I received an email from a friend yesterday. We had been talking back and forth about COVID and they were unconvinced that it was as contagious as it has been made out to be. They wrote, “Figured you might appreciate the first part of this email. I haven’t done my own homework on it, but the editor of this newsletter is someone familiar with big data and statistical measures; making those inferences…a little more accurate; hopefully.“
This is what the body of the article said:
The insights about COVID-19 and its spread published in the last three days have been incredible. The increase in data and transparency is allowing us to better understand this virus.
The University of Oxford just published some excellent research outlining how quickly the team from Oxford’s Evolutionary Ecology and Infectious Disease group believes COVID-19 has spread.
The assumptions used were all reasonable and backed by data already collected about COVID-19 and its basic reproductive rate (how fast it spreads). The team also assumed that COVID-19 first reached the U.K. by mid-January which is a conservative assumption.
The research models determined that COVID-19 has already infected somewhere between 36% and 68% of the U.K. population.
That’s not a typo.
It is highly likely that a third to more than half the population has had or has COVID-19… And most are asymptomatic. They don’t even know they have it.
This data is consistent with research just out of Iceland. Iceland is unique because it has tested nearly 3.4% of its entire population for COVID-19. On percentage basis, this is more than any other country in the world.
It found that half of those who tested positive for COVID-19 are asymptomatic, and the other half display “very moderate cold-like symptoms.” Only 30 have been hospitalized out of the country’s 1,086 confirmed infections, and there are no deaths as of the time of writing.
Additionally, Italy’s National Health Institute released data that confirmed more than 99% of all COVID-19 fatalities were people already suffering from underlying medical conditions. Only 0.8% of the fatalities were healthy adults.
To put that in context, only 86 healthy adults have died from COVID-19 there.
Given the research out of Oxford and the basic reproductive rate of COVID-19, it is not difficult to determine the exponential growth in the spread of COVID-19.
On March 1, I spoke with an infectious disease expert at Johns Hopkins who stated that the actual number of cases was likely 100–200 times larger than what was being reported.
But we don’t see those cases for the same reason as the research out of Iceland… The cases are asymptomatic or too mild to justify testing for it.
And here is why this latest research over the last few days is fantastic news for us all.
Johns Hopkins University reports that there are more than 741,000 confirmed cases. The reality is that there are almost certainly more than 74,100,000 cases.
That may seem scary, but it means that the actual mortality rate would be 0.047% or less. Compare that to the mortality rate from influenza in the 2017–2018 season (0.14%).
Every day, the data is giving us a lot to be optimistic about. The world is quickly building immunity to COVID-19.
Veritas in numeris.
Truth in numbers.
This proved to be an opportunity to politely flex my skeptical muscles, and I replied in kind, trying to try to reframe the above piece in a more realistic light.
Found the source material if you are inclined (attached).
In reviewing the paper, it seems that while the (now famous) Imperial model only looks at the info we know to be true and extrapolate forward – which we both agree is a necessarily inexact science – the Oxford model made an assumption/guess that the fatality rate was significantly less. From there, they try to get the curve to match other models and then see what the numbers would look like if they assume that they virus isn’t so deadly after all.
The article sited above seems a bit misleading, but such is the case with most things I read nowadays, from most/all sources, regardless of political slant/affiliation.
Some examples: “The University of Oxford just published some excellent research outlining how quickly the team from Oxford’s Evolutionary Ecology and Infectious Disease group believes COVID-19 has spread.” Nope – they show how another model, different than the Imperial-model can result in a similar curve and think it needs to be investigated with an abundance of testing that hasn’t yet been available. And LOL at the characterization of the “excellent research”. The only thing that is near-excellent is the outcome that the author is seeking to fit his biases.
And to call this “research” is a bit misleading as well. No new data was obtained. The article should read “Epidemiologists create an alternate model to explain hospitalization rates with more people infected but a lower fatality rate that still swamps existing medical infrastructure.”
“The research models determined that COVID-19 has already infected somewhere between 36% and 68% of the U.K. population.” This is also misleading. Models don’t make claims, researchers do. Inj this instance, the researchers arbitrarily input numbers into a model and it determines something. Nothing more. The researchers found an arbitrary number that creates a graph that seems to look similar to another graph and think, “Shit. That’s interesting, We should actually research this.” Now they are going to work on actual research with Universities of Cambridge and Kent in the UK to learn more. Cool science. I dig it.
Also, these numbers are incredibly inconsistent with the data that we actually have on the ground here in the US (and everywhere else too, I think). In CNY alone, we have over 3000 negative test results compared to approx 225 positive. The Oxford model would see positive findings in between 1000-2000. The virus is unlikely to behaving so differently here than elsewhere. Again, the more data the better, but ratios like we see across the country are what make the Imperial model more compelling, I think.
“On March 1, I spoke with an infectious disease expert at Johns Hopkins who stated that the actual number of cases was likely 100–200 times larger than what was being reported.” Fuck, an appeal to authority. I know that people appreciate this kinda thinking (eg climate science), but one should appreciate that if someone is often thinking that the outlier’s opinion is the right one, odds are that they are going to be wrong more than right. Consensus in science is important. So too are the outliers, those folks who don’t tow the line … but they shouldn’t be consistently be mistaken as authorities. Especially when, as I mention above, it doesn’t fit the reality of all the negative tests results we are seeing across the country. Nowhere are we seeing 50% positive rates among the symptomatic. It is hard to imagine that the asymptomatic would be significantly different.
“Additionally, Italy’s National Health Institute released data that confirmed more than 99% of all COVID-19 fatalities were people already suffering from underlying medical conditions. Only 0.8% of the fatalities were healthy adults … to put that in context, only 86 healthy adults have died from COVID-19 there.” Holy shit. Hypertension is considered a medical condition in this context. So too is Type 2 diabetes. To consider people who have controlled hypertension as “unhealthy” is plain silly. And there is a dangerous libertarian-leaning implication here that the “unhealthy” are somehow deserving of a poor outcome. Or maybe I am reading too much into it, but I don’t think so.
I know that there is probably a hint of truth to some of it, but it just doesn’t read the same to me as my friend, I think. I’m willing to be convinced otherwise, I just haven’t seen anything at compelling enough to flip my thinking yet. And that isn’t me hoping that people will die … I hope the models are wrong. I would love to advance our scientific understanding of where we erred if that were to be the case and I hope that we learn from this experience and are better prepared and have better strategies when something similar happens again in the future.
I shared all this with my friend, nearly verbatim. Their reply?
That’s what this article suggests. Hope you had a good first day at work.
I don’t think they’re angry or bitter. We have had similar conversations in the past. We are always cool after. That is why I enjoy spending time with them; they are one of only 3 non-internet friends that I can talk to about things like this. But – at the same time – I should have known better than to spend so much time trying to convince them of something that flies in the face of their biases and is so difficult to see clearly through a distorted lens of political ideology (in this libertarianism).