Does home COVID-19 testing bias COVID-19 vaccine effectiveness estimates?

COVID-19 home testing became widely available in early 2021. Care seeking for acute respiratory illness (ARI), and COVID-19 vaccine effectiveness (VE) estimates may be affected by home testing. In a test-negative case-control study, between January-May 2022, U.S. Flu VE Network outpatients were asked about their home COVID-19 testing before seeking care for ARI and were laboratory tested for SARS-CoV-2. Associations among home testing, care seeking and COVID-19 VE were assessed using multivariable logistic regression. Among 2,614 enrollees, home COVID-19 testing was significantly associated with current (≤6 months) COVID-19 vaccination (adjusted odds ratio (aOR=1.70; 95%CI=1.19-2.42); cough (aOR=1.69; 1.22-2.34) and having a college degree or higher (aOR=1.67; 1.37-2.03) and negatively associated with a prior positive COVID-19 test. COVID-19 illness was associated with cough (aOR=3.07; 2.04-4.61), contact with a COVID-19 case (aOR=2.41; 1.93-3.0), home testing (aOR=1.87; 1.53-2.29) and negatively associated with a prior positive COVID-19 test. In unadjusted modeling with only the association between current vaccination and COVID-19 infection, the OR was 0.77 (95% CI=0.66, 0.92); adjusting for patient race/ethnicity, age, days from onset to enrollment, prior COVID-19 illness and study site, the aOR was 0.65 (0.55, 0.78); estimated VE=35% (95%CI=22%-45%). Adding home testing to this adjusted model, VE was 37% (95%CI=13%-46%). Among patients who home tested, VE was 31% (13%-46%) versus 43% (24%-57%) among patients who did not. Patients reporting home testing for COVID-19 before seeking outpatient care for ARI differed from patients not using home tests, which could affect estimates of COVID-19 VE in some populations.