Context: Bioethicists have advocated the use of lottery allocation to allocate scarce health care resources during the COVID-19 pandemic and this was adopted in some settings including in the United States. There is limited evidence on public attitudes, and by extension the political acceptability, of this mode of allocation.

Methods: During the pandemic, we conducted a survey-based experiment to elicit the extent of public agreement with the use of lottery allocation as a mechanism compared with medical triage. Data were collected on 15,554 respondents from 14 economically and culturally diverse countries representing roughly half the world’s population. Respondents were randomly allocated (1:1) to two versions of a hypothetical question concerning ways to ration COVID-19 vaccine doses among nurses in a clinic): (i) an unweighted lottery; (ii) prioritization based on assessment by an independent committee of expert physicians. The main outcome was the level of agreement on a scale from 0 (“strongly disagree”) to 100 (“strongly agree”) and differences were stratified by a range of covariates.

Findings: Average levels of agreement for allocation by medical committee ranged from 52.88 (95%CI 50.59-55.18) in France to 71.27 (69.57-72.98) in India. Agreement with lottery allocation varied much more markedly from 21.75 (19.64-23.85) in France to 61.86 (60.24-63.49) in China. In every country there was a significantly higher levels of agreement with medical triage, but the difference in scores varied widely across countries from 6.7 (4.2-9.23) in China to 32.34 (28.86-35.81) in Spain. There was greater agreement with lotteries among males, those on higher income, those with the very lowest levels of education and those identifying with a right leaning ideology.

Conclusion: Globally, preferences for lottery allocation of scarce medical care varies widely. Successful implementation of lottery allocation in future will require greater engagement from health policy makers to ensure they can garner public support.