Incorporating Equity in Influenza Vaccination Strategies: An Assessment of Neighborhood Disadvantage and Influenza Outcomes

Abstract

The Centers for Disease Control and Prevention (CDC) estimates that influenza vaccinations prevented 9.8 million influenza-related illnesses, 120,000 influenza-related hospitalizations, and 7,900 influenza-related deaths in 2023 – 2024. However, influenza vaccination rates are declining and fall short of the Healthy People 2030 goal of 70%. Improving vaccination rates is important to reduce the influenza burden. In developing new strategies, public health leaders should apply lessons from COVID-19 vaccine allocation to prioritize equity in reducing respiratory infectious diseases and to prevent worsening health disparities. One approach to assessing disparities could use the area deprivation index (ADI), a composite Census block-group score that reflects neighborhood disadvantage. To our knowledge, few studies have evaluated neighborhood disadvantage and individual influenza outcomes in a nationally representative sample. This dissertation uses data from the CDC-funded US Flu Vaccine Effectiveness (Flu VE) Network, the largest network of ambulatory vaccine effectiveness studies, to assess the association between ADI and influenza outcomes and to conduct a cost-effectiveness analysis to inform routine vaccination programs that prioritize equitable strategies to increase vaccination rates. In Chapter 1, we examine associations between area deprivation index and influenza outcomes among 2023 – 2024 US Flu VE study participants. We constructed several modified Poisson regression models with robust standard errors to estimate risk ratios (RRs) for influenza outcomes (test positivity and vaccine uptake) by two ADI categorization types: relative and percentile-based. These exploratory analyses were conducted to estimate RRs across ADI categories, examine the impact of including Network site in the models, and assess the influence of covariates on the results. In Chapter 2, we use a retrospective cohort design to conduct modified Poisson regression analyses adjusted for age, sex, race, ethnicity, smoking exposure, history of a high-risk condition, month of illness onset, and calculate RRs for influenza outcomes by participants’ ADI quartile categories. Vaccine effectiveness by ADI quartile was also estimated. In Chapter 3, we analyze the cost-effectiveness of a hypothetical targeted influenza vaccination strategy to increase coverage among the most disadvantaged, using results from Chapter 2 to estimate equitable implementation. We built a decision tree model to estimate the health and economic outcomes of an intervention strategy over one year for 20 subgroups, based on disadvantage measured by ADI (4 groups, with Quartile 4 as most disadvantaged) and age (5 groups). To assess whether the intervention strategy improved equity, we calculated the incremental cost-effectiveness ratio across a range of inequality-aversion parameters, the Atkinson index, and the proportion of health benefits gained by the most disadvantaged. We found that ADI scores varied substantially across sites, and relative quartile categorization best balanced interpretability with sufficient population sizes. Study-site adjustment was outcome-dependent: site was retained for assessing associations between ADI and vaccine uptake but excluded for test positivity, where it over-adjusted for the socioeconomic signal ADI captures. Participants in the most disadvantaged neighborhoods were the least likely to be vaccinated and the most likely to test positive, though vaccine effectiveness did not differ significantly across quartiles. Lastly, we found the hypothetical targeted intervention was cost-effective (ICER $262/QALY against a $100,000/QALY threshold), and incorporating equity weighting strengthened the case: the equity-weighted ICER fell to $216/QALY, the Atkinson index dropped roughly 15%, and all incremental QALYs accrued to the most disadvantaged quartile. Neighborhood disadvantage is a meaningful and measurable driver of influenza disparities in the Flu VE Network. ADI should be operationalized as relative quartiles with outcome-specific site adjustment, and influenza prevention programs should prioritize the most disadvantaged communities, where targeted vaccination is both cost-effective and equity-improving. Future analyses should incorporate state-level vaccination policy to further reduce geographic residual confounding.