Reading Study Numbers: A Plain-Language Guide to Health Figures

Maybe a headline said a treatment "cuts major events by 30%," and you wondered whether that means you personally are 30% safer. Or maybe a study you were sent is full of percentages, p-values and "confidence intervals," and you feel you should already know what they mean. You are not behind. Most people were never taught this, and the numbers are often written to sound more certain than they are.

The honest short answer is that a study number only means something once you know what was counted, how many people were involved, how long they were followed, and what they were compared with. This guide takes you through those questions one step at a time, in the order you would meet them.

Step 1: Find out what is being counted

You are here because a number stopped you. Start with what it counts, since a single figure can combine several different events. Reading long numbers is itself effortful: in an eye-tracking study of adults indexed by PubMed, long Arabic numerals drew up to 7 times more eye fixations than matched long words.

A "30% fewer major events" figure can hide that one kind of event moved a lot while another barely moved. The Composite Endpoints guide shows how to find which events drove the result.

Next, ask whether the study tracked something that matters to you directly. A stand-in improving does not automatically mean the outcome you care about improved. The Surrogate Outcome guide lists the questions to ask, and the Health News Headline guide shows how headlines compress these details.

Once you know what was counted, you can ask how big the effect is. That is the next step.

Step 2: Turn a percentage into real people

The two can look very different, so you need both.

Here is an illustration with invented round numbers, not taken from any study. Suppose 2 people in every 100 in a comparison group had an event, and 1 person in every 100 in a treated group did. The relative reduction is 50%, but the absolute difference is 1 percentage point, or 1 fewer person in every 100.

The How to Calculate Relative Risk guide shows the arithmetic step by step, and the Relative vs Absolute Risk page compares the two side by side. The Number Needed to Harm guide explains why the time period and baseline risk matter.

Size alone does not tell you whether a result is reliable. That is the next question.

Step 3: Check how sure the result is

A p-value is a number that is often reported alongside a result. It does not measure how big the effect is. The P-Value guide explains what it can and cannot tell you.

Read the Confidence Interval guide for how to read one.

Statistical significance is a label applied to a result in a study. That is not the same as the result being large or useful to you. The Statistical Significance guide separates the two questions.

Certainty also depends on how the study was designed and who it included. That comes next.

Step 4: Check who was studied and how

The design of a study sets how far you can lean on its results. The Observational Study or Clinical Trial page explains what the design changes, and the Four Phases of Clinical Trials guide shows how trials are staged.

Size matters too. A 2017 systematic review indexed by PubMed looked at 26 published dietary supplement studies of heart rhythm. Sample sizes ranged from nine people to 206, and the authors concluded the studies were typically well performed but could be bolstered by design changes.

Small groups make results less stable, and a single study can disagree with the rest of the research. The Study Population guide asks who was included, and the Systematic Review guide explains what pooling studies can and cannot settle. The Evidence Downgrading guide lists the questions that lower how much weight a result deserves, and the Health Study Methods guide shows where to find these details.

A number can also be right in a study and still be misreported on its way to you. That is where tracing it helps.

Step 5: Trace the number back to its source

Numbers reach you through headlines, summaries and product pages, and each step can change them. Begin with the abstract, the short summary at the top of a study, and then read the full report. The Health Study Abstract guide shows what to look for, and the Trace a Health Claim guide walks you back to the original paper.

The result and the author's conclusion are not always the same thing. The Result vs Interpretation guide helps you separate them. If a study was registered with a plan, compare that study protocol with the published report to see whether the outcomes changed along the way.

A preprint is posted before peer review, so it has not yet gone through that check. The Preprint vs Peer-Reviewed guide explains how to tell them apart. Check who paid for the work and what conflicts were declared with the Funding and Conflict Disclosures guide, and look for later corrections with the Corrections and Retractions guide.

Supplement claims deserve extra care. A 2023 narrative review indexed by PubMed argued that FDA oversight of dietary supplements has been limited and linked to many adverse event reports. That is one review's argument, and you can weigh it alongside other sources. For the regulator's own information on supplements, see the FDA. The ClinicalTrials.gov record guide shows how to read a registered trial listing.

Once you have the right source, you can ask whether its result applies to your own situation.

Step 6: Decide whether the number applies to you

An accuracy claim for a health test describes how that test performed in a study, not what your own result means. The Test-Accuracy Claim guide separates study statistics from your personal situation. The Does It Apply to You guide gives you the questions to ask.

Reading numbers can be harder for some people. A study of 30 young adults with dyslexia indexed by PubMed found they had severe difficulty reading and writing numbers, compared with a typically developing group. If numbers are hard for you, ask for a plain-language explanation and take your time.

Not every summary carries the full picture. The Health Fact Sheet guide explains when a summary is not enough, and the Recommendation Strength guide shows how guidance rests on evidence of different certainty. When results are thin, the Insufficient Evidence guide explains what that means for you. If you are learning to read studies yourself, the Study Methods for Medical Students guide and the Top 5 Study Techniques guide cover study habits. This page is about health study numbers, so if you came for school math or Bible study, those guides may fit better.

With the numbers and their source in hand, you are ready for the last step.

Step 7: Take the numbers to your clinician

A clinician can check a study's numbers against your history and your situation. Bring the exact figure, the source, and the question you want answered. The Patient-Generated Health Information guide explains what to think about before sharing your own notes with a care team.

If a trial is cited, ask your clinician to help you read its registered record. The ClinicalTrials.gov guide shows what the listing includes. Do not start, stop or change any medication based on a number you read; ask your clinician first.

Your next step

1. Write down exactly what the number counts, and in how many people.
2. Ask for both the relative and the absolute figure.
3. Check the confidence interval and whether the study was a trial or observational.
4. Find the original source and compare its result with the headline.
5. Bring the figure, the source and your question to your clinician.

Every guide on this topic

Start with what is counted

Composite Endpoints: which events drove a combined result.

Surrogate Outcomes: stand-in measures versus outcomes you feel.

Outcome Measures: why the tracked outcome matters.

Health News Headlines: what a headline leaves out.

Percentages and people

Relative and Absolute Risk: what each figure hides.

How to Calculate Relative Risk: the arithmetic step by step.

Relative vs Absolute Risk: a side-by-side comparison.

Number Needed to Harm: how time period and baseline risk matter.

How sure is the result

P-Values: what they can and cannot tell you.

Confidence Intervals: how wide the range of results is.

Statistical Significance: unlikely chance versus practical importance.

Study design and who was included

Observational Study or Clinical Trial: what the design changes.

Four Phases of Clinical Trials: how trials are staged.

Study Population: who was included and why it matters.

Systematic Reviews: what pooling studies can and cannot settle.

Evidence Downgrading: questions that lower how much weight a result gets.

Health Study Methods: where to find design details.

Source and fine print

Health Study Abstracts: reading the summary at the top.

Trace a Health Claim: getting back to the original paper.

Result vs Interpretation: separating the finding from the conclusion.

Study Protocol vs Published Report: whether the plan and the report match.

Preprint vs Peer-Reviewed: what has and has not been checked.

Funding and Conflict Disclosures: who paid and what was declared.

Corrections and Retractions: how published studies are changed.

ClinicalTrials.gov Records: reading a registered trial listing.

Accuracy and whether it fits you

Health Test-Accuracy Claims: study statistics versus your result.

Does Health Information Apply to You: questions for fit.

When a Health Fact Sheet Is Not Enough: the limits of summaries.

Recommendation Strength: how certainty shapes guidance.

Insufficient Evidence: what thin results mean for you.

Questions for your clinician

Patient-Generated Health Information: what to ask before sharing notes.

Study habits

Study Methods for Medical Students: habits for reading research.

Top 5 Study Techniques: a practical study routine.

Health Evidence Literacy Guide is an independent publication. These guides are for education and do not replace care from a qualified professional.

By Health Evidence Literacy Guide Editorial Team