The potential benefits of delaying seasonal flu vaccine selections: a retrospective modeling study

Backgrounds and PurposeAntigenic match between selected vaccine virus and circulating virus crucial to achieve high vaccine effectiveness for seasonal flu. Due to the time-consuming process of producing eggs, vaccine candidate viruses are currently selected 5-6 months ahead of the flu season. New non-egg-based vaccine production technologies have emerged with the potential to improve production efficiency and to revise current vaccine formulation schedules. In this study, we aim to 1) identify the past flu seasons where the opportunity to improve vaccine decision existed if rapid vaccine production were available and to 2) quantify the impact of revising the current vaccine decision schedule, where new vaccine production technologies allow more time for specimen collection prior to vaccine virus selection. MethodsWe extracted the trend in the viral activity of season-predominant strain in three data points: when vaccine decision was made, in between vaccine decision and season starts, and after season starts. Between 2012 and 2020, we first identified the past flu seasons where the season-dominant strains presented increasing activity only after vaccine decisions had already been made in February for the Northern Hemisphere. Using an epidemiological model (SEIR) of season flu in the US, we evaluated the impact of updating vaccine decisions on the epidemic size and the number of flu hospitalizations in the United States. ResultsIn the past flu seasons between 2012 and 2020, the timing when the clades or subclades that predominantly circulated during flu season emerged varied by season. In particular, in 2013/14, season-dominant H3N2 subclade emerged after vaccine decisions were made, contributing to the mismatch between vaccine and circulating virus. If the H3N2 component of the vaccine were updated given the additional viral activity data collected after February, our simulation model showed that the updated vaccine could have averted 5,000-65,000 flu hospitalizations, depending on how much vaccine effectiveness could improve with matching vaccine virus. On the other hand, updating the B/Victoria vaccine component did not yield substantial change in flu burden in the 2019/20 season. ConclusionsWith rapid vaccine production, revising the timeline for vaccine selection can result in substantial epidemiological benefits, particularly at times when additional data help improve the vaccine effectiveness through better match between vaccine and circulating viruses.