Back

How many COVID-19 cases could have been prevented in the US if its interventions were as effective as those in China and South Korea?

Liu, K.; Song, Y.; Li, M.; Shen, Z.; Wang, M.; Wu, J.

2020-05-11 epidemiology
10.1101/2020.05.06.20092981 medRxiv
Show abstract

The COVID-19 [1] pandemic has forced governments to take measures to contain the spread of the disease [2]; however, the effects have varied significantly from one country to another contingent on governments responses. Countries that have flattened their coronavirus curves prove that interventions can bring COVID-19 under control. These achievements hold lessons, such as the strict social distancing and coordinated efforts of all government levels in China and massive testing in South Korea, for other countries battling the coronavirus around the world. In this work, we attempt to estimate how many COVID-19 cases could have been prevented in the United States (US) when compared with the USs actual number of cases assuming that on a certain date, the US took China-like or South Korea-like interventions and that these interventions would have been as effective in the US as in China and South Korea. We found that if that date was at the early stage of the outbreak (March 10), more than 99% (1.15 million) fewer infected cases could be expected by the end of the epidemic. This number decreases to 66.03% and 73.06% fewer infected cases with the China-like scenario and the South Korea-like scenario, respectively, if actions were taken on April 1, highlighting the need to respond quickly and effectively to fight the virus. Furthermore, we found that although interventions in both China and South Korea allowed the COVID-19 outbreak to be managed, the epidemic could still oscillate without strict large-scale lockdown measures, as shown in South Korea. Our results demonstrate that early effective interventions can save considerably more people from infection and provide a worldwide alert regard the need for swift response.

Matching journals

The top 10 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.