Relative Risk and Absolute Risk Are Different Numbers

The points below about relative and absolute risk reporting are ordered by how much difference they make, not by how often they get repeated.

Two descriptions of one result

Relative risk expresses how much a risk changes as a proportion of what it was before. Absolute risk expresses the change in terms of how many people out of a defined group are affected. Both can be entirely accurate descriptions of the same underlying finding, which is what makes the choice so powerful.

Relative figures sound larger whenever the starting risk is small, which is the case for most individual health risks. Which version is quoted tells you something about the intentions of whoever is quoting it.

An illustration with invented numbers

Take a purely hypothetical example, chosen for arithmetic rather than drawn from any real study. If a condition affects two people in ten thousand and a factor raises that to three, the relative increase is fifty per cent.

The absolute increase is one additional person in ten thousand, which most readers would judge quite differently. Both statements describe the same imaginary result, and one of them will appear in the headline. Working through this once with made-up numbers makes the pattern obvious in every real report afterwards.

Why baseline risk is decisive

A large relative change to a tiny risk remains a tiny risk, while a small relative change to a large risk matters greatly. Baseline risk depends on age, sex, family history, existing conditions and other factors specific to the person. A result reported for a population may correspond to very different absolute changes for different individuals within it.

This is why the same finding can be genuinely important for one group and irrelevant for another. Asking what my starting risk is converts a headline into something that can actually inform a decision.

Number needed to treat

This figure states how many people must receive a treatment for one of them to benefit. It translates trial results into a form that is unusually easy to grasp and unusually hard to spin. A low number indicates a treatment that helps many of the people who take it.

A high number indicates a treatment that helps a small minority, which may still be worthwhile for a serious outcome. The equivalent figure for harm is also calculable, and comparing the two is how benefit and risk get weighed.

Where this shows up in wellness

Supplement marketing quotes relative reductions from observational studies without mentioning the underlying rates. Food scares quote relative increases for exposures whose absolute contribution is very small. Screening promotions quote relative mortality reductions without the absolute figures or the harms of false positives.

Against the trial data, in each case the omitted number is the one that would let a person decide whether to care. The omission is legal, accurate and thoroughly misleading, which is the recurring shape of this whole subject.

What to ask when you meet a percentage

Ask a percentage of what, and out of how many people over what period. Ask for the absolute figures in both groups rather than the difference between them.

Against the trial data, ask whether the finding came from an observational study, where confounding can produce associations without causation. Ask whether the risk applies to people with your characteristics or to a broad population average. Any decision involving your own medical treatment should be made with a clinician who can supply your baseline.

Everything above, in order of what to do first

1. Two descriptions of one result. Relative risk expresses how much a risk changes as a proportion of what it was before.
2. An illustration with invented numbers. Take a purely hypothetical example, chosen for arithmetic rather than drawn from any real study.
3. Why baseline risk is decisive. A large relative change to a tiny risk remains a tiny risk, while a small relative change to a large risk matters greatly.
4. Number needed to treat. This figure states how many people must receive a treatment for one of them to benefit.
5. Where this shows up in wellness. Supplement marketing quotes relative reductions from observational studies without mentioning the underlying rates.
6. What to ask when you meet a percentage. Ask a percentage of what, and out of how many people over what period.

The takeaway

Ask what the risk was before and what it became. A percentage without a baseline is a number designed to be repeated, not understood.

An extraordinary mechanism needs better evidence than testimonials, and usually has less.