The impact of COVID-19 varied significantly by deprivation, ethnicity, and policy measures. We analyzed individual-level data on 12,310,485 first SARS-CoV-2 Pillar 2-PCR-confirmed infections, 439,083 hospitalizations, 107,823 deaths, and vaccination records in England from May 2020 to February 2022. Poisson regression models adjusted for demographic and temporal factors showed higher incidence rate ratios (IRRs) for severe outcomes in the most deprived areas compared to the least. We note higher IRRs for severe outcomes for all non-white relative to white ethnicities. The magnitude of IRRs for both deprivation and ethnicities declined from the wild-type to the omicron periods for severe outcomes. For infections, we observed IRRs above one for non-white ethnicities during the wild-type and alpha periods. Vaccination significantly reduced risks across all groups. For severe outcomes, preexisting health inequalities led to large and persistent disparities. For infections, measures must be structured with ethnicity and deprivation in mind early in a pandemic.