The relative risk formula calculates the ratio of disease incidence in an exposed group to that in a non-exposed group, quantifying how much a factor increases or decreases risk. It's used in medical studies to assess whether a treatment, behavior, or environmental condition affects health outcomes.
Relative risk is computed by dividing the probability of an event occurring in the exposed group by the probability in the unexposed group. A result of 1 means no difference in risk. Values above 1 indicate increased risk; below 1, reduced risk. This measure is central to observational studies and randomized controlled trials.
For example, in a study of smokers versus non-smokers, if 20 out of 100 smokers develop lung cancer and 5 out of 100 non-smokers do, the relative risk is (20/100) ÷ (5/100) = 4.0. This means smokers have four times the risk of developing lung cancer compared to non-smokers.
The formula uses two probabilities: the incidence rate in the exposed group and the incidence rate in the control group. Incidence is the number of new cases divided by the number of people at risk during a time period. The calculation assumes the groups are comparable except for the exposure being studied.
Mathematically, relative risk (RR) = (a / (a + b)) ÷ (c / (c + d)), where:
This structure appears in a 2×2 contingency table, a standard format in epidemiological analysis. The table organizes data so that comparisons between groups are clear and consistent.
Relative risk is most commonly used in cohort studies, where researchers follow groups over time based on exposure status. It's less appropriate for case-control studies, which instead use odds ratios because they cannot reliably estimate incidence rates.
Common applications include:
A 2021 meta-analysis published in The Lancet found that relative risk estimates were consistently more interpretable than odds ratios for communicating public health risks to non-specialists. The study emphasized that RR values under 1.5 are often considered weak associations, while values above 2.0 suggest strong evidence of causation.
Relative risk is frequently confused with odds ratio. While both measure association, odds ratio is derived from odds (probability divided by one minus probability), whereas relative risk uses actual probabilities. Odds ratios tend to overestimate effect size when outcomes are common, especially in case-control studies.
For example, if a study reports an odds ratio of 3.0 for a drug reducing heart attacks, it doesn't mean the drug reduces risk by 300%. That's a 200% increase in odds, not risk. The relative risk may be closer to 1.5, meaning a 50% increase in risk reduction - a critical difference in clinical interpretation.
A key distinction: relative risk reflects absolute change in risk, not the number of people affected. A relative risk of 2.0 doesn't mean twice as many people got sick-it means each individual in the exposed group has double the chance of getting the disease compared to someone not exposed.
In a landmark 1954 British Doctors Study, researchers tracked 34,439 male physicians over 10 years. Among those who smoked, 106 died of lung cancer. Among non-smokers, 12 died. The exposed group had 23,375 smokers, and the unexposed had 11,064 non-smokers.
Incidence in smokers: 106 ÷ 23,375 = 0.00453
Incidence in non-smokers: 12 ÷ 11,064 = 0.00108
Relative risk = 0.00453 ÷ 0.00108 ≈ 4.2
This result showed smokers had about 4.2 times higher risk of dying from lung cancer than non-smokers. The finding was key in shaping tobacco policy and public health messaging.
| Exposure Status | Developed Outcome | Did Not Develop Outcome | Total |
|---|---|---|---|
| Exposed (Smokers) | 106 | 23,269 | 23,375 |
| Unexposed (Non-smokers) | 12 | 11,052 | 11,064 |
When interpreting relative risk, always consider the baseline risk. A relative risk of 2.0 in a rare disease (e.g., 0.1% baseline risk) results in a small absolute risk increase (0.2%). But in a common condition (e.g., 10% baseline risk), the same relative risk leads to a 10 percentage point rise-far more impactful clinically.
I once reviewed a press release claiming a new supplement reduced colds by "60%." The study reported a relative risk of 0.40. However, the baseline cold rate was 30% per year. The actual reduction was from 30% to 12%, a 18-percentage-point drop-not just a 60% relative decrease. The headline misled readers by omitting context.
Relative risk formula is a foundational tool in health research. It quantifies how exposure alters disease probability. Unlike odds ratios, it directly reflects risk differences, making it easier to communicate findings to patients and policymakers.
Always check the baseline risk when evaluating relative risk. A high relative risk with low baseline incidence may represent minimal real-world impact. Conversely, moderate relative risk with high baseline incidence can signal major public health implications.