Learn

Relative Risk vs Absolute Risk: How to Read Health Numbers

Advertiser and Editorial Disclosures

Advertiser Disclosure

Wellness Pro News is supported by advertising. Some links and placements on this site are paid, and we may earn revenue when readers interact with them. Advertising never determines which topics we cover, which products we mention, or the conclusions our writers reach. Paid placements are labeled where they appear.

Editorial Disclosure

Our articles are researched and written by our editorial team and reviewed before publication. Health claims are checked against peer-reviewed research and recognized medical authorities, and we update articles when the evidence changes. Nothing on this site is medical advice or a substitute for care from a qualified clinician.

Emma Pollner

Emma Pollner, PhD, Clinical Research

Relative Risk vs Absolute Risk: How to Read Health Numbers

A headline says a habit "cuts risk in half." Another says the same habit lowers risk "by one percentage point." Both can describe the same result. The difference is whether the writer chose a relative figure or an absolute one. Knowing how each number is built makes health news much easier to read. This guide defines the main terms, works through one made-up example with round numbers, and lists what to look for when a story gives only one kind of figure.

Two ways to describe the same result

Absolute risk is the plain starting point. NIH News in Health, in an October 2016 article on understanding health risks, describes it as the chance of something happening, such as a health problem that might arise over a period of time. It is usually written as "2 in 100" or "2%."

Relative risk is a comparison. The same NIH article says relative risk compares the absolute risks of one group to another. The CDC textbook Principles of Epidemiology in Public Health Practice (Third Edition) calls this figure a risk ratio. It is found by dividing the risk in the group of main interest by the risk in the comparison group.

The CDC lesson also explains how to read the answer. A risk ratio of 1.0 means the two groups have identical risk. A ratio above 1.0 points to a higher risk in the group of interest. A ratio below 1.0 points to a lower risk.

Neither number is wrong. Absolute risk says how often something happens. Relative risk says how one group compares with another.

A made-up example with round numbers

The following illustration is invented. It is not from any study, and umbrellas were chosen so that no real health finding is implied.

Suppose 100 people carry umbrella design A for one year, and 2 of the umbrellas break. Another 100 people carry design B for the same year, and 1 breaks.

  • Absolute risk with design A: 2 in 100, or 2%.
  • Absolute risk with design B: 1 in 100, or 1%.
  • Absolute difference: 2% minus 1%, which is 1 percentage point.
  • Risk ratio: 1% divided by 2%, which is 0.5.
  • Relative difference: 1 point divided by the starting 2%, which is 50%.

So "design B cuts breakage by 50%" and "design B lowers breakage by 1 percentage point" are both accurate. NIH News in Health uses the same round numbers, 2 in 100 falling to 1 in 100, to make this point. It notes that the relative figure often sounds much greater than the absolute one, and it suggests focusing on absolute risk when numbers about risk are reported.

The terms and their formulas

Researchers who test a treatment use set names for these figures. A University of Nottingham teaching resource defines them using two inputs: the control event rate (CER), which is how often the event happens in the comparison group, and the experimental event rate (EER), which is how often it happens in the group being tested.

Term Plain meaning Formula Source
Absolute risk The chance of something happening over a period of time People who had the event ÷ all people in the group NIH News in Health (2016); Family Practice (2022)
Relative risk (risk ratio) Compares the risk in one group with the risk in another Risk in group of interest ÷ risk in comparison group CDC, Principles of Epidemiology, Third Edition
Absolute risk reduction The difference between the control event rate and the experimental event rate CER − EER University of Nottingham
Relative risk reduction The difference in event rates relative to the control event rate, shown as a percentage (CER − EER) ÷ CER University of Nottingham
Number needed to treat How many people would need to be treated for one of them to have the beneficial outcome 1 ÷ absolute risk reduction University of Nottingham

When the group of interest has the higher risk, the same arithmetic describes an increase instead of a reduction. The risk ratio is then above 1.0, as the CDC lesson explains.

The number needed to treat turns the absolute difference into a head count. In the umbrella illustration, the absolute difference is 1 in 100, or 0.01. One divided by 0.01 is 100. Put loosely, 100 people would need to carry design B for a year for one fewer umbrella to break.

Why the starting risk matters

A relative figure hides the starting point, which is often called the baseline risk. A 2022 article in the journal Family Practice describes the baseline risk as the risk in the comparison group. It notes that the relative risk reduction does not consider the actual risk in a population, while the baseline risk does influence the absolute risk reduction.

The umbrella example shows this. Change the made-up numbers so that 20 in 100 of design A break and 10 in 100 of design B break. The relative difference is still 50%. The absolute difference is now 10 percentage points instead of 1. The same "50%" covers both a small change and a large one.

This is why the National Cancer Institute's Division of Cancer Epidemiology and Genetics states that, for clinical and public health uses, absolute risk is typically more useful than relative risk.

Time frame and who was counted

A risk figure always belongs to a period of time. The NIH definition of absolute risk includes the words "over a period of time," and the NIH article suggests asking whether a risk applies over a lifetime or over a shorter period. "2 in 100 over one year" and "2 in 100 over thirty years" are very different statements.

The group matters too. NIH News in Health points out that if the people behind a statistic are not similar to you, or the category is very broad, your own risk may be different. The National Center for Complementary and Integrative Health makes related points in its Know the Science checklist for health news stories. The checklist asks whether enough people were studied, whether the participants were similar to you in ways that may matter, and whether the study was lengthy enough to show long-term benefits or risks.

How a number is worded also shapes how it lands. NIH News in Health reports that the way health statistics are described, or "framed," can influence how people understand them. It adds that pictures or diagrams are often the most understandable format.

What to check when a headline gives only a relative figure

  1. Look for the starting risk. Find how many people out of 100 or 1,000 had the outcome in the comparison group.
  2. Work out the absolute difference. Subtract one group's risk from the other's.
  3. Find the time frame. Check whether the risks cover one year, ten years or a lifetime.
  4. Check who was counted. See whether the people studied resemble the reader in ways that may matter.

People with a health condition or specific needs should talk with their own clinician or a registered dietitian about what a risk figure means for them.

Try the numbers yourself

Enter two risks below to see how they compare. The calculator shows the absolute difference and the relative difference side by side.

Risk calculator

Enter how many people out of each group had the outcome. The results show the same comparison as an absolute difference and as a relative difference.

Comparison group out of people
Other group out of people

  • Risk in the comparison group
  • Risk in the other group
  • Absolute difference
  • Relative difference
  • Risk ratio

For learning how the numbers work. It does not estimate anyone's personal risk.

Frequently asked questions

Is a relative risk figure misleading?

Not by itself. It is a correct comparison of two groups. It becomes hard to interpret when the starting risk is left out, because the same percentage can describe a very small or a very large absolute change.

What is the difference between "percent" and "percentage points"?

A percentage point is the plain gap between two percentages. Going from 2% to 1% is a drop of 1 percentage point. It is also a 50% relative drop, because 1 is half of 2.

Why does the number needed to treat change from study to study?

It is calculated as 1 divided by the absolute risk reduction, as the University of Nottingham resource shows. The absolute risk reduction depends on the baseline risk, so groups with different starting risks give different results.

Sources

All sources accessed November 16, 2026.

Last edited on
Emma Pollner
Emma Pollner

PhD, Clinical Research

Emma Pollner, PhD, has a background in clinical research. She writes and reviews our coverage of health research. LinkedIn: linkedin.com/in/emma-pollner

Was this article helpful? Let us know!