Simple yet Sharp Sensitivity Analysis for Any Contrast Under Unmeasured Confounding

TL;DR

This paper generalizes sensitivity analysis bounds for unmeasured confounding to any contrast, demonstrating their sharpness and practical attainability through theoretical proofs and real data examples.

Contribution

It extends previous work to any contrast, proving the bounds are sharp and practically attainable, enhancing the flexibility of sensitivity analysis methods.

Findings

Bounds are arbitrarily sharp and practically attainable.

The method applies to any contrast, not just risk ratio or difference.

Illustrated with real data demonstrating usability.

Abstract

We extend our previous work on sensitivity analysis for the risk ratio and difference contrasts under unmeasured confounding to any contrast. We prove that the bounds produced are still arbitrarily sharp, i.e. practically attainable. We illustrate the usability of the bounds with real data.