COVID-19: Down the Rabbit Hole of Death Rate Calculations - Part 3

Note: Written June 25th.

Before we continue, let's go over why I think this death rate is important to get right and to understand. From the NPR article:

The new CDC guide for modeling deaths and hospitalizations underscores how politicians, government officials and medical experts still disagree on how deadly the coronavirus is and how much caution to take when reopening the nation's economy.

If the CDC is telling our politicians, hey...its 0.27%...someone is going to whisper in their ear that that number is not that much higher than the flu, and nobody shuts down the economy for the flu do they!

If you are going to model how much death and hospitalizations we can expect, and you plug in that number of 0.26% for your Infection Fatality Rate (IFR) then you are going to calculate a lower number of deaths than what we will actually see.

How then should we look at the Infection Fatality Rate? How about we use data provided by the Imperial College of London's COVID-19 Response Team? Let's look at their report dated 24 May 2020

The number of deaths is then a function of the number of infections and the infection fatality rate (IFR).

Deaths = Number of Infected x IFR.

The CDC tells us this on their FAQ page:

The mortality rate is the percentage of people who died due to COVID-19 out of the total number of people with COVID-19 reported. Since this is an ongoing outbreak, the percentage can change daily. There are several reasons for this, such as there may be delays in reporting of additional confirmed cases of COVID-19 and not all COVID-19 cases will be detected.

That's the Case Fatality Rate. Basically it says, as of  recently, at least two out of every 100 positive tests will result in death.

That is consistent throughout the world. That's how the data works out using the formulas provided.

But there is a different story going on here that is also important, and that's the actual number of people who get COVID-19. This number includes all of those who were tested plus all of those who were asymptomatic and were not tested.

That's the IFR, and that's the number that seems to be the one that gets held up as "see, it's not that bad."

From a public health point of view both numbers are important I feel. The CFR tells me what my hospitals will expect and the IFR tells me how close we may be getting to herd immunity. Basically, the more people with antibodies the less people that can get the disease.

Why is there a difference in the number of reported cases and the number of people with the antibodies? Well it has to do with the testing that is done.

• The first type is a diagnostic test. This type of test tells you if you have a current infection by looking for parts of the virus itself. Swabs that take samples from the back of the nose, mouth, or lower respiratory tract are used for these tests. FDA-authorized diagnostic tests for SARS-CoV-2 are accurate for finding a current infection. This means a positive or negative result from a test is likely to give you a true test result.

• The second type of test is a serology (or antibody) test. These tests tell you if you had a previous infection by looking for antibodies in the blood.  Antibodies are proteins made by the immune system when a germ enters a person’s body. Our immune systems help us fight off germs and diseases. The test uses a blood sample to look for antibodies made in response to SARS-CoV-2 rather than looking for the virus itself. It usually takes 1-3 weeks for the body to make antibodies in response to an infection. We do not know how long the antibodies stay in the body after the infection is over. Serology tests have limited ability to diagnose COVID-19 and should not be used alone to diagnose COVID-19. Results from these tests should also not be used to make decisions about staffing or the ability of an employee to return to work, decisions about the need of available protective equipment (PPE), or the need to discontinue social distancing.

If we want to understand the IFR we need to know who has been infected and that requires the second test. However...there is some controversy in doing that:

The survey results, from Germany, the Netherlands, and several locations in the United States, find that anywhere from 2% to 30% of certain populations have already been infected with the virus. The numbers imply that confirmed COVID-19 cases are an even smaller fraction of the true number of people infected than many had estimated and that the vast majority of infections are mild. But many scientists question the accuracy of the antibody tests and complain that several of the research groups announced their findings in the press rather than in preprints or published papers, where their data could be scrutinized. Critics are also wary because some of the researchers are on record advocating for an early end to lockdowns and other control measures, and claim the new prevalence figures support that call.

Remember that comparison to the flu? From that same article:

A German antibody survey was the first out of the gate several weeks ago. At a press conference on 9 April, virologist Hendrik Streeck from the University of Bonn announced preliminary results from a town of about 12,500 in Heinsberg, a region in Germany that had been hit hard by COVID-19. He told reporters his team had found antibodies to the virus in 14% of the 500 people tested. By comparing that number with the recorded deaths in the town, the study suggested the virus kills only 0.37% of the people infected. (The rate for seasonal influenza is about 0.1%.) The team concluded in a two-page summary that “15% of the population can no longer be infected with SARS-CoV-2, and the process of reaching herd immunity is already underway.” They recommended that politicians start to lift some of the regions’ restrictions.

That article's date was April 21, 2020. Let's look at those numbers in Germany:

If you apply the same logic to the US, that is a 0.37% death rate and 15% of the population already infected, you would expect to see similar results, correct? We started at the same time and the virus should behave similar in terms of spread and asymptomatic. Let's look at the US numbers:

That looks different...well for one thing, in the US our politicians "started to lift some of the regions’ restrictions.:

What's going on?

Deeper into the rabbit hole...deeper still.

Part 4