Authors to correct influential Imperial College COVID-19 report after learning it cited a withdrawn preprint

A March paper by researchers at Imperial College London that, in the words of the Washington Post, “helped upend U.S. and U.K. coronavirus strategies,” cited a preprint that had been withdrawn.

Retraction Watch became aware of the issue after being contacted by a PubPeer commenter who had noted the withdrawal earlier this month. Following questions from Retraction Watch this weekend, the authors said they plan to submit a correction.

In March, the New York Times wrote:

The report, which warned that an uncontrolled spread of the disease could cause as many as 510,000 deaths in Britain, triggered a sudden shift in the government’s comparatively relaxed response to the virus.

American officials said the report, which projected up to 2.2 million deaths in the United States from such a spread, also influenced the White House to strengthen its measures to isolate members of the public.

Table 1 of the original Imperial College report, posted March 16, and table 3 of a resulting paper in Lancet Infectious Diseases, published on March 30, cite a preprint from researchers in China posted on February 10. 

But on February 21, apparently unbeknownst to the Imperial College researchers, the authors of the preprint withdrew it, replacing the PDF with this text:

Our manuscript was based on surveillance cases of COVID-19 identified before January 26, 2020. As of February 20, 2020, the total number of confirmed cases in mainland China has reached 18 times of the number in our manuscript. While the methods and the main conclusions in our original analyses remain solid, we decided to withdraw this preprint for the time being, and will replace it with a more up-to-date version shortly. Should you have any comments or suggestions, please feel free to contact the corresponding author.

‘We are confident that the results still hold’

On Sunday, Azra Ghani, co-corresponding author of the Lancet Infectious Diseases paper, thanked Retraction Watch for bringing the withdrawal to her attention, and said she and her colleagues would ask the journal for a correction but that they are “confident that the results would still hold.”

I have looked back and we did use the data on cases by age in our analysis from this preprint. We were aware on the grapevine that they had not sought the permission from Chinese authorities at the time (I can see an email about that in our records just after the original post) but I am afraid I did not check back at the time that we submitted our paper to LID (early March) so I was not aware that it had been withdrawn.

As the authors note “While the methods and the main conclusions in our original analyses remain solid, we decided to withdraw this preprint for the time being, and will replace it with a more up-to-date version shortly.” This appears to agree with what we heard about this being an issue with permissions. We were in contact with other researchers with access to those data in early February who shared very similar distributions so we had no reason to doubt it’s validity.

We will write to issue a correction to the Lancet ID to note that this citation was withdrawn. If the data were incorrect then this would influence our conclusion about the age-dependent pattern (but not the overall proportion) of infections that would be hospitalised. However, the epidemic has moved on and there are now many more datasets showing a very similar pattern to that estimated in our paper (for example https://www.medrxiv.org/content/10.1101/2020.04.20.20072413v1). Therefore we are confident that the findings still hold.

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.