From the Guidelines
Odds ratio and relative risk should be used in different contexts, with relative risk being preferred when incidence rates can be directly calculated, and odds ratio being more suitable for case-control studies or logistic regression analyses, as noted in the explanation of strengthening the reporting of observational studies in epidemiology 1. When interpreting research, it is essential to understand the difference between odds ratio and relative risk, as they measure association in epidemiology. Odds ratio is the ratio of the odds of an outcome in an exposed group to the odds in an unexposed group, while relative risk is the ratio of the probability of an outcome in an exposed group to the probability in an unexposed group. Some key points to consider when choosing between odds ratio and relative risk include:
- For rare outcomes (less than 10% incidence), odds ratios and relative risks are numerically similar, as discussed in the context of laboratory medicine best practices systematic review and meta-analysis 1.
- As the outcome becomes more common, odds ratios increasingly overestimate the relative risk.
- Relative risk is preferred when incidence rates can be directly calculated, such as in cohort studies or randomized controlled trials.
- Odds ratios are more appropriate for case-control studies where incidence cannot be directly measured, or in logistic regression analyses, as explained in the explanation of strengthening the reporting of observational studies in epidemiology 1. When interpreting research, be aware that odds ratios tend to exaggerate the effect size compared to relative risk when outcomes are common, which could lead to overestimating treatment effects or risk factors, potentially affecting clinical decision-making and patient counseling about risks and benefits, as highlighted in the discussion of attributable risks and population attributable fraction 1.
From the Research
Odds Ratio vs Relative Risk
The odds ratio (OR) and relative risk (RR) are two commonly used measures of association in medical research.
- The OR is often reported in the medical literature as a measure of association between exposure and outcome, but it is the RR that people more intuitively understand as a measure of association 2.
- The RR can be directly determined in a cohort study by calculating a risk ratio, while in case-control studies, the OR provides a reasonable approximation of the RR when the outcome occurs in less than 10% of the unexposed population 2.
- However, when an outcome is common (i.e., 10% in the unexposed group), the OR will exaggerate the RR 2.
Key Differences
- The OR and RR compare the relative likelihood of an event occurring between two groups, but the RR is easier to interpret and is consistent with general intuition 3.
- Covariate adjustment is easier for an OR, and the OR avoids ambiguity by being invariant to the labeling of the outcome measure 3.
- The choice between OR and RR depends on the study design and the outcome being measured, with simple rules available for when and how to use each statistic 4.
Interpreting Results
- When reading research that summarizes data using ORs or RRs, it is essential to be aware of the limitations of both measures 3, 5.
- Researchers should report both a relative and an absolute measure, presenting these with appropriate confidence intervals to aid in the interpretation of results 5, 6.
- A sound understanding of ORs, RRs, and other measures of association and risk quantification is foundational in the clinician's ability to interpret, synthesize, and apply available evidence from the medical literature 6.