Investigating the Impact of COVID-19 Vaccination on Hospitalization Risk, Variant Emergence, and Mortality Inequities

Abstract

COVID-19 is one of the most significant infectious disease threats in modern history, resulting in millions of hospitalizations and deaths globally. Vaccinations against COVID-19 first became available in late 2020 and are estimated to have averted nearly 20 million deaths in the first year alone.1 Yet, the consistent emergence of new COVID-19 variants coupled with waning of both vaccine- and infection-induced immunity has led to persistent morbidity and mortality even among vaccinated populations. Despite extensive research to support the effectiveness of vaccines for preventing infection, hospitalization, and death, less is understood about the role of vaccination in more complex circumstances. This dissertation addresses research gaps in three such nuanced situations. In Manuscript 1, I use generalized additive models to estimate hospitalization risk over time among COVID-19 vaccinated patients in a large Minnesota health system. I then examine differences in hospitalization risk trajectories across six COVID-19 risk factors: number of comorbidities, age, race/ethnicity, gender, prior COVID-19 infection, and community disease incidence. And finally, to contextualize the results I compare absolute risk of hospitalization to the threshold used to define a severe influenza season. I find that in the first year and a half after vaccination there are two peaks in hospitalization risk occurring at 14- and 33-weeks post-vaccination. I also find markedly higher risks of hospitalization for people with comorbidities, and evidence that the risk trajectory among people with one comorbidity or with two or more comorbidities differs from the risk trajectory for people with no comorbidities. Similarly, I find that older age groups have increased hospitalization risk and potential differences in hospitalization risk trajectories by age group. Finally, I find that the risk of hospitalization within certain vaccinated groups quickly exceeds thresholds for a severe influenza season. In Manuscript 2 I compare differences in risk of infection from newly emerging COVID-19 variants between unvaccinated, monovalent vaccinated, and bivalent vaccinated COVID-19 cases. I find that both monovalent and bivalent vaccinated cases, were less likely than the unvaccinated population to be infected with the BA.4/5 variant than other variants circulating at the time. I also find that the bivalent vaccinated population was more likely than the unvaccinated population to be infected with the emerging BQ or XBB variants than the BA.4/5 variant. These results offer two key insights for healthcare and public health practitioners. The first, is that while recently vaccinated people have lower risk of infection and hospitalization overall, they may experience disproportionate risk as new variants emerge. And second, that vaccinations exert selection pressure that may facilitate the emergence of new variants and more work is needed to understand the role that vaccines play in SARS-CoV-2 evolution. In Manuscript 3, I examine how differences in vaccination by race/ethnicity affected mortality disparities in Minnesota in the first year and a half of vaccine availability. Using a decomposition approach, I find that more favorable vaccination profiles in the BIPOC population reduced mortality disparities, but this reduction was not enough to outweigh disparities from other sources. In addition, I found that the bulk of mortality disparities were driven by differences within the unvaccinated populations and within the vaccinated population the mortality disparities were much smaller and at times reversed. These findings indicate that vaccinations played an important role in reducing mortality disparities in Minnesota and that a substantial portion of the observed mortality disparities were driven by differences between BIPOC and White unvaccinated populations rather than by differences is maintaining booster doses. Although additional work is needed to extend these findings to future populations, efforts to improve initial vaccine uptake are likely to have an outsized benefit for reducing COVID-19 mortality disparities by race/ethnicity.