How to Read Health News: Telling Good Evidence From Bad

Every week brings a headline promising that something ordinary either causes cancer or prevents it. Most of those stories are built on real research, which is what makes them so effective at misleading people. The problem is rarely fabrication. It is the gap between what a study measured and what a headline claims it proved. Here is a practical toolkit for closing that gap yourself, in about the time it takes to read the article.

Start with what kind of study it was

Not all evidence carries the same weight, and the type of study tells you most of what you need to know before reading a word of the findings. At the bottom sit test tube and animal studies, which are useful for understanding mechanisms and terrible for predicting what happens in people. A great many exciting health headlines describe something that happened to cells in a dish or to mice given doses no human would ever consume.

Above those sit observational studies, which follow groups of people and look for associations. They are valuable, especially for questions you cannot ethically randomise, but they cannot establish cause. Then come randomised controlled trials, where people are randomly assigned to an intervention or a comparison, which is the design that can actually demonstrate cause and effect. At the top sit systematic reviews and meta-analyses that pool many trials. When a headline describes a single small study contradicting a body of existing evidence, the existing body is usually still right.

The association trap

The most common error in health journalism, and in health marketing, is treating an association as a cause. Observational studies routinely find that people who do X have less of Y, and the honest interpretation is almost always complicated by confounding: the people who do X differ in a dozen other ways that also affect Y.

The classic example is the long-held belief that moderate drinkers were healthier than non-drinkers, which largely dissolved once researchers noticed the non-drinking group contained people who had stopped because they were ill. The useful habit is to ask, whenever you see an association, what kind of person does this thing, and what else is true about them. Watch the verbs too: careful reporting says linked to or associated with, while the headline above it often says causes. Those are different claims, and the headline writer rarely reads the methods section.

Absolute versus relative risk

This one distinction defuses a large share of alarming headlines. A study reporting that something raises your risk by 50 percent is reporting a relative change, and relative changes say nothing about how likely the outcome was to begin with. If the baseline risk was 2 in 10,000, a 50 percent increase takes it to 3 in 10,000. That is a real effect and a trivial one in daily life.

So whenever you see a percentage, look for the underlying numbers. Good reporting gives both: the relative change and the absolute change. Anything quoting only the relative figure is either being lazy or is trying to make a small effect sound frightening. The same logic works in reverse for supplements and interventions, where a large-sounding percentage improvement often rests on a tiny absolute difference in a surrogate measure that nobody would notice in their own body.

The details that separate solid studies from weak ones

Once you know the study type and the risk framing, a short checklist covers most of what remains. None of it requires statistical training.

• How many people: twenty participants tells you very little; thousands over years tells you rather more.
• How long: a six-week study cannot say anything about long-term outcomes.
• Who was studied: results from young male athletes may not transfer to a sixty-year-old with diabetes.
• What was actually measured: a change in a blood marker is not the same as fewer heart attacks.
• Who funded it: industry funding does not invalidate research, but it belongs in your assessment, particularly for supplements.
• Was it peer reviewed: preprints and conference abstracts have not yet been through that filter.

The surrogate marker point is worth dwelling on. Many trials measure something convenient, such as a change in cholesterol, inflammation markers or bone density, because measuring actual events like heart attacks or fractures takes years and thousands of participants. A change in the marker is a reasonable hint. It is not proof that the outcome people care about improved.

Practical habits that take under a minute

Read past the headline to find the study itself, and note whether the article links to it at all. Reputable outlets link the source, and its absence is informative. Check whether the study was in humans. Look for the sample size and duration, which are usually stated in a single sentence. Then ask whether the claim being made matches what was measured.

Beyond individual stories, the better move is to check whether the claim fits the broader evidence, and a handful of sources do that work well. The Cochrane Library publishes systematic reviews across medicine, and the NHS and major academic centres maintain plain-language summaries that reflect the weight of evidence rather than the newest single paper. For supplements specifically, independent evidence databases that grade the strength of research are far more useful than a manufacturer's citation list. If a claim appears in one exciting article and nowhere in those sources, that mismatch is your answer.

A note on your own health decisions

The point of all this is not to become cynical about research. It is to stop making changes to your health on the strength of a single study reported by someone working to a deadline. Real shifts in medical consensus happen slowly, across many studies, and they rarely arrive as a surprise announcement about one food.

Before acting on any health claim, particularly one involving stopping a medication, starting a supplement or making a major dietary change, discuss it with a doctor or pharmacist who knows your situation. This is general information rather than medical advice. The skill worth building is not knowing every answer, it is recognising when a claim has not yet earned your behaviour change.

Frequently asked questions

How can I tell if a health study is reliable?

Check the study type, how many people took part, how long it ran, whether it was done in humans, what outcome was actually measured, and whether it has been peer reviewed. Single small studies contradicting established evidence are usually wrong.

What is the difference between relative and absolute risk?

Relative risk describes the size of a change compared with a baseline, while absolute risk describes the actual chance of the outcome. A 50 percent increase on a baseline of 2 in 10,000 is still only 3 in 10,000.

Why do health headlines keep contradicting each other?

Because each new study is reported as a verdict rather than as one piece of a larger body of evidence. Individual studies vary, and consensus only emerges from many of them together.

Are animal studies useful?

They are useful for understanding mechanisms and for deciding what to test next in people, but they frequently fail to translate. Doses are often far higher than any human intake, and mice are not small humans.

Read next

• Do You Actually Need Supplements? An Honest Guide

  Applying the same evidence filter to the supplement aisle.

• Longevity Supplements: A Reality Check

  A case study in claims outrunning the research.

• Blood Test Results Explained: A Plain-English Guide to the Common Numbers

  Reading your own numbers with the same scepticism.

This article is for general informational purposes only and is not medical advice. Always consult a qualified healthcare professional about your individual situation, and never start or stop a medication without your doctor. See our Medical Disclaimer.