How To Interpret That ‘1 in 8 Recovered COVID Patients Die’ Study

The study, which has not yet been peer-reviewed, has troubling implications. But headlines focusing solely on deaths miss the broader point about "long COVID."

  • Published 21 January 2021
Image via David Cliff/NurPhoto via Getty Images
As governments fight the COVID-19 pandemic, Snopes is fighting an “infodemic” of rumors and misinformation, and you can help. Read our coronavirus fact checks. Submit any questionable rumors and “advice” you encounter. Become a Founding Member to help us hire more fact-checkers. And, please, follow the CDC or WHO for guidance on protecting your community from the disease.

On Jan. 18, 2021, several media outlets covered a preliminary draft of a large-scale study investigating the long-term outcome of COVID-19 patients in the U.K. who were hospitalized, recovered, and released. While still in pre-print form and therefore not yet peer-reviewed, it is a potentially significant contribution to COVID-19 literature, using a massive trove of data and conducted by researchers from the U.K.’s Leicester University and the Office for National Statistics.

Coverage of the study primarily focused on death. “One in eight recovered COVID patients die from illness complications within 5 months,” wrote the New York Post in its headline. This headline is somewhat misleading, as it doesn’t sufficiently define what it means by “recovered COVID patients.” In the case of this study, it refers to patients with severe enough cases of COVID to warrant hospitalization. If you contracted a mild case of COVID-19 and beat it, that does not mean you have a 1-in-8 chance of dying in the next five months.

The broader message of the work, though, is that what is increasingly being referred to as “post-COVID syndrome” (PCS) or “long COVID” — an enigmatic series of medical complications that can persist indefinitely, span multiple organ systems, and present itself as entirely different or new medical emergencies — warrants more attention and public health planning. “Our findings suggest that the long term burden of COVID-related morbidity on hospitals and broader healthcare systems is likely to be substantial,” the researchers concluded.

Below we explain these findings in detail and explain how the study was designed and conducted.

What Did the Study Find?

Using data collected by the U.K.’s National Health Service (NHS), the researchers first looked at a pool of 47,780 individuals who were hospitalized with COVID-19 and released between January and August 2020. Within that pool of cases, 29.4% were readmitted to the hospital and 12.3% (i.e., nearly 1 in 8) died within a 150-day period following their discharge.

Knowing these raw numbers tells only part of the story. What is more informative is how much of a risk these COVID cases present compared to people hospitalized for other causes. By comparing their pool of COVID cases to a pool of non-COVID related hospitalizations that occurred at the same time (more on this below), researchers found that the COVID cases presented 3.5 times a higher risk of readmission and 7.7 times higher risk of death compared to non-COVID hospitalizations.

The researchers then analyzed the medical conditions driving these post-COVID hospitalizations or deaths. In addition to high rates of respiratory illness, the study found that rates of multi-organ dysfunction were much higher in the COVID group compared to the control group — significantly diabetes and “major adverse cardiovascular events” (MACE). 

How Was the Study Conducted?

This study is what’s referred to as a retrospective matched cohort study. Such a study is retrospective in that it relies on preexisting data collected not specifically for the research question being addressed. In this case, their data — medical records of people hospitalized with and without COVID — came from NHS’ Hospital Episode Statistics database and a collection of records from the NHS’ General Practice Extraction Service created for studying COVID known as the Data for Pandemic Planning and Research database. Death records came from the U.K.’s National Office of Statistics.

A matched cohort study is a way to compare a population of people exposed to a certain variable (in this case COVID-19) to a population that is as similar as possible to the exposed population with the exception of that one variable. In this case, every one of the 47,780 individual COVID cases was paired with a case (found within those massive collections of records) from the same general time period with the same “demographic and clinical characteristics.”

Such a design reduces the possibility of confounding variables, making it easier to see, from a statistical standpoint, a meaningful comparison between COVID hospitalizations and non-COVID hospitalizations. This group of non-COVID cases is the control group on which the comparative risk of readmission and/or death is based.

Why Is the Study Important?

As science writer Ed Yong wrote in The Atlantic in August 2020, there has been a significant delay in recognizing so-called “long-hauler” cases of COVID:

Our understanding of COVID-19 has accreted around the idea that it kills a few and is “mild” for the rest. That caricature was sketched before the new coronavirus even had a name; instead of shifting in the light of fresh data, it calcified. It affected the questions scientists sought to ask, the stories journalists sought to tell, and the patients doctors sought to treat. It excluded long-haulers from help and answers.

This study, should its findings hold up under peer-review, would be one of the largest conducted that quantifies the risk of re-hospitalization or death following COVID hospitalization. It provides initial and sobering evidence about the danger —  and reality —  unique to COVID-19 long-haulers and PCS. “There’s been so much talk about all these people dying from Covid … but death is not the only outcome that matters,” Charlotte Summers, a lecturer in intensive care medicine at Cambridge who was not involved in the research, told The Guardian. “The idea that we have that level of increased risk in people – particularly young people – it means we’ve got a lot of work to do.”

“Urgent research is required to further understand the risk factors for PCS,” the authors of the study concluded, “so that treatment provision can be better targeted to demographically and clinically at-risk populations.”