Back

Global projections of potential lives saved from COVID-19 through universal mask use

Gakidou, E.; IHME COVID-19 Forecasting Team,

2020-10-11 public and global health Community evaluation
10.1101/2020.10.08.20209510 medRxiv
Show abstract

BACKGROUNDSocial distancing mandates (SDM) have reduced health impacts from COVID-19 but also resulted in economic downturns that have led many nations to relax SDM. Until deployment of an efficacious and equitable vaccine, intervention options to reduce COVID-19 mortality and minimize restrictive SDM are sought by society. METHODSA susceptible-exposed-infectious-recovered (SEIR) deterministic transmission model was parameterized with data on reported deaths, cases, and select covariates to predict infections and deaths from COVID-19 through March 01, 2021. We explore three scenarios: a "non-adaptive" scenario where neither mask use or SDM adapt to changing conditions, a "reference" where current national levels of mask use are maintained and SDM reintroduced when deaths rise, and an increase in mask use to 95% coverage levels ("universal mask"). We reviewed published studies to set priors on the magnitude of reduction in transmission through increasing mask use. RESULTSMask use was estimated at 59.0% of people globally on October 19, 2020. Universal mask use could avert 733,310 deaths (95% UI 385,981 to 1,107,759) between October 27, 2020 and March 01, 2021, the difference between the predicted 2.95 million deaths (95% UI 2.70 to 3.35) in the reference scenario and 2.22 million deaths (95% UI 2.00 to 2.45) in the universal mask scenario over this time period. CONCLUSIONSThe cumulative toll of the COVID-19 pandemic could be substantially reduced by the universal adoption of masks before the availability of a vaccine. This low-cost, low-barrier policy, whether customary or mandated, has enormous health benefits with presumed marginal economic costs.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.