Cohort Studies Practice Test

The method being tested is a sensitivity measure that tells you how strong an unmeasured confounder would have to be to explain away the observed association. The E-value does this by providing the minimum strength of association that an unknown confounder would need to have with both the exposure and the outcome to nullify the observed effect, under the usual study assumptions. It helps you gauge the robustness of findings in observational studies without having measured that confounder. This is not what a Kaplan-Meier curve does—that curve shows survival probabilities over time and doesn’t quantify how much unmeasured confounding could bias results. It’s also not the hazard ratio, which is a measure of effect size between groups but doesn’t itself assess potential unmeasured confounding. And a Mann-Whitney test compares distributions between groups without addressing confounding at all. So the E-value is the tool that explicitly addresses how unmeasured confounding could influence the observed association. For intuition, if you observe a risk ratio of, say, 2.0, the E-value would be 2.0 + sqrt(2.0*(2.0-1)) ≈ 3.41. That means an unmeasured confounder would need to have a fairly strong association (risk ratio around 3.4) with both the exposure and the outcome to fully explain away the observed association. The larger the E-value, the more robust the finding is to potential unmeasured confounding.