Fig. 4

Impact of including (blue) or excluding (orange) each age group in a vaccination policy, measured as reductions in cumulative cases, hospitalizations, and deaths in Rhode Island by 30 June 2021. With nine age classes in the model, there are 2⁹−1=511 possible age-based vaccination strategies where the vaccine supply is equally distributed among the participating age groups. The vaccination campaign in this figure ends on 4 March 2021 and the total vaccine supply is enough to vaccinate 300,000 people (28.3% coverage). Vaccine efficacy half-life here is 360 with slope = 2. Each violin plot shows the distribution across 256 (blue) or 255 (orange) strategies which include or exclude the corresponding age group. Each strategy was evaluated by taking the median of ten runs with ten different sets of parameters drawn from the inference [18] posterior distributions. Campaigns which cover the 30–39 age group would reduce the median of cumulative cases by 0.67% and cumulative hospitalizations by 0.06% compared to those not covering this age group. The median of cumulative deaths drops by 3.58% when targeting the 80+ age group when compared to strategies that do not include the 80+ age group. Results for Massachusetts shown in Additional file 1: Figure S13