Understanding Risk Ratio (Relative Risk)
A risk ratio (relative risk) is a measure that compares the probability of an outcome occurring in an exposed group versus an unexposed group, calculated by dividing the risk in the exposed group by the risk in the unexposed group. 1
Definition and Calculation
- Risk ratio (RR) represents the ratio of disease risk in the exposed group to that in the unexposed group 2
- The measure captures the strength of association between an exposure and disease 1
- It is calculated as: Risk in exposed group ÷ Risk in unexposed group 2
Interpretation
Basic Interpretation
- An RR of 1.0 indicates no difference in risk between groups (no association) 1
- An RR greater than 1.0 indicates increased risk in the exposed group 1
- An RR less than 1.0 indicates decreased risk (protective effect) in the exposed group 1
Clinical Significance
- If the relative risk is far from 1, it is less likely that the association is due to confounding 1
- Relative risks tend to be more consistent across studies and populations than absolute measures 1
- For example, similar relative risks were obtained for cardiovascular risk factors across men in Northern Ireland, France, USA, and Germany, despite substantially different underlying coronary heart disease rates 1
Critical Distinction: Relative Risk vs. Absolute Risk
Why Both Matter
- Both absolute and relative risk information should be presented together, as relative risk alone can be misleading 3
- Reducing risk from 2 in a million to 1 in a million yields an RR of 0.50, as does reducing risk from 20% to 10%, but the latter has vastly greater public health and personal impact 3
Clinical Application
- When assessing adverse drug effects, the absolute number of additional cases per unit time is often more clinically relevant than the relative risk 1
- For instance, 10 years of hormone replacement therapy results in 5 additional breast cancers per 1,000 users of estrogen-only preparations (RR can be translated to this absolute measure) 1
- A woman with baseline risk of 12/1,000 per year gains twice the absolute benefit of a woman with 6/1,000 per year risk, even with identical relative risk reduction 3
Common Pitfalls and Caveats
Presentation Bias
- Framing benefits in relative rather than absolute terms can bias perception of treatment effectiveness, making benefits appear more favorable than they actually are 3
- This is why guidelines recommend presenting both measures together 3
Time Dependency
- The same relative risk sustained over different time periods produces different absolute risk reductions 3
- An intervention with RR of 0.50 reduces absolute risk by 1.5% over 5 years but 3% over 10 years for someone with 6/1,000 per year baseline risk 3
Comparison with Odds Ratio
- Risk ratio differs from odds ratio, though they approximate each other when outcomes are rare (<10%) 2, 4
- When outcomes are common (≥10%), the odds ratio will be further from 1.0 than the risk ratio and can exaggerate the effect 5, 4
- Risk ratios are more intuitively understood and consistent with general intuition compared to odds ratios 5, 6
Statistical Properties
- Risk ratios have a mathematical property called collapsibility, meaning the RR size won't change if adjustment is made for a non-confounding variable 5
- Widely used statistical models like logistic and Cox proportional hazards regression are based on ratio measures 1
- Confidence intervals should always accompany risk ratios to indicate precision and uncertainty of the estimate 1, 7