Research Workbench

Risk-ratio E-value sensitivity

Quantify a risk-ratio-scale unmeasured-confounding benchmark.

Your entries stay in this page’s memory and are not submitted to us. Downloads are generated locally. Reloading or leaving the page clears your entries. Do not enter patient information.

Risk ratio, confidence limit closest to the null

Supply a positive risk ratio and the confidence limit closest to one. For RR>1 use the lower limit; for RR<1 use the upper limit. A limit crossing one gives an interval E-value of one.

The starting values are hypothetical. Results update as you edit. Use aggregate or synthetic inputs only.

Calculation results

Point-estimate E-value: 3.4142136
Closest-limit E-value: 1.9244998
Scale reminder: Risk ratios only; not a generic odds-ratio or hazard-ratio conversion.

How to interpret the output

For an observed RR above one, E=RR+sqrt(RR×(RR−1)); protective ratios are inverted. The benchmark concerns joint confounder–exposure and confounder–outcome risk-ratio associations conditional on measured covariates. It does not address selection bias, measurement error or prove causality. Do not directly enter a common-outcome odds ratio.

Method reference

Source checked 2026-11-06. Read the reproducible calculator validation report.