In April 2021, a COVID-19 outbreak occurred at a correctional facility in rural Virginia, USA. Eighty-four infections were identified among 854 incarcerated persons by facilitywide testing with reverse transcription quantitative PCR (qRT-PCR). We used whole-genome sequencing to link all infections to 2 employees infected with the B.1.1.7α (UK) variant. The relative risk comparing unvaccinated to fully vaccinated persons (mRNA-1273 [Moderna, https:// www.moderna.com]) was 7.8 (95% CI 4.8–12.7), corresponding to a vaccine effectiveness of 87.1% (95% CI 79.0%–92.1%). Average qRT-PCR cycle threshold values were lower, suggesting higher viral loads, among unvaccinated infected than vaccinated cases for the N, E, and S genes. Vaccination was highly effective at preventing SARS-CoV-2 infection in this high-risk setting. This approach can be applied to similar settings to estimate vaccine effectiveness as variants emerge to guide public health strategies during the ongoing pandemic.