Comparison of survey questions to define long COVID: Implications for prevalence and disparities.

Objective: To understand whether variation in survey-based Long COVID estimates is partially due to how and when survey questions are asked. Methods: We compared Long COVID prevalence using distinct questions within a population-based, longitudinal survey of adults with confirmed SARS-CoV-2 before June 2022 in Michigan. Results: In our sample (n = 3826), 17.0 % reported symptoms for 90+ days at baseline, a median of 4.4 months after COVID-19 onset. A median of 18.4 months after COVID-19 onset, 24.5 % reported ever experiencing Long COVID, 16.9 % reported current Long COVID, and 10.8 % reported diagnosed Long COVID. Among adults without 90-day symptoms at baseline, 17.3 % reported ever Long COVID at follow-up. Relatedly, among adults with 90-day symptoms at baseline, 31.1 % reported they never had Long COVID at follow-up. After adjustment for reinfection, respondents who were Hispanic (vs. White) or lower income (<$75,000 vs. $75,000+) had greater odds of reporting baseline symptoms but never experiencing Long COVID at follow-up. Conversely, Black (vs. White) respondents had greater odds of reporting ever Long COVID at follow-up without baseline symptoms. Conclusion: Surveys should employ several questions to define Long COVID and interpret findings within the context of factors likely contributing to discrepancies, including reinfection, stigma, awareness, and care-seeking behaviors.