INTRODUCTION: Understanding of COVID-19-related disparities in the U.S. is largely informed by traditional race/ethnicity categories that mask important social group differences. This analysis utilizes granular information on patients' country of birth and preferred language from a large health system to provide more nuanced insights into health disparities. METHODS: Data from patients seeking care from a large Midwestern health system between January 1, 2019 and July 31, 2021 and COVID-19-related events occurring from March 18, 2020 to July 31, 2021 were used to describe COVID-19 disparities. Statistics were performed between January 1, 2022 and March 15, 2023. Age-adjusted generalized linear models estimated RR across race/ethnicity, country of birth grouping, preferred language, and multiple stratified groups. RESULTS: The majority of the 1,114,895 patients were born in western advanced economies (58.6%). Those who were Hispanic/Latino, were born in Latin America and the Caribbean, and preferred Spanish language had highest RRs of infection and hospitalization. Black-identifying patients born in sub-Saharan African countries had a higher risk of infection than their western advanced economies counterparts. Subanalyses revealed elevated hospitalization and death risk for White-identifying patients from Eastern Europe and Central Asia and Asian-identifying patients from Southeast Asia and the Pacific. All non-English languages had a higher risk of all COVID-19 outcomes, most notably Hmong and languages from Burma/Myanmar. CONCLUSIONS: Stratifications by country of birth grouping and preferred language identified culturally distinct groups whose vulnerability to COVID-19 would have otherwise been masked by traditional racial/ethnic labels. Routine collection of these data is critical for identifying social groups at high risk and for informing linguistically and culturally relevant interventions.