Infectious Disease
How should we select test-negative controls? A causal perspective in the era of multiplex respiratory testing Christopher Boyer* Christopher Boyer Cleveland Clinic Lerner College of Medicine of Case Western Reserve University
The test-negative design (TND) is widely used to estimate vaccine effectiveness (VE) for respiratory pathogens by comparing vaccination odds among symptomatic individuals who test positive (cases) versus those who test negative (controls). A central—and increasingly consequential—design choice is which test-negative illnesses constitute valid controls. With rapid multiplex PCR panels, investigators can now observe a rich spectrum of non-focal pathogens among test-negative patients, creating both opportunity (better characterization of control illness) and risk (including invalid controls that amplify bias). Building on recent work clarifying identifying assumptions under which the standard TND odds ratio recovers a causal estimand, we show how multiplex testing can inform the control-selection problem: “test-negative illness” becomes a mixture of pathogen-specific illnesses, each of which may violate assumptions differently. We propose a causal taxonomy of test-negative pathogens and a pragmatic selection strategy guided by (i) vaccine irrelevance (exclude pathogens affected by the focal vaccine, including plausible cross-protection), (ii) vaccine-correlation threats (exclude or adjust for pathogens influenced by other vaccines whose uptake is correlated with the focal vaccine), and (iii) testing-process comparability (avoid pathogens whose diagnosis or care-seeking is altered by alternative testing channels, such as at-home rapid tests). We also discuss differences between pathogens that serve as exposure proxies, i.e. proxies for unmeasured confounders of pathogen exposure and vaccination, versus testing proxies, i.e. proxies for unmeasured confounders of receiving a test and vaccination. Throughout, we illustrate how better information about test-negative illness can inform better design of test negative studies.
