Identification and estimation of vaccine effectiveness in the test-negative design under equi-confounding (2504.20360v3)
Abstract: The test-negative design (TND) is frequently used to evaluate vaccine effectiveness in real-world settings. In a TND study, individuals with similar symptoms who seek care are tested for the disease of interest, and vaccine effectiveness is estimated by comparing the vaccination history of test-positive cases and test-negative controls. Traditional approaches justify the TND by assuming either (a) receiving a test is a perfect proxy for unmeasured health-seeking behavior or (b) vaccination is unconfounded given measured covariates -- both of which may be unrealistic in practice. In this paper, we return to the original motivation for the TND and propose an alternative justification based on the assumption of odds ratio equi-confounding, where unmeasured confounders influence test-positive and test-negative individuals equivalently on the odds ratio scale. We discuss the implications of this assumption for TND design and provide alternative estimators for the marginal risk ratio among the vaccinated under equi-confounding, including estimators based on outcome modeling and inverse probability weighting as well as a semiparametric estimator that is doubly-robust. When the equi-confounding assumption does not hold, we suggest a sensitivity analysis that parameterizes the magnitude of the deviation on the odds ratio scale. We conduct a simulation study to evaluate the empirical performance of our proposed estimators under a wide range of scenarios.