Do What I Mean
March 23rd, 2020
Just like everybody else COVID-19 is on my mind constantly these days. In this post I look at what is going on according to main stream media, what other views experts may have, and I try to figure out what is happening.
I am not an expert: I am not a virologist, nor a data scientist. So, my intention is to rely on those who are. But between them and me are the media, politics and public opinion, which may obscure my view if I am not careful. So, let's dig in.
My heart goes out to all the victims of this virus. I wish strength to those who are sick, those who have lost or fear the loss of a loved one, and the care-givers who are working hard to help them. You deserve our support!
These 11 charts [March 17] nicely summarize the data we see in the news every day. The gist:
|infection rate (R0)||1.3||2.3|
|incubation time (days)||1-4||1-14|
In his address [March 16, French] to the nation the French President Emmanual Macron said: "We are at war, in a health war, of course: we are not fighting against an army or against another Nation. But the enemy is there, invisible, elusive, advancing. And that requires our general mobilization."
If statements like this don't strike fear in the hearts of citizens, nothing will...
This prompted me to see if I could find out more.
The death rate of COVID-19 is estimated by the WHO to be 3.4%, based on numbers from Wuhan. But newer reports [March 20] suggest the number was much lower there, more like 1.4%.
The Centre for Evidence-Based Medicine (CEBM) of the University of Oxford explains that the Infection Rate Fatality (IFR) differs from the Case Fatality Rate (CFR) in that it aims to estimate the fatality rate in all those with infection: the patients who have been tested positive (cases) and those with an undetected disease (asymptomatic and not tested group). "Our current best assumption, as of the 22nd March, is the IFR is approximate 0.19%." What this means is clearly visible in Germany where 24,873 people have been tested positive and 94 have died (0.4%): the point here is that Germany performs 160,000 COVID-19 tests every week which means they find far more infections than any other country, but their fatality rate is much lower.
Ioannidis refers to the Diamond Princess Cruise Ship as one situation where an entire, closed population was tested (six deaths occurred out of 705 who tested positive constituting a CFR of 0.85%. All six deaths six occurred in patients > 70. No one under 70 died): "Projecting the Diamond Princess mortality rate onto the age structure of the U.S. population, the death rate among people infected with Covid-19 would be 0.125%." (the article continues with further nuances). This is confirmed by this study: comparing deaths onboard with expected deaths based on naive CFR estimates using China data, they estimated CFR 1.1% and IFR 0.5%.
To put things in perspective, the seasonal flu in the Netherlands (17 million people) cost 9,500 deaths in season 2017/2018 [Dutch] and hardly anybody noticed. Compare this to the 179 deaths from COVID-19 in The Netherlands now (March 22), or the 5476 deaths in Italy (Lombardy has 10 million inhabitants) and the 3275 deaths in China (Wuhan has 11 million). The season 2017/2018 was a severe flu season in The Netherlands but other seasons [Dutch] still show significant numbers.
We have had similar outbreaks of viruses in the past, like the 2009/2010 swine or Mexican flu pandamic with estimates of 700 million-1.4 billion confirmed cases and 151,700-575,400 deaths. These are staggering numbers, but in The Journal of the American Medical Association as early as Sept 8 2010 Edward A. Belongia and colleagues report numbers that suggest that the swine flu were no worse than the seasonal flu. The WHO confirmed in 2019 that swine flu ended up with a fatality rate of 0.02%. And this pandemic also saw the Mexican government closing most of Mexico city and clinics in some areas being overwhelmed by infected people.
CEBM conclude that evaluating CFR during a pandemic is a hazardous exercise, and high-end estimates must be treated with caution as the H1N1 pandemic highlights that original estimates were off by a factor greater than 10.
Flatten the curve and Social Distancing are the current strategies of choice. CEBM states: "Although limited, the best available evidence appears to support social distancing measures as a means of reducing transmission and delaying spread. Staggered and cumulative implementation of these interventions may prove most effective. The timing and duration of such measures is critical." but also: "The effect of restricting and cancelling mass gatherings and sporting events on infectious diseases is poorly established and requires further assessment. The best-available evidence suggests multiple-day events with crowded communal accommodations are most associated with increased risk. Mass gatherings are not homogenous and risk should be assessed on a case-by-case basis."
Michael Levitt, an American-British-Israeli biophysicist who won the 2013 Nobel prize for chemistry has monitored [March 20] the virus in China since the early days. He predicted the current decline in number of infections two weeks before it happened and now predicts the virus will disappear from China by the end of March. This explanation [March 11, Dutch] by Applied Statistics professor Casper Albers shows why the exponential model to predict new cases doesn't fit the data, but that logistic regression does and probably provides a better model.
Notice the definition of the infection rate R0:
the expected number of cases directly generated by one case in a population where all individuals are susceptible to infection.
The key words here are "susceptible to infection". Of course, in time not all individuals will be susceptible to infection anymore and the virus will die out. And the more contagious a virus is, the sooner this will happen.
The social media bombard us with cries of panic and pleas for help from doctors and nurses from Italy, but we also get this interview [Feb 26, Italian] with Maria Rita Gismondo, the director of the Laboratory of Clinical Microbiology, Virology and bio-emergencies in Milan, who states "There is a bombardment of news that foment fear, there has been a collective brainwashing", but COVID-19 "is little more than a normal flu", and "we are not at war". For this, Gismondo has received serious flack [22 March, Italian] but that hasn't changed her mind.
Current data cannot tell us yet how bad COVID-19 actually is. It may be bad, but it may also be comparable to the common flu. We just don't know yet.
That makes we wonder why the response is so strong and why COVID-19 has lead to draconian measures like totally locking down entire countries. I can come up with a number of possible explanations:
From the looks of it, I get the impression that COVID-19 is fast. The numbers may be similar to those of the common flu, but they are achieved in weeks instead of months. That pushes us to also respond fast. Don't think, act!
In marketing FUD, for Fear, Uncertainty, Doubt, is a well-known strategy. It pushes people towards the perceived safer option.
These days we are used to big data giving immediate answers. We are used to be in control: when you're sick, go to the doctor. The doctor knows best.
So, when we see experts without answers and doctors panicking, we start to panic too, and flock to anyone who pretends to have a solution.
Even during the 2009 swine flu pandemic, social media were not as ubiquitous as they are today. Back then we got updates a couple of times a day, mostly text, sometimes with a photo. These days we get updates almost live and with video. That has a significantly different impact by instilling a constant sense of urgency.
If we got similarly frequent updates about the common flu during flu season, we might respond as we do now. In fact, similar maps exist, but probably don't get the same amount of traffic as those for COVID-19 (this map shows the global numbers and this map shows the Dutch situation).
Moreover, even more than in tradition news sources, on social media panic mongers get the same amount of screen time as experts, if not more. Experts are difficult, while demagogues provide snappy sound bites. We bother less and less with the longer, more complicated, but also more nuanced articles.
I feel like we live in times where small vocal groups have a strong grip on the media and politicians.
In The Netherlands the official guide lines are:
Schools have been closed since March 16. Not because experts thought it necessary, but because [Dutch] "society voted with their feet". In other words: the government had no other choice, because school leaders wanted to close and parents kept their kids at home.
Also, on social media people are shamed for visiting parks, letting their children play outside, and generally being social. Usually by Karens on Twitter, celebrities in talk shows or other non-experts. They seem to feel the need to 'educate' their fellow citizens and demonstrate their superior correctness.
But also local mayors and individual medical professionals feel the need to be stricter that the official guide-lines. They seem to think they know better than multi-disciplinary teams of experts that weighed the pros and cons of all options. As a software developer, I know that local optimization may have disastrous global side-effects.
We just don't know yet, and anybody who suggest they do are lying.
My two cents, and remember I am no expert, so this is not much more than divination and is intended primarily for myself to look back and see where I was most wrong:
Most importantly: Let's keep calm and use our heads.