The Breakthrough Reality Check: Seven Questions to Ask Before Sharing a Medical Headline

When a health headline turns out to be overblown, the blame usually lands on journalists. The evidence tells a more uncomfortable story. In a 2014 study in the BMJ, researchers compared hundreds of university press releases with the research papers and news stories linked to them. A substantial share of the press releases already contained exaggeration: overstated health advice, cause-and-effect claims drawn from studies that could only show correlation, or animal findings presented as if they applied to people. When the press release exaggerated, the news usually did too. When it did not, news exaggeration was much less common.

So overstatement can enter at every link in the chain, from lab to press office to newsroom to the group chat. The last link is you. Before you share the next “breakthrough”, seven questions can separate the genuinely important from the merely exciting. They take about two minutes, and you do not need a science degree.

Why a checklist beats gut instinct

Medical headlines are designed to feel urgent. Hope, fear and novelty are powerful, and they make it easy to skip the details that decide whether a result matters. A short routine slows you down just enough. It is also fairer to the science: most early findings are not wrong, they are simply early, and the questions below help place a result at the right point on its journey.

A headline tells you something happened. The seven questions tell you what, to whom, and how sure anyone can be.

The seven questions

1. Was it tested in people, or in cells or animals?

This single question resolves a surprising number of headlines. Results in cells in a dish or in mice are preclinical: essential first steps, but many treatments that look promising at this stage never show the same benefit in humans. Look for words like “in mice”, “in vitro” (meaning in a dish) or “cell lines”, often buried several paragraphs down. Our guide on how to read a biotech breakthrough goes deeper.

2. Was there a fair comparison?

If people who got a treatment did well, the question is: compared with whom? The strongest evidence usually comes from a randomized controlled trial, in which participants are assigned by chance to the new treatment or to a comparison such as standard care or a placebo (an inactive dummy treatment). Without a comparison group, improvement could reflect natural recovery, other care, or the hopes of those involved. Observational studies, which simply follow people over time, can reveal associations but struggle to prove cause and effect.

3. How many people, and for how long?

Size and duration matter. The FDA describes Phase 1 trials as typically involving 20 to 100 people and focusing on safety and dosage, Phase 2 as up to several hundred people, and Phase 3 as roughly 300 to 3,000. The agency also estimates that only around a third of drugs move from Phase 2 to Phase 3. An exciting Phase 1 result is real information, but it is an early chapter. Our explainer inside a clinical trial walks through each phase.

4. What was actually measured?

Did the treatment help people live longer or feel better, or did it change a number on a lab test? The FDA calls the latter a surrogate endpoint: a measurement used in place of a direct measure of how a patient feels, functions or survives. Some surrogates, like blood pressure for stroke risk, are well validated. Others are only “reasonably likely” to predict benefit and need confirmation. Shrinking a tumor on a scan, for example, is not automatically the same as living longer.

5. How big is the benefit, in absolute terms?

“Halves the risk” sounds dramatic. But a relative risk change tells you nothing without the starting point. As Cancer Research UK explains, a tenfold increase from 2 in 100,000 to 20 in 100,000 is still a small absolute risk. Look for “how many people out of 100 (or 1,000)” and whether the difference is large enough to matter in daily life.

6. Where did the claim come from?

Not all sources are equal. A peer-reviewed journal article has been checked by independent experts, though peer review is far from infallible. A preprint has not: medRxiv, a major preprint server for health research, warns that its preprints have not been certified by peer review and should not guide clinical practice or be reported as established information. Conference abstracts often contain limited data, and company press releases describe results in the company’s own words. None of these is worthless; each deserves a different level of confidence.

7. Who benefits, and what do independent experts and regulators say?

Check who funded the work and whether authors or spokespeople have financial interests. That does not make a result false, but it is context. Then look for independent comment, replication by other groups, and regulatory status. A drug in a trial is not an approved treatment; the FDA’s drug development process ends with review and post-market monitoring, not a headline.

A worked example

Here is an invented headline, for practice only: “New compound reverses aging and wipes out tumors.” Run it through the questions. The underlying paper, you find, studied mice (question 1) with a small comparison group (2) over a few weeks (3), measured tumor size and some blood markers (4), reported relative changes (5), was posted as a preprint (6), and was promoted by a startup whose founders are co-authors (7). None of that makes the research bad. It makes the headline wrong. A fair version might read: “Compound slows tumor growth in mice in early, unreviewed study.”

Note that no single study can show a “cure for cancer”. Cancer is many different diseases, with different treatments and outcomes, as we explain in Is There a Cure for Cancer?, and claims about reversing aging deserve special skepticism, as our field guide to slowing aging sets out.

Sharing responsibly

You do not have to stop sharing science news. You can share it better:

• Add the context the headline left out: “in mice”, “early trial”, “not yet peer-reviewed”.
• Link to the original study or an independent explainer, not only the most breathless story.
• Be especially careful with posts that might lead someone to change or stop a treatment. Those decisions belong with a clinician.
• If you later learn the story was overstated, say so. A correction helps the people who saw your first post.

The pocket version

Before you share, ask:

1. Was it tested in people, or in cells or animals?
2. Was there a fair comparison?
3. How many people, and for how long?
4. What was actually measured?
5. How big is the benefit, in absolute terms?
6. Where did the claim come from?
7. Who benefits, and what do independent experts and regulators say?

Ready to practice on real-style headlines? Try our interactive Breakthrough Reality Check.