BackgroundThe availability of high-dose (HD) influenza vaccine for seniors should decrease influenza-related hospitalization. Studies to date show a range of mostly moderate increased HD vaccine effectiveness (VE). While a ‘healthy vaccinee’ phenomenon can inflate VE, for influenza and particularly an HD vaccine targeted at frailer adults, an ‘at-risk vaccinee’ bias may deflate VE estimates. We assessed senior HD vaccine effectiveness against influenza-related hospitalization by linking immunization registry records to hospitalizations. We also examined whether adding strata typically ignored in case-control matching schemas, such as residence areas, exact age, and provider biases, would increase VE. MethodsFor the 2016–17 influenza season in the Portland metropolitan area, the differential VE for the HD vaccine in preventing PCR-confirmed influenza hospitalization was assessed by a nested series of models across matching strata. For an exact match for high-dose and standard-dose seniors, matching elements included exact age, gender, residence type, race-ethnicity, provider bias, and residence area (zipcode). ResultsAs a first step, a simple aggregate comparison of influenza-related hospitalization risk showed no added HD effectiveness. For the nested models, adding strata increased VE. In the final model, among 23,712 matched pairs of HD to SD vaccinated seniors, the HD vaccine was 30.7% (95%CI: 8–48%) more effective in preventing influenza-related hospitalization. ConclusionFor this study, the high-dose influenza vaccine provided superior protection for seniors against influenza hospitalization. Including matching elements as exact year of age and residence zipcode all added to the calculation of VE. As a warning, non-matched or overly simple matched VE study designs may substantially under-estimate VE.