Variation over Time in Child and Neighborhood Characteristics Associated with COVID-19.

Introduction: To examine the associations between child and neighborhood characteristics and incidence of COVID-19 infection during the first 19 months of the pandemic. Study Design: We utilized individual electronic health record data and corresponding census tract characteristics for pediatric SARS-CoV-2 cases (age <18 years) from March 23, 2020 to September 30, 2021 with molecular tests resulted at a children's health system in Colorado. We compared associations between individual SARS-CoV-2 cases and census tract SARS-CoV-2 positivity rates over three time periods (TP1: March-September 2020; TP2: October 2020-March 2021; TP3: April-September 2021) using multinomial logistic regression for individual associations and negative binomial regression for census tract associations. Results: We included 7498 pediatric SARS-CoV-2 cases and data from 711 corresponding census tracts. Spanish preferred health care language was associated with SARS-CoV-2 positivity for TP1 (odds ratio [OR] 4.9, 95% confidence interval [CI] 3.7-6.5) and TP2 (OR 2.01, 95% CI 1.6-2.6) compared with TP3. Other non-English preferred health care language was associated with SARS-CoV-2 positivity in TP1 (OR 2.4, 95% CI 1.4-4.2). Increasing percentage internationally born in a census tract was associated with SARS-CoV-2 positivity for TP1 (multivariable incident rate ratio [IRR]=1.040, p<0.0001), TP2 (multivariable IRR=1.028, p<0.0001), and in all TP combined (multivariable IRR=1.024, p<0.0001). Discussion: Our study is notable for the identification of COVID-19 disparities among children in immigrant families and communities, particularly early in the pandemic. Addressing disparities for immigrant communities requires targeted investments in public health infrastructure.