BC: SHOE LESS JOE JACKSON COMES TO IOWA
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
Executive Summary
Right now 2 million Americans are infected with the coronavirus.
The total U.S. death toll by April 15th will be more than 20,000.
We estimate that 80 thousand of the 2 million infected Americans will be
hospitalized over the next 2 weeks.
Thesis:
Three parameter estimates are needed to predict the number of infections and
number of deaths over the next 3 weeks: infection fatality rate, infection growth rate,
and the number of days between initial infection and resolution (either death or recovery)
of the infection.
1. We Know estimate that the coronavirus’s fatality rate is ~0.8%.
This means 1 out of every 125 infected people will die.
We know that almost all countries had problems with testing and identifying all infected people.
There are two exceptions to this: South Korea and Japan’s Princess Diamond cruise ship.
South Korea tested more than 320,000 people and identified 8652 infections.
The total number of deaths was 94. This means South Korea’s case fatality rate is 1.09%.
We believe there are still a considerable number of South Koreans who were asymptomatic
and weren't tested. So, we estimate that the actual fatality rate is anywhere from 0.5% and 1%.
In early February the Princess Diamond cruise ship was quarantined in Japan after
one of the passengers tested positive. This was a bad idea for passengers as a total
of 712 passengers were eventually infected and 7 of these people died.
As far as I know all 3000+ passengers of this cruise ship were tested, so we have a reliable
dataset with pretty accurate number of infections and number of deaths. The case fatality rate
on Princess Diamond is 0.983%. We know that the fatality rate is higher among older people.
Assuming that the median age of passengers on Princess Diamond is greater than America’s,
which is 39, we can estimate that the new coronavirus’ fatality rate will be around 0.8% in
America (maybe a little lower, but this is a nice round number).
2. This paper estimates that an infection takes around 23-24 days to resolve.
The first 5-6 days the patient doesn’t show any symptoms. It takes an average of 5 days
between the onset of symptoms and hospitalization. Finally it takes about 2 weeks between
hospitalization and death (about 10 days in ICU).
3. There is no accurate direct way of calculating the infection growth rate because there
is a huge variation in the number of people we are testing. [We were testing only a few people
a couple of weeks ago, so we identified only a small number of infections. In recent days we
started testing a large number of people (especially in New York) and now our case count is
growing at alarming rates.] However, we can assume that the death rate is constant and we
use the change in deaths to estimate the infection growth rate. There is a 23-24 day lag between
an infection and the resolution of that infection. This means the growth rate in the number of deaths
today is a very good estimate of the infection growth rate 23-24 days ago.
On March 19th, the U.S. reported a total of 205 deaths. That figure was 85 on March 16th and 47
on March 13th. This means the number of deaths doubles about every 3 days. This also means
that the number of infections were doubling every 3 days on February 25th (the people who are
dying today were infected on February 25th)." On March 19th, the U.S. reported a total of 205 deaths.
That figure was 85 on March 16th and 47 on March 13th. This means the number of deaths doubles
about every 3 days. This also means that the number of infections were doubling every 3 days on
February 25th (the people who are dying today were infected on February 25th).
So, here is my simple mathematical model.
"For the sake of argument, I am going to assume that all 205 American deaths occurred on March 19th
and all of these people were infected on February 25th (this assumption simplifies calculations,
we don’t need a complex model to have a thorough understanding of what is going on).
For the sake of argument, I am going to assume that all 205 American deaths occurred on March 19th
and all of these people were infected on February 25th (this assumption simplifies calculations,
we don’t need a complex model to have a thorough understanding of what is going on).
Our estimate for the fatality rate is 0.8%; this means for every death we have 125 infections. Since
we have 205 total deaths, there must have been 205 times 125 total infections on February 25th.
That’s 25,625 infected people. If you understand this part of the calculation, the rest of our analysis
is pretty straightforward. Our estimate for the fatality rate is 0.8%; this means for every death we have
125 infections. Since we have 205 total deaths, there must have been 205 times 125 total infections on
February 25th. That’s 25,625 infected people. If you understand this part of the calculation,
the rest of our analysis is pretty straightforward.
The number of infected people doubles every 3 days. So, on February 28th the number of infected
people doubled to 51,250 (let’s round it down to 50,000). Three days later, on March 2nd, the number
of infected people doubled again to 100,000.The number of infected people doubles every 3 days.
So, on February 28th the number of infected people doubled to 51,250 (let’s round it down to 50,000).
Three days later, on March 2nd, the number of infected people doubled again to 100,000.
There were 100K infected people on March 2nd in America. We know that 0.8% of these people will
die by March 26th. That means our death toll will be 800 on March 26th [you can verify the accuracy of our model on March 26th by comparing the actual death toll to our estimate] Our model tells us that the number of infections doubled again on March 5th, reaching 200,000. Our model also tells us that the number of infected people was 400,000 on March 8th, 800,000 on March 11th, and 1.6 million on March 14th.
These calculations imply that the American death toll will be 12,800 on April 7th. To put that in
perspective, yesterday, the total death toll in Italy was 3400 and 3000 in China.
I know that these are just estimates, but even if my estimates are off by 50%, we will have still
twice as many coronavirus deaths as China 2.5 weeks from now.
On March 14th, various municipalities and agencies started introducing social distancing.
The practice eventually began to be suggested or required in the hardest hit parts of the nation.
The good news is that we started cancelling schools and closing down restaurants around March 14th.
So, the number of total US infections isn’t doubling every 3 days anymore. Unfortunately, the horse
is already out of the barn. As of March 14th, one out of every 200 Americans is already infected.
Italy put the entire country under lockdown 12 days ago, yet its death toll is still increasing
exponentially. That’s because there is a 24 day lag between an infection and its resolution.
We haven’t put our country under a lockdown yet.
Except a few educated people, no one has any idea that there are already around 2 million
infected people in America today and the American death toll will exceed 15,000 in just 24 days.
If we don’t take strict measures, we will be reporting 1000 deaths per day in just 3 weeks.
The attacks on 9/11 killed around 3000 people. We will be reporting a 9/11 every three days.
That’s why we say “hell is coming”.
This is a mathematical certainty. It is inevitable.
Assuming that we were able to put an end to the coronavirus outbreak after a 2 month lockdown,
we will still have to implement strict quarantine measures towards international travelers.
Tourism accounts for 2.9% of our GDP. We will have to deal with high unemployment figures,
bankruptcies, and ballooning budget deficits for some time to come. The bad news is that almost
every country in the world will experience a recession. |
