The effect of outcome choice on vaccine dose allocation: a modelling study

In the early stages of a pandemic, vaccine rollout strategies are generally based on reducing the burden on healthcare services and/or deaths. Quality-adjusted life-years are often not taken into account. With COVID-19 as an example, we used mathematical modelling to investigate how the optimal strategy for allocating vaccines is affected when life-years are the focus rather than deaths, specifically whether single-dose or two-dose prioritisation is more effective.

We extended a previously published model to estimate the number of life-years saved and deaths averted following a one-dose or two-dose COVID-19 vaccination schedule in England at the beginning of the vaccine rollout. We included updated estimates for the risk of mortality due to COVID-19 and vaccine uptake levels for different population groups.

The vaccine dose allocation strategy considered optimal changed depending on the outcome of interest. For lower numbers of available doses (between three and ten million), it was more beneficial to prioritise the one-dose strategy, regardless of the outcome. As the number of available doses increased above ten million, the two-dose strategy became optimal at a point determined by the vaccine effectiveness and the chosen outcome measure.

Rather than focussing on deaths, which is the approach taken in many existing COVID-19 models, targeting life-years can lead to different vaccination policies being considered optimal. We argue that - alongside infections, hospitalisations and deaths - life-years lost should be regarded as a primary measure of disease impact in the development of future vaccine programmes.