Maybe it started with a headline, a friend’s post, or a product page that said “studies show” something. Now you are holding a health claim and asking whether it has anything to do with you. That is the right question, and most health claims never answer it on their own.
Here is the short answer: a study result applies to you only as far as you resemble the people in it, and the size of any benefit depends on the numbers behind the headline. This guide walks you through checking both, then turns what you find into questions for a clinician.
| What this guide does | Sorts a health claim into who it describes, what kind of source it is, and which numbers to check |
| Where to start | A claim you saw, a recommendation, or a study you cannot yet read |
| Who decides | You and a clinician, especially for any treatment, medicine or supplement you are considering |
| Main sources | Published reviews indexed in PubMed by the U.S. National Library of Medicine |
| What it does not do | Tell you whether a product works for you, or diagnose any condition |
Where are you starting?
Step 1: Find out who the claim was written for
You are at the start: you have a claim in front of you and no clear sense of whether it fits your life.
Begin with the people. Start by checking who the study enrolled, including their age, health status and how many took part. The guide on why who was studied matters explains each of those checks.
Recommendations work the same way. A recommendation is written for a group with certain traits, not for every reader, and the guide on reading a health recommendation shows how to tell the difference.
What this means for you: treat the claim as a statement about a group until you can say who that group is and how you differ from it.
Read next
Questions to Ask When Health Information May Not Fit You: a step-by-step way to turn a finding into one question.
Once you know who a claim describes, the next question is what kind of document you are holding.
Step 2: Work out what kind of source you are holding
You may be looking at a study, a guideline or a fact sheet. Each answers a different question, and mixing them up is a common reason a claim seems to say more than it does.
| Source | What it is | What it tells you |
|---|---|---|
| Study | One research team’s report on a group of people | What happened to the people studied, under the study’s conditions |
| Guideline | Recommendations based on reviewed evidence | What is advised for a defined group after relevant benefits, harms, evidence quality and costs are considered |
| Fact sheet | A plain-language summary for the public | A general picture that may not match your medicines, history or risks |
The 2018 spine-care review in PubMed shows the process clearly. Its authors adapted guidelines written for high-income countries for low- and middle-income communities, weighing benefits, harms, evidence quality, cost and feasibility. The same recommendation can look different once the setting changes.
A recommendation is a judgment built on evidence, and how firm that judgment is becomes a separate question, covered in the guide on recommendation strength and evidence certainty. When a fact sheet describes what is generally true, check it against the limits in when a health fact sheet is not enough.
The systematic review covered in Step 4 evaluated 26 published studies. The answer to which type of evidence is strongest depends on your question, so this guide gives no single ranking; the research library’s guide to meta-analyses and systematic reviews explains how those methods work.
What this means for you: before you act on a claim, name the source type. A fact sheet and a trial answer different questions.
Read next
How to Tell a Study, Guideline, and Patient Fact Sheet Apart: what each one can and cannot answer for you.
Once you know the source type, the numbers inside it are the next thing to check.
Step 3: Turn a percentage in a headline into real numbers
The headline says a thing cut risk by half, and it sounds decisive. The arithmetic behind that sentence can be much smaller than it sounds.
Here is an illustrative example, not taken from any particular study. Suppose 2 out of every 100 people in a group had an event over a year without a treatment, and 1 out of every 100 had one with it. The absolute difference is 1 person in 100. The relative difference is 50%, because 1 is half of 2. Both statements are true, and the headline uses only the second.
Ask which numbers the study reports before you decide the claim is big. The guide on how to calculate relative risk walks through the math, and a confidence interval shows the range of values the true result is likely to fall within.
What this means for you: when you see a percentage, ask compared with what, and how many people out of 100.
Read next
Relative vs. Absolute Risk Explained: why the same result can sound very different.
Numbers tell you the size of an effect. The next question is whether the product or treatment behind the claim was tested in a way you can see.
Step 4: Check product and supplement claims against the research
Many supplement pages say the studies back them. Your job is to find out what those studies tested, in whom, and for how long.
A 2023 narrative review in Innovations in Pharmacy argues that dietary supplements get less pre-market review than prescription medicines, and that weaker oversight has produced more harm than benefit. A narrative review is the authors’ summary of cases and arguments, not a controlled test. It reports 79,071 adverse events linked to supplements, reported to the Center for Food Safety and Applied Nutrition between 2004 and 2021.
Heart safety gets its own check. A 2017 systematic review in the Journal of Dietary Supplements examined 26 published studies of supplements and the QT interval, a timing measure on an electrocardiogram (ECG), a recording of the heart’s electrical rhythm. Sample sizes ranged from nine people to 206, and eight studies found a significant change in QT/QTc intervals. More than 30% showed some ECG changes, which the review says calls for continued safety assessment.
When you read a supplement page, check these three things:
- The dose the study tested, compared with the dose in the product
- The group studied, and whether it resembles you
- The length of the study, and whether it measured the outcome you care about
For a plain walk through what ingredient studies do and do not show, read the research library’s guide to ingredient evidence versus marketing claims. If a seller says its product is supported by research, that is the seller’s statement, and the cited study is the thing to check. Ask the seller which study, which dose and which group it describes.
What this means for you: a supplement can have a label and a study and still carry an open safety question. Check both before you decide.
Read next
What Insufficient Evidence Means in Plain Language: what it means when studies are too few to settle a question.
If an option still looks possible after these checks, the clinician visit is where it gets sorted for your situation.
Step 5: Sort your real options with a clinician
You have narrowed the claim and want to know what options exist. This guide looks at three kinds of options covered in the sources here: lifestyle changes, prescription treatments and supplements.
Lifestyle. A 2020 review in Diabetologia looked at exercise for blood sugar control in people with type 2 diabetes. It found that combined aerobic and resistance training improved glycaemic control (how well blood sugar stays in range) more than either type alone, and that the benefit of aerobic exercise seemed to rise with intensity. The review says the most beneficial regimen is still unknown.
Prescription options with evidence. A 2017 narrative review in the Journal of Pain and Symptom Management covers two topical treatments for peripheral nerve pain that it describes as licensed. In a head-to-head comparison with pregabalin, an oral standard treatment, the capsaicin 8% patch was reported as noninferior across a range of causes. Noninferior means not meaningfully worse than the standard, by a margin set in advance. The lidocaine 5% plaster gave similar relief for post-herpetic neuralgia (nerve pain after shingles) but did not meet its noninferiority criteria.
Whether any option fits you depends on your history and medicines, which a clinician can check. Ask before starting or stopping any medicine.
Supplements. Bring the label and the study you found to the visit, and ask what the clinician thinks of them for your situation. The guide on tracking questions and follow-up after a visit helps you keep the answer.
What this means for you: bring the claim, its numbers and your question, not a conclusion you already drew.
Read next
How to Prepare Questions for a Preventive-Care Conversation: a short list to bring to your visit.
With your questions ready, the last step is making sure they survive the visit.
Step 6: Keep your questions alive and follow through
You now have a question list. This step makes sure it does not fade after the appointment.
A 2026 review in Stroke describes low-cost tools such as default appointments, digital reminders and simple workflow changes. It says these have boosted sticking with medicines, exercise and follow-up in some programs, mostly early pilots. Pilots are not proof for your situation, but they show why a written plan helps.
If you keep your own log of symptoms or readings, label who recorded it and when before you share it. The guide on home notes before sharing with a care team lists the questions to answer first.
What this means for you: a question you write down is one you can actually answer at your next visit.
Read next
Keep an Open-Questions List for Unclear Health Info: how to hold onto questions that are still open.
Your next step
- Write the claim in one sentence, and name who it was written for.
- Name the source type: study, guideline or fact sheet.
- Ask for the absolute numbers and the number of people studied.
- Check any product claim against the study it cites, including dose and group.
- Take your question list and your numbers to a clinician before deciding on any treatment or supplement.
This guide does not use a numbered level scale, because level systems differ between courses and organisations. For the system your program uses, ask a librarian or clinician which one applies.
Every guide on this topic
When you are starting with a claim
- Questions to ask when health information may not fit you: one question at a time.
- Why who was studied matters: participants and what they share with you.
- How to read a health recommendation: who it was written for.
- When a health fact sheet is not enough: where general summaries stop.
- Study, guideline and fact sheet: telling the three apart.
- What insufficient evidence means: when studies cannot settle a question.
- Funding and conflict disclosures: what a study’s funding statement tells you.
- Tracing a health claim to its original source: finding the study behind a headline.
- Study result versus author interpretation: what the data show and what authors say.
- How to read a health study abstract: the short summary at the top of a paper.
- Preprint versus peer-reviewed study: how the two differ.
- Correlation and causation in health research: why linked is not the same as caused.
- Observational study versus clinical trial: what each design can show.
- Four phases of clinical trials: how a treatment is tested in stages.
- Which clinical trials pay the most: a guide to paid trial participation.
- Reading a ClinicalTrials.gov record: what a registry entry lists.
When you are checking the numbers
- Relative vs. Absolute Risk Explained: why percentages can mislead.
- How to calculate relative risk: the steps, with the arithmetic shown.
- Statistical significance and practical importance: a result can be real and still small.
- P-values in health studies, in plain language: what the number means.
- Confidence intervals in plain language: the range around an estimate.
- Number needed to harm and baseline risk: why time period and starting risk matter.
- Surrogate outcomes versus patient outcomes: a marker is not always the outcome you care about.
- Why the study outcome measure matters: what the study counted.
- Composite endpoints and which events drove the claim: one number, several events.
- Reading a health test accuracy claim: accuracy is not a diagnosis.
- Study protocol versus published report: how the plan and the paper can differ.
- Health study corrections, expressions of concern and retractions: when a paper changes after publication.
When you are judging how strong the evidence is
- Why evidence gets downgraded: five questions in plain language.
- Recommendation strength vs. evidence certainty: two separate questions.
When you are studying the skill itself
- Top 5 Study Techniques: A Practical Guide: study habits for reading research.
- Study methods for medical students: learning to read research well.
When you are preparing for and tracking your visit
- How to prepare questions for a preventive-care conversation: a short list to bring.
- How to track questions and follow-up after a visit: keeping the answer.
- Home notes before sharing with a care team: labeling your own logs.
- Keep an open-questions list for unclear health info: holding onto unanswered questions.
Health Evidence Literacy Guide is for education and does not replace care from a qualified professional.
By Health Evidence Literacy Guide Editorial Team