Home / Evidence & Research Literacy
Absolute and Relative Effects
Absolute and relative effects describe the same comparison in different ways, so both are needed to understand the size of a reported benefit or harm.
#Two views of the same comparison
A risk is the probability of a specified outcome in a specified group. A risk ratio divides the risk in one group by the risk in the comparison group. A risk difference subtracts one risk from the other. Relative and absolute measures describe different aspects of the same comparison.
Read which outcome is counted, which groups are compared and how long they were followed. The sign of a difference depends on the order of subtraction. Whether an increase is helpful or harmful depends on the outcome: recovery and illness are not interchangeable events.
Evidence: Cochrane Handbook, Chapter 6: effect measures / Cochrane Handbook, Chapter 15: interpreting effects and uncertainty
#A percentage is not a percentage point
Consider a hypothetical comparison in which an adverse outcome occurs in 4 out of 100 people in one group and 3 out of 100 in the other, over the same follow-up. The risk ratio is 3 divided by 4, or 0.75. That is a 25% relative reduction. The absolute reduction is 1 percentage point, equivalent to 10 fewer events per 1,000 people.
These are two descriptions of the same numbers, not two different benefits. This is an arithmetic example, not a result for a medicine. Report the risks in both groups alongside the relative change so readers can see the scale of the comparison.
Evidence: Cochrane Handbook, Chapter 6: effect measures / Cochrane Handbook, Chapter 15: interpreting effects and uncertainty
#Baseline risk changes the absolute effect
A relative change cannot tell you the absolute difference without the comparison risk. Applying a study estimate to another population requires a suitable baseline risk and an assumption that the relative effect applies there. Neither assumption should be hidden.
Keep the population, comparator, outcome and time frame with the estimate. A risk from one setting may not describe another setting. Cochrane recommends presenting both relative and absolute effects for binary outcomes, with uncertainty, rather than treating one headline percentage as the full result.
Evidence: Cochrane Handbook, Chapter 6: effect measures / Cochrane Handbook, Chapter 14: certainty of evidence
#Check the measure and the limits
An odds ratio is not a risk ratio. Odds compare the probability of an event with the probability of no event; the two ratios can differ substantially when events are common. Do not translate an odds ratio directly into a percentage change in risk.
Effect size also does not establish cause by itself. Read study design, bias, uncertainty, benefits and harms before drawing a conclusion. A large relative reduction can still be a small absolute difference. Neither measure alone decides whether an intervention is appropriate for a person.
Evidence: Cochrane Handbook, Chapter 6: effect measures / CDC Field Epidemiology Manual: analyzing and interpreting data / Cochrane Handbook, Chapter 15: interpreting effects and uncertainty
Source note
The sections above were checked against the linked sources. No clinical review has been performed. This is general research education, not a clinical guideline.