ABSTRACT
Early in the pandemic, in the North American epicenter, we investigated associations between sociodemographic factors and rates of pediatric COVID-19 diagnoses in a non-clinical setting and whether symptoms varied by child age. From 20 April-31 August 2020, COVID-19-related data were collected on 2694 children aged ≤ 18 years living in households participating in the New York University Children's Health and Environment Study. We examined differences in rates of subjective and objective diagnoses according to sociodemographic characteristics and differences in reported symptoms by child age. Children of women who were non-Hispanic White, had private health insurance, higher income, or more education were more likely to be diagnosed via WHO criteria or healthcare provider. Children of women who were Hispanic or Asian, reported low income, had less education, or were/lived with an essential worker were more likely to test positive. Older children were less likely to experience cough or runny nose and more likely to experience muscle/body aches, sore throat, headache, and loss of smell or taste than younger children. In conclusion, relying on subjective disease ascertainment methods, especially in the early stage of an outbreak when testing is not universally available, may misrepresent the true prevalence of disease among sociodemographic subgroups. Variations in symptoms by child age should be considered when determining diagnostic criteria.