Two physician-entrepreneurs who own a string of urgent care clinics in Bakersfield claim the data they have collected from testing for coronavirus proves that the current “stay-at-home” and physical distancing orders need to end and that COVID-19 is “just like the flu.” Their YouTube video interview from last week has been viewed more than five million times. It has garnered attention from the likes of Elon Musk who tweeted “Docs make good points” and Fox News host Laura Ingraham who played excerpts Monday night. But their claims have sparked wide controversy from experts in academia who say their assumptions are flawed, their sample biased, and their extrapolations “implausible.” The Kern County health office has stated it does not agree with their claims. And yesterday their original video was removed for what YouTube says are violations of its “community guidelines.”
The doctors claim their sample of 5,213 tests for coronavirus, in which they found 340 positives (6.5%), can be extrapolated to not only Kern County (population 700,000) but to the entire state of California (population 39.5 million). They claim that this shows that the coronavirus is more widely spread than previously thought, is less virulent, and is just like the flu. But experts like Carl Bergstrom, an evolutionary biologist and expert in modeling biological systems, say the physicians’ statements and assertions are flawed. Writing on Twitter, Dr. Bergstrom noted that estimating the proportion of infected patients from an urgent care clinic sample is a bit like “estimating the average height of Americans from the players on an NBA court.”
The doctors, using what they say is data from Norway and Sweden additionally claimed there is “no statistical difference” in deaths in the two countries, despite the differences in how Sweden and Norway have approached the pandemic. Norway has closed schools and businesses and mandated social distancing while Sweden has chosen to keep its businesses and society open. The data would suggest otherwise. Google searches for “Sweden COVID-19 cases” and “Norway COVID-19 cases” yielded the number of reported COVID-19 cases for each country along with the number of reported deaths as of April 28. The rate of deaths per capita in each country might shed some light on the doctors’ claim of no differences. For Norway and Sweden the per capita mortality rate for COVID-19 are 0.1% and 0.2% respectively. Epidemiologists estimate the severity of an outbreak by calculating the Case Fatality Rate (CFR) – dividing the number of deaths by the number of confirmed cases. Using the data from the Google search, Norway’s CFR is 0.3% (205 deaths / 7,599 confirmed cases) and Sweden’s is 1.2% (2,355 deaths / 19,621 confirmed cases) - four times higher. When statistical tests are used to compare these last two rates, they are very significantly different.
Beyond the implausibility of their calculations and extrapolations, the Bakersfield doctors make statements about statistical modeling that are false. They claim in one video that because the number of deaths and cases is lower than the models originally predicted, the models must have been wrong. This is a logical error in that models only model based on current conditions and the assumptions built in. Good modelers are aware of their assumptions, build them into the model and then when predicting biological processes that can change over time, adjust their models using new data. And this is exactly what the most commonly quoted models have been doing. The doctors in Bakersfield criticize modeling as divorced from reality because they treat patients every day thus they know what is “really” going on. (They also criticize Dr. Anthony Fauci saying that he has not seen a patient in twenty years.) It may be that they know some things the modelers do not, but since they only know who they see in their Bakersfield clinics, their observations may be biased and are small in number.
Finally, on a number of points in the video they assert that COVID-19 is “just like the flu.” They assert that because we don’t shut down the society because of the flu, ergo we should not be doing it for the coronavirus. But coronavirus is not the flu. The flu may kill upwards of 40-50,000 individuals in the U.S. annually, but that is over an eight month flu season. The COVID-19 pandemic has killed over 55,000 people in the U.S. in just over two months. The mortality rates for the flu have been suppressed by a combination of antiviral treatments, vaccinations, and the build-up of immunity in the community. Additionally, there is no evidence (like there is for the SARS-COV-2 virus) that the flu is spread by infected but asymptomatic individuals. And although experts believe that we may be “over the hump” in this wave of coronavirus deaths, we can still expect thousands more deaths before the fall. And some expect a second wave will hit this fall (on top of the beginning the next flu season), before we have been able to develop effective treatments and possibly a vaccine that will provide some increased level of immunity from infection.
So, in sum, the doctors in Bakersfield make irrational arguments and use their limited and biased sample to make arguments that should not be believed.