What counts as strong evidence in women's health

Reviewed by Emma Carter, Senior Health & Nutrition Editor

Health advice seems to flip every week: a food is good for you, then it is not; a supplement is a breakthrough, then a waste of money. Most of that whiplash comes from treating every study as equal. They are not. Once you can tell strong evidence from weak, the noise gets much easier to filter, and you do not need a science degree to do it.

Not all evidence is equal

Researchers rank evidence in a rough pyramid, from the weakest at the bottom to the strongest at the top. Higher up means the finding is less likely to be down to chance, bias or wishful thinking.

Researchers formalize this with a system called GRADE, which rates certainty as high, moderate, low or very low. Randomized trials start high and observational studies start low, then the rating is downgraded for problems in five areas: study design limitations, inconsistency between studies, imprecision, indirectness, and publication bias (Am J Epidemiol, 2024). That is why a study's design is only the starting point, not the verdict.

A single glowing testimonial sits at the bottom for a reason. It might be genuine, but it cannot tell you whether the result was the treatment, the placebo effect, or luck.

What a meta-analysis actually is

Near the top of the pyramid sit systematic reviews and meta-analyses, and the two often get muddled.

A systematic review gathers every good-quality study on a question using a clear, repeatable method, then weighs them up together. A meta-analysis is the extra step some reviews take: pooling the numbers from those studies into one combined result. Because a meta-analysis draws on many more people than any single trial, it can detect real effects that smaller studies miss, and it can show when a popular claim does not hold up. Reviews from organisations like Cochrane are widely treated as a benchmark, because they follow strict methods to reduce bias.

Why one study rarely settles it

A single study is where a question opens, not where it closes. Several things explain why:

• Small studies mislead. A trial of 20 people can throw up a striking result by chance that disappears in a larger one.
• Findings need to replicate. A result you can trust is one that shows up again when other teams test it independently.
• Correlation is not cause. Observational studies can show that two things travel together without one causing the other. Women who take a supplement may simply be healthier to begin with.
• "Statistically significant" is not the same as important. A result can be statistically real yet so small it makes no practical difference to your life.
• Who paid matters. Studies funded by a company selling the product are worth reading with extra care.

This is exactly why marketing outpaces science so often. A brand can build a campaign on one small, favorable study while the fuller picture, seen across many studies, is far more modest. You can see the gap in guides like collagen, where the marketing runs ahead of the evidence, compared with something like vitamin D, where the case is better established.

Questions to ask about any health claim

You can pressure-test almost any headline with a few questions:

1. Is this one study, or a review of many?
2. Was it done in people, or only in cells or animals?
3. How many people took part, and for how long?
4. Did it compare against a control group?
5. Who funded it, and what were they selling?
6. Has anyone repeated the finding?

How we use this at VeriNourish

This hierarchy shapes everything we publish. We lean on systematic reviews, meta-analyses and guidance from bodies like the NHS and the NIH before we lean on any single study, we name our sources so you can follow the evidence yourself, and we say plainly when the evidence is weak or mixed rather than dressing a hopeful guess as a fact. You can read more in our editorial policy. The aim is not to sound certain; it is to be honest about how strong the evidence actually is.

Frequently asked questions

What is a meta-analysis?

A meta-analysis pools the results of many separate studies on the same question and combines them statistically into one overall answer. Because it draws on far more people than any single study, it usually gives a clearer, more reliable estimate of whether something works and by how much.

What is the difference between a systematic review and a meta-analysis?

A systematic review gathers and appraises all the good-quality studies on a question using a defined method. A meta-analysis is the statistical step some systematic reviews take to combine those studies into a single number. Every meta-analysis sits inside a review, but not every review runs one.

Why can't I trust a single study?

One study can be small, chance-driven or flawed, and results often shrink or vanish when others try to repeat them. A finding becomes trustworthy when it holds up across several studies and, ideally, in a systematic review. Single studies are where questions start, not where they end.

What is the strongest kind of health evidence?

A well-conducted systematic review or meta-analysis of randomised controlled trials sits at the top of the evidence hierarchy. Below it come individual randomised trials, then observational studies, then lab or animal work, then expert opinion and anecdote at the bottom.

References

1. Assessing the certainty of the evidence in systematic reviews: importance, process, and use · Brignardello-Petersen R, Guyatt G, American Journal of Epidemiology, 2024
2. Levels of evidence: an introduction · Centre for Evidence-Based Medicine, University of Oxford
3. The evidence-based medicine pyramid · Students 4 Best Evidence, Cochrane