You may have seen the headlines: Smoking protects people from the new coronavirus. But this is one case where you shouldn't believe everything you read.
The false narrative put out by a number of news outlets highlights an underlying problem. Scientific studies can be complicated, and it’s easy for non-experts, including reporters, to misread their findings.
The data on smoking and COVID-19 is buried in a wide-ranging report from Oxford University and the London School of Hygiene & Tropical Medicine.
The authors of the May 7 report analyzed the health data of 17.4 million U.K. adults between early February and late April. Their goal was to “discover the key factors associated with death from COVID-19,” according to a London School of Hygiene & Tropical Medicine release.
Some of the risk factors the study highlights are unsurprising. If you’re over the age of 80 or have heart disease of diabetes, your risk of dying from the virus scientists call SARS-CoV-2 is higher than for younger, healthier people. Men are more likely to die than women.
One apparent data point is surprising, however. “Those who had never smoked were twice as likely to die from the virus compared to current cigarette-users,” the Daily Mail tabloid reported. “It is the latest in a growing line of studies to suggest smokers have a lower risk of becoming seriously ill with COVID-19, which researchers describe as ‘weird.’”
A French study from early April also seemed to indicate a relationship between smoking and a higher COVID-19 survival-rate.
But the Daily Mail and other media that have latched onto the “smoking helps” notion are misreading the data from these studies, epidemiologists told The Daily Beast. Smoking will not help you beat SARS-CoV-2, these experts stressed. Countless studies over many decades have firmly proved that smoking doesn’t really have any medical benefits.
Still, the emergence of the pro-smoking narrative isn’t shocking. “This type of misinterpretation is bound to happen when these studies are released to a broader-than-normal audience,” Theresa MacPhail, an author and medical anthropologist, told The Daily Beast.
The U.K. study, one of the biggest of its kind, was a “really heroic effort at bringing together real-time data on a huge number of people,” Eleanor Murray, a Boston University epidemiologist, told The Daily Beast. “But it was not an attempt to identify things that cause or protect against death from COVID.”
“The media has largely misreported the results of this study by interpreting them as cause and effect,” Murray added. “The smoking result is one that has been particularly widely circulated and misinterpreted.”
The basic problem is mistaking correlation for causation. Just because many smokers beat COVID-19 doesn’t mean smoking is the reason they beat it. And that means you shouldn’t take up smoking in order to improve your chances of surviving the pandemic.
One term is partially responsible for the widespread misreading of the U.K. report, Murray explained. “This study was looking for what, in epidemiology, we call ‘risk factors.’ That term probably sounds to a lot of people like something that puts you at higher—or lower—risk for death or disease.”
“But the way we use it in epidemiology there is a specific technical meaning. A risk factor is something that could help us predict what might happen to someone even if it’s not actually a cause,” Murray added.
Again, that’s correlation not causation. “Although current smokers, all things considered, seem to be less likely to die in hospital from COVID than never-smokers, this study doesn’t tell us anything about whether smoking is the causal factor,” Murray stressed.
The bad readings of the U.K. study aren’t an isolated phenomenon. Reporters have also mischaracterized the novel coronavirus’ tendency to mutate. The New York Post wrongly reported that SARS-CoV-2 is mutating to become less dangerous.
Meanwhile, the Los Angeles Times incorrectly claimed that the virus is mutating to become more dangerous. Both examples of inaccurate reporting seemed to hinge on the reporters’ poor understanding of technical jargon in complicated scientific studies.
“Media (with a few exceptions) should be forbidden to write about mutations,” Nathan Grubaugh, a Yale geneticist, tweeted jokingly.
Cause and effect in diseases and their risk factors is an equally treacherous subject for non-specialists. “Most of us aren't that good at understanding statistics in the first place,” MacPhail said. “Forget about things like ‘causal inference.’"
Murray recommended that reporters and casual readers ask themselves several questions before drawing conclusions from scientific studies. The first is: What was the goal of this research?
“If the goal is not to understand cause and effect, then think about whether this study is telling you something new,” Murray said.
Her point is that you already know smoking is bad for you. If you read a COVID-19 study and conclude that smoking is some kind of antiviral elixir, it’s probably because you’re reading the study wrong.
