Estimates of the impact on COVID-19 deaths of unequal global allocations of vaccines

During 2021, COVID-19 vaccinations were delivered much more rapidly in some countries than in others. Ethical principles would have suggested allocating available vaccines to people by age, irrespective of where they live, because mortality risks from COVID-19 are much higher for older people. The World Health Organization recommended initial allocations of vaccines to countries based on their total population size, in part due to uncertainty about how COVID-19 would affect different countries.

This paper estimates how many people would have died from COVID-19 up to 31 October 2021 if either of these allocation rules had been applied, compared to estimates of actual COVID-19 deaths. The estimates suggest that allocating vaccines by age would have resulted in between 500,000 and 1,500,000 fewer deaths globally (with a best estimate of 1,090,000 fewer deaths), while allocating vaccines between countries based on national population sizes would have reduced total deaths globally by between 450,000 and 2,100,000 (with a best estimate of 1,440,000 fewer deaths).

Most low-and middle-income countries would have seen reductions in deaths, with the greatest absolute numbers in large middle-income countries (especially Bangladesh, India and Indonesia). More deaths would have taken place in many high-income countries, with the greatest absolute numbers in the United States and Turkey, and the greatest percentage changes in Arabian Peninsula countries, Israel and some island states. In most European Union countries, deaths would not have differed much if vaccines were allocated by age, because they would have received more vaccine doses during the early months of 2021 but fewer later in the year.

Although allocation of vaccines by age should intuitively lead to fewest deaths, the estimated deaths would have been even lower if vaccines were allocated based on population size. Allocation by population would have directed disproportionate numbers of vaccines to a set of countries – especially India, Bangladesh and Indonesia – which experienced large outbreaks due to the Delta variant in 2021 after having previously limited infections through “ flattening the curve”.

Sequencing of vaccination by age in national vaccination rollouts is critical to maximizing the numbers of lives saved. The estimated gains from fairer global vaccination allocation would be greater if high-income and upper-middle-income countries did not sequence vaccinations by age cohort, and would be lower if lower-middle-income and low-income countries did not vaccinate most of their elderly before the general population, whether due to policy choices or people not accepting vaccines made available to them or logistical difficulties in vaccine delivery.

The estimates correspond to a reduction of between 8.5% and 10.7% (with a best estimate of 10.4%) of the total estimated actual deaths from COVID-19 between 1 January and 31 October 2021, if vaccines were allocated by age, and between 7.8% and 15.0% (with a best estimate of 13.6%) of total estimated actual deaths, if vaccines were allocated based on national population sizes. These percentages are small, despite the large differences in vaccine deployment between countries, because the mostly high-income countries which vaccinated their populations faster have disproportionately large numbers of elderly people. If a future SARS-CoV-2 variant, or a future pandemic, were to have fatality rates that are similar across age groups, or that are higher for children and young adults, then unequal global allocation of vaccines would have a much more severe effect on overall global mortality than it has for COVID-19 so far.