Absolute Versus Relative Risk, and the One Question That Cuts Through the Hype

When a headline says a treatment "cuts your risk by half," the most useful question you can ask is: half of what? A relative number tells you the proportion by which a risk changed. An absolute number tells you how many people that change actually touches. The same study can honestly report a 50 percent drop and a one-in-a-thousand drop at once, because those describe the identical result from two angles. The relative figure makes the news. The absolute figure tells you whether to care. This is educational, not medical advice, so use it to read the numbers better, then talk through your situation with your clinician.

I have spent years on the producing side of these numbers as well as the reading side. My doctoral research at the Lund University Diabetes Centre is on the genetics of type 2 diabetes, where risk is the whole vocabulary, and I co-authored a meta-analysis in Diabetes Care that pooled effect sizes across many studies. Pooling teaches you fast that a percentage floating free of its baseline is almost meaningless. The number that survives scrutiny is anchored to how common the event was in the first place.

What is the difference between absolute and relative risk?

Here is the short, quotable version. Relative risk compares the chance of an event between two groups as a ratio or a percentage change. Absolute risk is the plain chance of the event in a group, expressed as a number out of some total, like 2 in 100. Relative risk reduction is the percentage drop between the two groups; absolute risk reduction is the simple subtraction of one plain chance from the other.

Those two ideas sound close, and that closeness is where readers get misled. A relative change can look enormous while the absolute change is tiny, because it hides the baseline. If a risk falls from 2 percent to 1 percent, the relative reduction is 50 percent, which sounds like a breakthrough. The absolute reduction is one percentage point, the part you can actually feel. Both statements are true. Only one tells you the scale.

A neutral worked example with plain numbers

Let me use invented round numbers so nothing rides on a particular product. Imagine a trial of 2,000 people in two equal groups of 1,000, watched for some unwanted event over five years. In the comparison group, 40 have the event, an absolute risk of 4 percent. In the treated group, 30 have it, an absolute risk of 3 percent.

From those two arms the derived measures follow directly: a relative risk reduction of 25 percent, an absolute risk reduction of one percentage point, and a number needed to treat of about 100.

Now read that two ways. The relative risk reduction compares the rates: going from 40 events to 30 is a drop of 10 out of 40, which is 25 percent. The absolute risk reduction is simpler, 4 percent minus 3 percent, or one percentage point. So the honest summary is that this approach reduced the event by 25 percent in relative terms and by one person in a hundred in absolute terms. A press release reaches for the 25 percent. A careful clinician holds both.

There is a third number that ties the example together and is, in my view, badly underused. If one extra person in a hundred avoids the event, you would need to treat about 100 people for one to benefit. That figure is the number needed to treat, the inverse of the absolute risk reduction. A number needed to treat of 100 means 99 people took on whatever the treatment involves, its cost and inconvenience and side effects, without getting the payoff being measured. That is not a reason to refuse it. It is the context that lets you decide.

Why relative numbers dominate the headlines

The pull toward relative figures is rarely dishonesty, and the people writing these summaries are mostly working in good faith. Relative numbers travel well. They are larger, they sound more decisive, and they hold steady across populations with very different baseline risks.

That convenience is also the catch. Because a relative number ignores the baseline, it says nothing about how much room there was to move. Drop a 40 percent baseline by a quarter and you have spared 10 people in a hundred. Drop a 0.4 percent baseline by the same quarter and you have spared one person in a thousand. Identical relative figure, hundredfold difference in what it does for real people.

How to translate any risk claim in under a minute

You do not need statistics training for this. You need to refuse a percentage without its baseline. When you meet a claim like "lowers risk by a third," ask a few plain questions. A third of what starting number? If nobody can tell you the baseline risk, the relative figure is unreadable, and you should treat it as incomplete rather than impressive. What is the risk with and without the intervention, each stated as a number out of 100 or out of 1,000? Subtracting those two absolute risks gives you the absolute reduction directly. And how many people would need the treatment for one to benefit, which is simply one divided by the absolute reduction expressed as a fraction.

The same discipline works for harms. A scary "doubles your risk" deserves the identical interrogation. Doubling a 1 in 10,000 risk lands you at 2 in 10,000, a real change and still a very small number. Doubling a 1 in 10 risk is a different conversation. Direction without magnitude is half a fact.

Where this matters most

Two situations deserve extra care. The first is prevention in people who feel well, where baseline risk is often low and a striking relative reduction can shrink to a modest absolute one. The second is screening and risk-marker testing, where a "raised risk" can sound alarming while the absolute chance over a meaningful timeframe stays small. In both, the relative figure frightens or excites, and the absolute figure should drive the decision alongside your own values and your clinician's read of your particular picture.

None of this makes relative risk a villain. For comparing the strength of an intervention across groups it can be the right tool. The error is presenting it alone, stripped of the baseline that gives it meaning. So keep the habit simple. Whenever you are handed a percentage about your health, ask for the plain numbers underneath it: how common was the event, how common is it now, and how many people share the effort for each one who gains. A claim that answers those questions is worth your attention. One that cannot is not yet finished.