Recently, two Doctors operating a free clinic called Accelerated Urgent Care in Bakersfield, CA stepped forward to criticize the mortality rate of Covid-19, saying the numbers indicated that it was no worse than the annual flu. Their video was critiqued by the experts, universally lampooned, and then YouTube removed the video for violation of policy, namely dissemination of false or misleading information. These two were lionized by Dr. Phil McGraw (which should be a red flag for everyone, right there), and conservative media, of course. But were they right? I looked for a table of their data, but couldn't find it. That's because they didn't provide one. So instead I found those few remaining outlets which still offered the video and watched the whole thing through, taking notes. If you want to follow along with me, one of the websites which still shows the video can be found here. Here's what I found:
They're idiots.
Erickson does most of the talking during the press conference which the video shows. He begins by saying that it is right to quarantine the sick, but not right to quarantine the healthy. This is empirically false. Quarantining can both keep a disease in, OR keep a disease out! It is simply, plainly false to say that quarantining should only be done one way to keep a disease in! Prevention is done on both sides!
And this guy is a doctor?
But Erickson was only getting started. He then begins to quote numbers, and I'm pretty good with numbers, so here's where I started writing things down furiously. He points out that 5,213 people in his clinic had tested positive for Covid-19. Of those, 340 tested positive, for a ratio of 6.5%. No problems with the math, there.
He then goes on to point out that there were 33,865 confirmed Covid-19 cases in California, out of 280,900 people tested. That's 12% positive. Still no math errors yet.
It is here that Erickson takes a wild leap. He points out that there are 39.5 million people in California, and extrapolates, based on his 12% positive test rate cited, that there are 4.7 million cases of Covid-19 in California.
That's positively DAFT! With the virus currently spreading, and the population nowhere near reaching herd immunity, in California or anywhere else, it is ridiculous to assume the virus had already spread to 12% of everyone in California at that time! Yes, it's fair to say that many people who did not have symptoms hadn't been tested, and it's also fair to say that because of limited testing, it's impossible to say how far exactly the virus had spread, but calculating 12% of the entire state presumes that the virus spread uniformly, instead of breaking out in "hot spots."
Had Erickson been referring to tests of the antibodies which react to Covid-19, he might have had an argument. Then, he could have said, "X number of people have been confirmed to have the virus already, with Y number of people showing symptoms and Z having died." From there, he could have extrapolated based on population. But he does not do this. He bases his numbers based on those who have been tested so far. With testing limited, that translates almost entirely to people showing symptoms of an illness. It still wouldn't have been 100% accurate, because extrapolating never is, but it would have been closer. Instead, he only takes numbers available from people who had the virus right then and there, and those numbers were provided by the hospitals or clinics which did the testing.
That's a little bit like calculating the average height of people based on measurements taken from players in the NBA.
Then Erickson really goes off the deep end. He takes the extrapolated number, 4.7 million cases of Covid-19, and calculates mortality based on the verified number of Covid-19 deaths, 1,227. This gives him a mortality rate of 0.03% (actually 0.026106). This is ridiculous! He extrapolated the denominator, but not the numerator! He extrapolates, because testing is limited and many people show no symptoms, that 12% of all of California is infected. But by that same logic, he must also factor in that many people who died from Covid-19 were never recorded. Now, that's difficult to fathom, but if one uses the same extrapolation technique Erickson applied, then out of 4.7 million cases in California, we should expect a mortality figure of 170,290. That's a lot more than 1,227!
Let's re-do Erickson's calculations using today's numbers to illustrate what went wrong. Erickson did his calculations and announced them back on April 22. Today's date is May 5th. During that two-week span of time, the number of confirmed cases in California has increased to 54,937, and the number of deaths has nearly doubled to 22,254. The number of people tested is spotty, but the L.A. times reports it as 747,874 people tested so far. That's 7.3% positive. Extrapolate from this against the overall population of California, like Erickson did, and we get 2.9 million Covid-19 cases. 22,254 dead out of 2.9 million is 0.77%! That's still pretty small, but a whole lot bigger than 0.03!
See the problem?
Erickson's tests at his own clinic showed a positive test percentage of 6.5%, 340 positives out of 5,215 people. Suppose he had presumed his own clinic's numbers to be more accurate than California's numbers and used those? These would have given him a mortality rate of 0.05%. Still pretty low, but bigger than 0.05.
What Erickson did was deliberately go with the calculation which would give him the largest possible denominator, while ignoring and/or minimizing the numerator! There's only one way to describe that: politics got put before science!
In the video, Erickson goes on to do the exact same mistake with New York's numbers. Here, one of the reporters couldn't keep quiet any more, and pointed out that such extrapolations probably wouldn't be accurate. Erickson's reaction is to brush off the reporter's concerns and keep right on going. He does the same deliberate miscalculation on New York's numbers, getting an absurdly low percentage (garbage in, garbage out), and then does the same thing with the numbers nationally.
Even the reporters knew more about math than this guy!
One does not need to be a math genius to see the errors, here. One would expect medical doctors wouldn't make a math error this basic. But really, many medical professionals don't bother much after trigonometry, and forget about certain things in favor of learning the names of various medications, memorizing anatomical nomenclature, learning about protein folding and polysaccharide chains... It is literally drinking from a fire hose. Some basic math is bound to fall by the wayside.
When we start getting some testing numbers in for the antibodies relating to Covid-19, and get some measure of how many people have truly been infected with this disease already, we can revisit Erickson's calculations and get a more accurate figure.
In the meantime, I wouldn't visit their clinic in Bakersfield if my life depended on it!
Eric
*
No comments:
Post a Comment