Question explored with the scientific record
Warum ist das absolute Risiko aussagekräftiger als das relative Risiko?
Relative risk tells you how much the chance shrinks. Absolute risk tells you what you actually face.
A relative risk of 0.70 sounds dramatic until you see the baseline. In one meta-analysis of 21 statin trials, the relative drop in all-cause death was 9%, but the absolute drop was only 0.8% [2]. Same drug, same trials, two very different pictures. The relative number hides how rare the event already was.
The clearest demonstration is pravastatin in kidney disease. The relative effect was similar across kidney function groups, but because people with moderate chronic kidney disease started at higher risk, their absolute benefit was larger: a 6.3% absolute reduction in cardiovascular events versus a 2.3% reduction in all-cause death [1]. A Korean cohort found the same pattern: relative benefits were similar across risk groups, so absolute gains were bigger in the high-risk group [3]. And the same logic applies to harms: PPI-attributable cases concentrated in the highest-risk 40%, with a risk difference of 2.94% versus 1.07% in the lowest 20% [4].
| Measure | What it tells you | Example from records |
|---|---|---|
| Relative risk | How much the chance changes | 9% relative drop in death [2] |
| Absolute risk | The real change in your odds | 0.8% absolute drop in death [2] |
| Absolute benefit, high-risk | Bigger when baseline risk is high | 6.3% cardiovascular reduction [1] |
The reason is arithmetic. A 50% relative reduction on a 1-in-100,000 risk saves half of one person per 100,000. The same 50% on a 1-in-100 risk saves 50 per 100. Relative risk flatters interventions aimed at rare events and understates them where risk is high. Absolute risk also lets you weigh harms against benefits on the same scale, which relative risk never does.
My call: absolute risk is more informative because it tells you what actually changes for a real person, and relative risk alone can make a tiny effect look large. Confidence: high.
Sources used 4
-
Effect of Pravastatin on Cardiovascular Events in People With Chronic Kidney Disease
A post hoc analysis of three randomized pravastatin trials (PPP) shows that pravastatin reduces cardiovascular events and all-cause mortality in people with or at high risk for coronary disease who have moderate chronic kidney disease (GFR 30–59.99), with similar relative effect…
DOI: 10.1161/01.CIR.0000143892.84582.60 -
Statin Use and the Risk of All-cause Mortality
Statins reduced all-cause mortality in some trials, but a meta-analysis found only a 0.8% absolute reduction, leaving the effect uncertain and prompting calls for trials with all-cause mortality as the primary endpoint.
DOI: 10.3904/kjm.2023.98.1.4 -
Statin use and outcome risks according to predicted CVD risk in Korea: A retrospective cohort study
In Korean adults without CVD, regular statin use was associated with lower major CVD events and mortality; relative benefits were similar across predicted risk, so absolute benefits were greater at higher predicted risk.
DOI: 10.1371/journal.pone.0245609 -
Individualized prevention of proton pump inhibitor related adverse events by risk stratification
In five prospective cohorts (2,079,724 participants), proton pump inhibitor use was associated with increased risk of 15 of 30 leading global-burden diseases, with dose-response relationships and risk stratification concentrating PPI-attributable cases in high baseline-risk indi…
DOI: 10.1038/s41467-024-48007-8