Allocation of COVID-19 Vaccines Under Limited Supply

This paper considers how to allocate Covid-19 vaccines to different age groups when limited vaccines are available over time. The disease dynamics is specified by an age-structured SAPHIRE model whose parameters are estimated by the standard least square method using the epidemic data from New York City. We derive optimal static allocation policies with different objectives under different amounts of daily available vaccines, and examine several dynamic allocation heuristics including old-first policy, infection-first policy, myopic policy, death-weighted myopic policy, and two-day myopic policy. For static policies, our numerical study shows that to minimize the total deaths, it is optimal to allocate limited vaccines to the oldest group first and then the younger group if there are capacities remaining. In contrast, to minimize the total confirmed cases, the optimal static policy allocates a considerable portion of vaccines to younger groups even if the daily available vaccines are very limited. The optimal static policies achieve a much smaller number of confirmed cases and deaths compared to two benchmark policies: a uniform allocation policy that allocates available vaccines equally to each age group, and a proportional allocation policy that allocates available vaccines proportionally to the population of each age group. For dynamic allocation policies, the myopic policy and the two-day myopic policy have similar performance and significantly outperforms the other dynamic heuristics and the static policies in terms of the confirmed cases and deaths.