How to read a study, or the headline about one
Menopause research gets reported badly in a small number of predictable ways, and six questions catch nearly all of it.
1. Who was studied? Average age, health status, whether they were already postmenopausal. A finding in one group may not transfer to another, and this field has a long history of results being applied to populations they did not study.
2. What was actually measured? A symptom score, a biomarker, a self-report, or a clinical outcome. These are not the same thing and headlines routinely swap one for another.
3. How big is the effect in absolute terms? This is the question that catches the most. A rise from 1 in 1,000 to 2 in 1,000 is a "100 percent increase" and one extra case per thousand people. Relative risk without absolute numbers is the standard way a small finding is made to sound enormous, in both directions.
4. Was it a comparison or an experiment? Observational research shows associations, which may have other explanations. A randomised trial supports causal claims. Both are useful; they are not interchangeable.
5. Who funded it, and what do the authors declare? Not disqualifying. Worth knowing.
6. Does the headline match the paper? Frequently it does not, and the abstract is public and takes three minutes.
And the rule underneath all six: a single study is one input. It does not overturn guidance, and it is not a reason to change treatment. That decision belongs with your prescriber.
The specific failures this field produces
Applying a result to the wrong population. The most consequential failure in menopause research reporting, and the reason the age and health status of participants is the first question rather than a detail.
Relative risk with no denominator. "Doubles the risk" is meaningless without knowing what it doubled from. Ask for the two absolute numbers, and if the article does not have them, the paper does.
Turning an association into a cause. "Women who did X had less Y" is compatible with X causing less Y, with less Y causing X, and with the kind of person who does X differing in a dozen other ways.
Reporting a surrogate as an outcome. A change in a measurement is not automatically a change in how anyone feels or what happens to them.
Small studies reported as findings. A promising result in forty people is a reason to do a bigger study.
A single study framed as a reversal. Bodies like NICE and professional societies update guidance by weighing the whole evidence base, and a new paper enters that process rather than replacing it. What NICE NG23 actually says.
Checking one yourself, in ten minutes
- Find the paper. The article usually names the journal; if it names nobody, that is itself information.
- Read the abstract. Population, method, main result, and the numbers.
- Look for absolute figures. If only percentages are given, look in the results section for the counts.
- Read the limitations paragraph. Authors are usually candid there, and it is the paragraph coverage skips.
- Compare with the headline. Note the gap.
Doing this three or four times permanently changes how you read health news, which is a better outcome than any individual answer.
Where this matters for tracking
The same discipline applies to what you conclude from your own record, on a smaller scale.
Your own data is observational, and n equals one. A symptom improving after you started something is compatible with the thing helping, with regression from a bad stretch, with the season changing, and with several other explanations. That is not a reason to distrust your record. It is a reason to record what else was happening, and to let a clinician interpret it.
It is also why HormoneLog does not tell you what your data means. No score, no threshold, no "your symptoms improved by 40 percent." A number like that would carry all the errors above in a smaller package. What menopause symptom scales are, and what they are not. · When HRT does not seem to be working.
Where to start instead of a headline
Your own health service's pages, which are written to be read by patients and reviewed periodically.
Guidance documents like NICE NG23, which are public, written for clinicians, and considerably more specific than any summary of them.
Professional societies, for practitioner directories and position statements.
The paper itself, when a headline matters to you. It is nearly always findable and the abstract is nearly always readable.
What to take to your first appointment. · What a menopause specialist actually is. · Why tracking perimenopause symptoms is worth it.
More in this section
About HormoneLog
HormoneLog is an iOS app from Baker Ventures LLC for tracking perimenopause and menopause symptoms and keeping an HRT record: what you took, at what dose, from when, and what changed afterwards. It is built for the appointment, where "it has been bad lately" is worth far less than a dated record of frequency and severity. The app is in development.
HormoneLog is a record-keeping tool. It does not diagnose, does not score you, and makes no claim about symptom relief. The guides on this site link to NICE, the NHS, the Menopause Society and peer-reviewed sources, and say plainly where the evidence is thin. Everything here is free and needs no account.
Questions and answers
How do you tell whether a health headline is reliable?
Ask who was studied, what was measured, how big the effect was in absolute terms, whether it was a comparison or an experiment, who funded it, and whether the headline matches what the paper says. Most misleading coverage fails at the third or the sixth.
What is the difference between relative and absolute risk?
Relative risk describes the change compared with a baseline, so a jump from 1 in 1,000 to 2 in 1,000 is a 100 percent increase and one additional case per thousand people. Reporting the first without the second is the most common way a small finding is made to sound enormous.
Why does the population studied matter so much?
Because a finding in one group may not transfer to another. Average age, whether participants were already postmenopausal, and existing health conditions all change what a result means, and menopause research in particular has been misapplied across groups it did not study.
Does observational research prove cause?
No. Observational studies show associations, which can have other explanations including that people who do one thing differ systematically in other ways. That does not make them useless; it makes them a different kind of evidence from a trial.
Should I change treatment because of a news story?
No. A single study, however well reported, is one input into a decision that depends on your history and circumstances, and that decision belongs with your prescriber rather than with an article.
How can I check a headline myself?
Find the paper, read the abstract, and look for the population, the comparison, the absolute numbers and the limitations paragraph. That takes ten minutes and it resolves most of the question.
HormoneLog. “How to read a study, or the headline about one.” Baker Ventures LLC, September 7, 2026. https://hormonelog.bakerventuresstudio.com/guides/how-to-read-a-menopause-study/