Absolute versus relative risk: read the numbers behind a claim
A hypothetical worked example shows why a percentage reduction needs a baseline risk, a defined outcome and a time frame before it becomes useful.
Absolute risk describes how often an outcome occurs in a specified group over a specified time. Relative risk compares that risk between groups. To understand a percentage reduction, ask: compared with what baseline, for which outcome, in whom and over how long?
The numbers below are invented solely to explain the arithmetic. They are not results from a trial, estimates of your risk, or evidence that any test or treatment works.
The same hypothetical result, described three ways
Imagine two groups of 1,000 people followed for the same five years. A defined unwanted outcome occurs in 20 people in Group A and 15 in Group B.
| Description | Calculation | Result |
|---|---|---|
| Absolute risk in Group A | 20 divided by 1,000 | 2% over five years |
| Absolute risk in Group B | 15 divided by 1,000 | 1.5% over five years |
| Absolute risk reduction | 2% minus 1.5% | 0.5 percentage points, or 5 fewer in 1,000 |
| Relative risk, B compared with A | 1.5% divided by 2% | 0.75 |
| Relative risk reduction | 1 minus 0.75 | 25% |
“25% lower risk” and “5 fewer people in 1,000 over five years” describe the same hypothetical difference. Neither is a promise about an individual. The difference also does not establish causation without information about how the groups were formed and studied.
The NCI definition of relative risk describes a comparison between risks. NICE's risk communication recommendations favour absolute risk and natural frequencies when explaining benefits and harms. The arithmetic here is our own example, not a NICE or NCI study result.
Change the baseline and the absolute difference changes
In a second hypothetical example, use the same 25% relative reduction but start at 4 events per 1,000 over five years. The comparison becomes 4 versus 3 per 1,000: an absolute difference of 1 per 1,000, or 0.1 percentage points.
The relative reduction is unchanged; the absolute difference is smaller. This is why a headline percentage without the baseline is incomplete. These examples say nothing about whether a particular intervention would produce the same relative effect in another population.
Keep these fields beside a percentage claim
Print this short reading sheet or copy it into your notes. Leave missing information marked unknown.
| Question | Answer from the source |
|---|---|
| What exact outcome is counted? | ______ |
| Who was studied, and what was the comparison? | ______ |
| How many events occurred out of how many people? | ______ |
| Over what time, using the same time frame for both groups? | ______ |
| What is the absolute difference? | ______ |
| How uncertain is the estimate? | ______ |
| What harms or burdens were measured alongside benefit? | ______ |
What this calculation cannot settle
A risk difference does not show whether the research applies to you. It does not replace looking at study design, uncertainty, harms or practical consequences. A change in detection rate also cannot simply be relabelled a reduction in deaths or an increase in lifespan: those are different outcomes.
Use the longevity evidence worksheet to examine the whole claim and screening versus diagnostic testing to identify the purpose of a proposed test. Discuss individual decisions with an appropriate clinician.
Sources checked on 12 October 2026. All numerical examples are explicitly hypothetical; this article is editorial education, not licensed clinical review.
Sources
Applying the evidence
A verdict about a claim is not an individual screening plan. Read the screening guide and discuss your history, prior results and possible follow-up with a qualified clinician.