RR / OR / NNT Risk-Measure Calculator
Enter a 2×2 table to compute relative risk (RR), odds ratio (OR), absolute risk difference (ARD), relative risk reduction (RRR) and number needed to treat (NNT), with 95% confidence intervals for the ratios and the risk difference. Common for interpreting cohort, case-control and RCT outcomes. Computed locally; data never uploaded.
① Enter the 2×2 table
Rows = groups (exposed/treatment vs control), columns = outcome (event vs no event). Enter counts in each cell.
| Event + | No event − | |
|---|---|---|
| Group 1 (exposed/treatment) | ||
| Group 2 (control) |
How to use & methodology
RR vs OR — which should I use?
RR (relative risk) is the ratio of two groups' incidence risks, intuitive, for cohort studies and RCTs. OR (odds ratio) is the ratio of odds, for case-control studies and logistic regression. When the outcome is uncommon (<10%) they are close; when common, OR diverges further from 1 and must not be read as RR.
What are ARD, RRR and NNT?
ARD is the absolute difference in risk between groups; RRR is the relative risk reduction (1−RR); NNT is the number to treat to avoid one extra outcome = 1/|ARD|. ARD and NNT reflect absolute benefit and are closer to clinical decision-making than RR/OR alone.
What if a cell is 0?
A zero cell makes OR/RR become 0 or divide by zero. This tool automatically applies the Haldane–Anscombe correction (+0.5 to each cell) before estimating and flags it. This is standard practice, but a zero cell usually means a small sample, so conclusions need caution.
What does a CI crossing 1 (or risk difference crossing 0) mean?
A 95% CI for RR/OR crossing 1, or the risk-difference CI crossing 0, indicates no statistically significant difference at the 0.05 level. In that case the NNT confidence interval extends to infinity, which the tool flags as "CI includes ∞".