Heterogeneity in Racial and Ethnic Disparities in COVID-19 Severity Among Pediatric Inpatients in a National...

The COVID-19 pandemic magnified longstanding racial and ethnic disparities in pediatric health, but it is unclear which populations experienced the largest disparities. Our objective was to determine whether disparities in COVID-19 severity differed with respect to patient factors analyzed as effect modifiers.

Using data from the Premier Healthcare Database, this retrospective cohort study included encounters among inpatients < 19 years old from April 2020 through September 2022 in the USA with a COVID-19 diagnosis. Outcomes of COVID-19 severity were intensive care unit (ICU) admission and ventilator use. Comparisons between Black and White patients and Hispanic and White patients were adjusted for confounders using propensity score weights, reported as risk differences (RDs) in percentage points, and tested for heterogeneity (interaction P) across subgroups of effect modifiers such as complex chronic conditions (CCCs) and insurance status.

Of 8947 pediatric inpatients with primary COVID-19 diagnosis, 3858 were White, 2153 were Black, and 2936 were Hispanic. Among children with a CCC, 14.3% of Black inpatients required a ventilator compared to 9.8% of White inpatients; among children without a CCC, 3.2% of both Black and White inpatients required a ventilator (RDs 4.5 vs 0.0; interaction P = .013). Comparisons of Hispanic and White inpatients showed a similar trend in ventilator use, with larger disparities among inpatients with CCCs and no difference among those without CCCs (RDs 2.7 vs -0.7; interaction P=.031). Among children with government insurance, 25.9% of Black inpatients were admitted to the ICU compared to 20.8% of White inpatients; among children with private insurance, Black and White inpatients had comparable ICU admission rates of 20.0% and 21.4%, respectively (RDs 5.1 vs -1.4; interaction P = .025).

Among hospitalized children, racial and ethnic disparities in COVID-19 severity were largest for those with CCCs or government insurance. These results can help identify target populations for interventions to reduce inequity.