Lucent Explainers

Reading a Clinical Trial Result: Relative vs Absolute Risk

3 November 2026

Most drug headlines report relative risk, which describes the size of the improvement compared with the old risk, not compared with zero. The gap between that number and absolute risk, the real change in your chances, is where headlines get their shine. Once you know which one you are looking at, a trial result stops being persuasive and starts being useful.

The Trick Behind the Headline

Say a new drug cuts the risk of a heart attack by 50 percent. That sounds dramatic. But 50 percent of what? If the drug lowers heart attack risk from 2 in 100 people down to 1 in 100 people, that is indeed a 50 percent reduction, because 1 is half of 2. The relative risk reduction is 50 percent.

But look at what actually changed for a person taking the drug. Out of 100 people, one fewer person had a heart attack. That is an absolute risk reduction of 1 percentage point. Both numbers describe the same result. One sounds like a breakthrough. The other sounds modest. Neither is wrong, but only one tells you what to expect for yourself.

Journalists and press releases favour relative risk because it is almost always the bigger number. A company reporting 'cuts risk by half' gets more attention than one reporting 'reduces risk from 2 percent to 1 percent'. The statistics are identical. The impression they leave is not.

What Absolute Risk Actually Means

Absolute risk is the plain chance of something happening to you, stated as a fraction or percentage of a group. If a condition affects 4 people in 1,000 untreated, and 2 people in 1,000 treated, the absolute risk reduction is 2 in 1,000, or 0.2 percentage points.

This number answers the question a patient actually has in mind: how much does this change my odds? It does not get inflated by starting from a small base, because it reports the change in real terms rather than as a ratio.

The catch is that absolute risk reduction looks small for rare conditions even when the drug works well. A treatment that halves the risk of a disease affecting 1 in 10,000 people will reduce absolute risk by about 0.5 in 10,000. That is a genuinely useful effect at a population level, but it will never produce an eye-catching headline.

A Worked Comparison

Take a hypothetical cholesterol drug tested over five years. In the placebo group, 3 percent of participants have a heart attack. In the treatment group, 2 percent do. The relative risk reduction is 33 percent, because 2 is a third less than 3. The absolute risk reduction is 1 percentage point.

Now imagine the same relative effect in a higher risk group, where 30 percent of the placebo group have a heart attack and 20 percent of the treatment group do. The relative risk reduction is still 33 percent. But the absolute risk reduction is now 10 percentage points, ten times larger.

This is why relative risk alone can mislead in both directions. It can make a tiny absolute effect look large in a low risk population, and it can hide a genuinely large absolute effect in a high risk population behind the same headline number. You cannot judge the real-world size of a benefit without knowing the baseline risk it started from.

Why Trials Report It This Way

Relative risk is not reported purely for marketing. Statistically, it tends to stay more consistent across different populations than absolute risk, which makes it genuinely useful for researchers comparing results across studies or applying findings to groups with different starting risks.

The problem is not the statistic itself but which number gets quoted in public. Academic papers usually report both relative and absolute measures, often with confidence intervals attached to each. Press releases and news summaries tend to keep only the relative figure, because it is the one more likely to get a reader's attention.

Regulators and some medical journals have pushed back on this by asking for absolute numbers alongside relative ones, and for a plain statement of baseline risk. Where that information appears, it is usually buried a few paragraphs into the results section rather than in the abstract or the headline.

What To Ask When Reading One

Before accepting a relative risk figure, ask what the baseline risk was. A 50 percent reduction means something completely different depending on whether the starting risk was 2 percent or 20 percent. If a news article or press release gives you only the relative number, treat it as an incomplete sentence.

Look for the absolute numbers in the paper itself, usually stated as event rates in each group, such as '14 of 500 patients' against '28 of 500 patients'. From there you can calculate both the relative and absolute reduction yourself and judge which one matters for your situation.

Also check whether the trial measured something you actually care about, such as heart attacks or deaths, or a proxy measure like cholesterol level or blood pressure. A large relative reduction in a proxy measure does not guarantee a matching reduction in the outcome that affects your life.

Number Needed To Treat

One figure that cuts through the confusion is the number needed to treat, often shortened to NNT. It tells you how many people would need to take the treatment for one additional person to benefit, calculated as 1 divided by the absolute risk reduction.

In the first worked example above, with an absolute risk reduction of 1 percentage point, the NNT is 100. That means 100 people would need to take the drug for five years to prevent one extra heart attack. In the higher risk example, with a 10 percentage point reduction, the NNT is only 10.

NNT is popular with doctors and increasingly appears in patient information leaflets because it translates statistics into something closer to a real decision. A low NNT, such as 10, suggests a treatment with a meaningful personal benefit. A high NNT, such as 500 or more, suggests the benefit is real but thinly spread across a large population.

Common questions

Does a small absolute risk reduction mean the drug is not worth taking?

Not necessarily. A small absolute reduction can still be worthwhile if the treatment is cheap, has few side effects, and is being given to a very large number of people, since small individual effects add up across a population. The decision depends on weighing that benefit against cost, side effects, and how the risk compares with other things you could do instead.

Why do news articles almost always use relative risk?

Relative risk numbers are usually larger and sound more dramatic, which makes for a more compelling headline than a modest absolute figure. It is rarely dishonesty, more a habit of reporting the number that reads best, so the responsibility falls on the reader to seek out the absolute figures in the original study.

How do I find the absolute risk if a study only gives relative risk?

Look in the results section for the actual event counts in each group, something like '12 out of 600 patients' versus '24 out of 600 patients'. Subtracting the two percentages gives you the absolute risk reduction directly, without needing any further statistics knowledge.

Is number needed to treat the same for every patient?

No, NNT is an average calculated across the people in a particular trial, and it changes depending on a person's baseline risk. Someone with higher baseline risk than the trial population will typically benefit more, meaning a lower effective NNT for them, and someone with lower baseline risk will benefit less.

Can relative risk and absolute risk ever point in different directions?

They always point the same direction, since both describe the same underlying improvement, but they can disagree sharply on how large that improvement looks. This is exactly why checking both numbers matters more than trusting either one alone.

Lucent, once a week

We read the new research and write up what actually holds, in the same plain language as this page. No jargon, no hype, three papers a week.

Read the latest issue

Read next