to how long and how well you live: more strength, more energy, better sleep, fewer years of illness. We track the big health podcasts, new papers in the top medical journals and university press releases, and pick what people are talking about. Then we check how strong the evidence is, so you can tell a real breakthrough from an early hint. Some claims don't hold up, and we say so, but our main job is to find the ones that do. We state the claim the way you'd hear it ("40% of the weight you lose on Ozempic is muscle"), not a softer version that's easier to defend or attack.
How we read a study
Every study we cite comes with four facts, because they decide how much it can tell you:
- Human or animal. A result in mice is a reason to run a human study, not a reason to change what you do.
- How many people. A study of 16 people can find a big effect by chance. A study of 10,000 rarely does.
- Randomized or observational. A randomized trial can show cause. An observational study can only show that two things go together.
- How long. A 12-week trial can't tell you about the next 10 years.
We also say who paid for a study when it's the company that sells the product. We open every study we cite and check each number against it.
Our grades
| Grade | Means |
|---|---|
| Solid | Several good human studies, including randomized trials, agree |
| Promising | Some human trial evidence, but small, short or not yet repeated |
| Weak | Only observational links, animal studies, or studies that conflict |
| Unproven | No direct human evidence either way |
| Not supported | Good human evidence says it doesn't work or isn't true |
Each article opens with the claim and the grade, then shows the studies, gives the case for the other side, and says what new evidence would change the grade.
Updates
When a new study changes the picture, we update the article, change the grade if needed, and note the date. We don't write a second article on the same claim.
What we save when you search or send feedback
When you use the search box we save the words you typed and the basic details your browser sends with every request: the browser name, language, country, and whether the page ran JavaScript. We use them to tell people from AI assistants and bots. We don't save your IP address. We read the searches to decide what to check next.
The same goes for the "Was this helpful?" buttons and the "What health claim should we check next?" box under each article: we save your answer or suggestion, which article it came from, and the same browser details. Never your IP address, and we don't ask for your name or email. Please don't put personal details in a suggestion.
Not medical advice
These articles explain what the research shows. They aren't advice for you personally. Talk to your doctor before starting, stopping or changing a treatment.