Comparing the Flu with COVID-19
I have a FB friend who's going to be economically hurt by social distancing policies such as bar and restaurant closings in Chicago. Quite understandably, she was asking about how the COVID-19 virus compares to the flu, so I wrote this up for her.
The Death Rate and Infection Rate for the Flu
First, here's the CDC webpage on the number of Flu Cases in the US.
----------------------
Flu Cases in the US (from 10/1/2019 to 3/7/2020)
2019-2020 U.S. Flu Season: Preliminary Burden Estimates
https://www.cdc.gov/flu/about/burden/preliminary-in-season-estimates.htm
-----------------------
They provide a high and low estimate of the number of illnesses, deaths, etc. I'll use the high estimate to illustrate:
Flu Illnesses: 51,000,000 (i.e. 51 million or 51M or 51,000K)
Deaths: 55,000 (i.e. 55K)
Death Rate: 0.108% = 55K/51000K
That's a death rate of about 1/10th of a percent.
There are about 327,200,000 people in the US (or 372.2M)
So, the infection rate is about 15.6%
Infection Rate = 51M/372.2M
The Death Rate for COVID-19
Now for the death rate for COVID-19. I'm using data from Worldometers. The URL will be in the reference section.
The Confirmed Fatality Rate (CFR) that Worldometers reports is based on this equation:
CFR = deaths/(deaths + recovered_cases)
or
CFR = deaths/closed_cases
Note: closed_cases = deaths + recovered_cases
Here's the data for China and other (i.e. countries outside of china) as of 3/14/2020.
China:
Total Recovered: 65,573
Total Deaths: 3,189
Closed Cases: 68,762
CFR: 0.046 (i.e. 4.6%)
Other Countries:
Total Recovered: 8,153
Total Deaths: 2,426
Closed Cases: 10,579
CFR: 0.229 (i.e. 22.9%)
Total all countries:
Total Recovered: 73,726
Total Deaths: 5,615
Closed Cases: 79,341
CFR: 0.071 (i.e. 7.1%)
So, the CFR (or death rate) for COVID-19 is much higher than it is for the Flu.
The big question is whether the infection rate for COVID-19 will be less than, about equal to, or bigger than the infection rate for the flu (which will be between 11% and 15% this past year according to the CDC).
Now, Anglea Merkel made a statement recently that the COVID-19 infection rate for Germany will be between 60% and 70%. I'm not sure how she came up with these numbers, but if accurate then that's disturbing given the CFR for Covid-19 so far. (See the reference section.)
So far, China has had a much lower infection rate, but then again they really clamped down on their population. Anyway, the governments in Europe and USA are trying to reduce the number of people who end up getting infected. They're also trying to "flatten the curve" which basically means they trying to avoid an early spike in cases that could overwhelm the health care system. Right now, the best tool they have is social distancing, which is going to take an economic toll. When the USA gets testing up and running, that will help control the virus, too.
See the reference section for a YouTube video that explains the math of exponential growth using Covid-19 as an example. The video is based on the WorldMeters data.
Thoughts on the various death rates for COVID-19
A couple of days ago, some FB friends were discussing various reported death rates for the covid-19 virus (e.g. the WHO's report of a 3.4% death rate). This helps explain the Confirmed Fatality Rate (CFR) and how it relates to testing (i.e. you'll get different CFRs depending on whether your country is doing a lot of testing for covid-19 or not).
Here's the worldometers webpage on the COVID-19 death rate. Note that there are a number of ways to compute the death rate.
https://www.worldometers.info/coronavirus/coronavirus-death-rate/
When considering the question of the death rate, the three questions I ask are:
- What is in the numerator?
- What is in the denominator?
- Are these figures predictions of what will [supposedly] happen in the future, or are these figures of what has [supposedly] happened in the past?
So, if you are just looking at closed_cases (i.e. deaths + recovered_cases) in the denominator, then you're looking at past data. So, one way to calculate the Confirmed Fatality Rate (CFR) with past data is:
CFR = deaths / (deaths + recovered_cases)
or
CFR = deaths / closed_cases
Note that this version of a "death rate" does not say anything about the percentage of a population that might get infected in the future or the number of currently active cases. The current "past data only" CFR is based on the data we have right now. It would be nice to know what the CFR will be at some point in the future (e.g. when the crisis is over), but getting that figure now would involve time travel. :-)
Anyway, this worldometers webpage provides a discussion of various ways of computing the death rate (e.g. the WHO's estimate etc.)
I should add one more point about "past data."
Situation A: You're dealing with a population where everyone is being tested (and the test works), then the past data for recovered_cases makes sense.
Situation B: You're dealing with a population where there is no testing (i.e. people show up at the hospital with obvious and often severe symptoms) then the past data for recovered_cases is problematic. When you go to calculate the CFR, you'll get a higher rate of death (than in Situation A) because really sick people are showing up at the hospital. On the other hand, you have a smaller number of "cases" (than in Situation A) because you aren't doing any testing.
Obviously, you are in a better position to control the virus if you are in Situation A. In this case, you can identify people who are infected and take steps to ensure that they do not infect others.
Thoughts on the data coming out of China
The YouTube video on Exponential Growth and Epidemics focuses on countries outside of China. They use the WorldoMeters website, which has numbers for China, too. But some folks question these numbers.
Maybe that's because the data coming out of China for the total number of cases (80,849) is suspect for a country with 1,386 million people. That's an infection rate of only 0.0058%
Merkel says that she's expecting an infection rate for Germany of between 60% and 70%. The Exponential Growth and Epidemics video may cast some light on where she's getting these figures. It has to do with when the "growth factor" for new cases levels off. At this point, the exponential growth curve starts to become an "S" curve (i.e. a logistic curve).
As for myself, I like to read articles and reports with opposing points of view.
References
----------------------------------------
Begley, Sharon
Lower death rate estimates for coronavirus, especially for non-elderly, provide glimmer of hope
By SHARON BEGLEY
MARCH 16, 2020
https://www.statnews.com/2020/03/16/lower-coronavirus-death-rate-estimates/
In a rare piece of good news about Covid-19, a team of infectious disease experts calculates that the fatality rate in people who have symptoms of the disease caused by the new coronavirus is about 1.4%. Although that estimate applies specifically to Wuhan, the Chinese city where the outbreak began, and is based on data from there, it offers a guide to the rest of the world, where many countries might see even lower death rates.
The new figure is significantly below earlier estimates of 2% or 3% and well below the death rate for China based on simply dividing deaths by cases, which yields almost 4%. While it is still higher than the average 0.1% death rate from seasonal flu, it raises hopes that the worst consequence of the coronavirus will be uncommon.
-----------------------------------------
Beshears, Fred
worldometers_coronavirus.pdf (3/17/2020)
https://drive.google.com/open?id=1slHZaQVnr7AmIG-59PdAxv1ndbuz5K8t
I've been maintaining a spreadsheet on the WorldOMeters coronavirus data. Here's a PDF of that spreadsheet. To better understand the data presented here, check out the video Exponential Growth and Epidemics. To go to the source, go to the WorldOMeters website for the coronavirus.
-----------------------------------------
Coronavirus Act Now Tool
According to the webpage for Coronavirus Act Now:
"""
The model predicts the last day each state can act before the point of no return. The model is intended to help make fast decisions, not predict the future.
"""
I cannot speak for the accuracy of the predictions it offers. However, it would be nice to get a tool like this from the CDC. I would have more faith in their predictions.
In any case:
For each state, it looks at four scenarios:
- No Action, Current Trends Continue
- 3 Months of Texas-style delay/social distancing
- 3 Months of California-style "shelter-in-place"
- 3 Months of Wuhan-style Lockdown
For each of the above scenarios, it shows the following:
- the estimated cumulative infected as a percent of the population
- the estimated date hospitals overload
- the estimated number of deaths
https://covidactnow.org/
-----------------------------------------
Exponential Growth and Epidemics
https://www.youtube.com/watch?v=Kas0tIxDvrg
- The data used in this video came from the worldometers website.
- It looks at the exponential growth curve of covid-19 cases outside of China.
- Note that in the real world, exponential growth curves don't to on forever. Eventually they turn into logistic curves (i.e. S curves, which level out).
- The video also looks at the question of how one determines when an exponential growth curve has hit its inflection point.
-----------------------------------------
Fink, Sheri
White House Takes New Line After Dire Report on Death Toll
Federal guidelines warned against gatherings of more than 10 people as a London report predicted high fatalities in the U.S. without drastic action.
By Sheri Fink
3/16/2020
https://www.nytimes.com/2020/03/16/us/coronavirus-fatality-rate-white-house.html
Sweeping new federal recommendations announced on Monday for Americans to sharply limit their activities appeared to draw on a dire scientific report warning that, without action by the government and individuals to slow the spread of coronavirus and suppress new cases, 2.2 million people in the United States could die.
-----------------------------------------
Flu Cases in the US (from 10/1/2019 to 3/7/2020)
2019-2020 U.S. Flu Season: Preliminary Burden Estimates
by the Center for Disease Control (CDC)
https://www.cdc.gov/flu/about/burden/preliminary-in-season-estimates.htm
------------------------------------------
Kaletsky, Anatole
FMB: Notice that Kaletsky has dramatically changed his view since March 10th. His latest article, dated March 19, states that western countries are facing a medical disaster. The number of new cases is not leveling off the way it did in China (outside of Hubei provence), Sinapore, and South Korea. Western countries are also facing an economic disaster, but he believes that this will actually be the easy part.
Averting Economic Disaster Is the Easy Part
ANATOLE KALETSKY
Mar 19, 2020
Based on China’s experience with COVID-19, the fiscal cost of comprehensive compensation for lost income could reach 10% of annual GDP, and as much as 25% of GDP in the US and Europe if the epidemic turns out to be worse there, which now looks likely. These may seem like mind-boggling sums, but they can be financed in several ways.
https://www.project-syndicate.org/commentary/government-compensation-for-covid-19-losses-by-anatole-kaletsky-2020-03
LONDON – Contrary to my initial expectations, the spread of the COVID-19 coronavirus around the world is not following the relatively benign trajectories experienced in China outside of Wuhan and Hubei province, and in South Korea, Singapore, and the rest of Asia. Instead, across Europe – and very likely in the United States, too – the spread of the virus increasingly resembles the path it took in Hubei.
This threatens both medical and economic disaster. But while it may be too late for policymakers to avert a public-health crisis, it is still possible to implement the fiscal and monetary measures needed to prevent an economic catastrophe. To do this, they will need to go much further than the monetary steps announced by the US Federal Reserve and the Trump administration’s proposals for untargeted tax cuts and cash handouts thus far.
Initially, I expected that the number of cases in European Union countries would converge toward the 10-100 patients per million seen in Asia outside Hubei, and that the US might follow a roughly similar pattern. The fact is, however, that Italy, France, Spain, and other EU countries are not experiencing the slowdown in the rate of change of acceleration (the second derivative of velocity, or the “jolt” in mathematical parlance), seen in Wuhan and the rest of China by this stage in the epidemic.
--------------
COVID-19 by the Numbers
Mar 10, 2020
ANATOLE KALETSKY
Callous as it may sound, the economic and political impact of the coronavirus pandemic will ultimately be determined by the epidemiological and clinical data. Fortunately, in this case, the relevant statistical trends are developing in a much less alarming way than panicked media headlines might suggest.
https://www.project-syndicate.org/commentary/covid19-data-do-not-support-global-alarmism-by-anatole-kaletsky-2020-03
LONDON – One death is a tragedy; a million deaths is a statistic. Whether or not Stalin actually said this, it is a cruelly accurate description of economic reality. The global panic about the new coronavirus, COVID-19, is understandable, because every premature death is a human tragedy. But, callous as it may sound, the economic and political impact of this pandemic will ultimately be determined by the numbers. Fortunately, in this case, the relevant ones are developing in a much less alarming way than panicked media headlines might suggest.
Anatole Kaletsky is Chief Economist and Co-Chairman of Gavekal Dragonomics. A former columnist at the Times of London, the International New York Times and the Financial Times, he is the author of Capitalism 4.0: The Birth of a New Economy in the Aftermath of Crisis, which anticipated many of the post-crisis transformations of the global economy. His 1985 book, Costs of Default, became an influential primer for Latin American and Asian governments negotiating debt defaults and restructurings with banks and the IMF.
-----------------------------------------
Khan Academy
Estimating actual COVID 19 cases (novel corona virus infections) in an area based on deaths
by Khan Academy
https://www.youtube.com/watch?v=mCa0JXEwDEk
FMB: Very nice explanation about how you can take current reported deaths from COVID-19 in your area and work backwards to estimate how many infected people are in your area. It highlights the importance of getting testing underway in the USA. Otherwise, many people will seriously underestimate the number of infected people in their area because the reported number of deaths seem low. You need to do the math, and Khan Academy helps with that.
------------------------------------------
Merkel, Angela
Angela Merkel on Infection Rate in Germany
Most people in Germany will get COVID-19, aim is to slow its spread: Merkel
3/11/2020
https://www.channelnewsasia.com/news/world/germany-covid19-coronavirus-angela-merkel-12527596
BERLIN: Up to 70 per cent of the population is likely to be infected with the coronavirus that is currently spreading around the world, German Chancellor Angela Merkel said on Wednesday (Mar 11), adding that since there was currently no cure, the focus had to be on slowing its spread.
"When the virus is out there, the population has no immunity and no therapy exists, then 60 to 70 per cent of the population will be infected," she told a news conference in Berlin.
-----------------------------------------
MinutePhysics
How to tell if we're beating COVID-19
by MinutePhysics
https://www.youtube.com/watch?v=54XLXg4fYsc
-----------------------------------------
Worldometers Data
Covid-19 Case Data for Countries Outside of China
https://www.worldometers.info/coronavirus/coronavirus-cases/#case-tot-outchina
Comparing Various Covid-19 Death Rates
https://www.worldometers.info/coronavirus/coronavirus-death-rate/
Comments
Post a Comment
Ask me anything here...