Coronavirus: Is the CDC's CFR % even realistic?

On Tuesday, I wrote a post looking at a couple projections that were made regarding the CFR % and IFR % of coronavirus. This is what the CDC estimate was via a Reason article:

According to the Centers for Disease Control and Prevention (CDC), the current "best estimate" for the fatality rate among Americans with COVID-19 symptoms is 0.4 percent. The CDC also estimates that 35 percent of people infected by the COVID-19 virus never develop symptoms. Those numbers imply that the virus kills less than 0.3 percent of people infected by it—far lower than the infection fatality rates (IFRs) assumed by the alarming projections that drove the initial government response to the epidemic, including broad business closure and stay-at-home orders.

Here we have the CDC saying the CFR % is 0.4%. I think the general consensus is that the CFR % for the seasonal flu is 0.1%. This indicates that coronavirus is 4x more deadly than the seasonal flu.



The below is my reasoning for why I believe the CDC will be revising their coronavirus CFR % up in the near future.

I decided to do some back of the envelope calculations. I went to the section of the CDC website where they keep stats on flu deaths. If you click on the link, you'll see estimated deaths for flu seasons from 2010-2011 to 2018-2019. The average deaths per years comes out to 37,462. If coronavirus is really just 4x more deadly than the flu, that would indicate that we should see around 149,848 deaths due to coroanvirus.

Now there are a number of caveats here. A few that I can think of include:

1. Technically, it looks like the CFR % for the flu is really 0.13%. I got that by adding up the symptomatic column and divided it by the death column in the CDC link.

2. I'm making the assumption that the symptomatic cases for both the flu and coronavirus are around the same when I come up with the 149,848 deaths estimates. Of course, that is a seriously wrong assumption, but that kind of makes my point that I'm about to make about why I think the CFR % for the coronavirus is seriously off.

I'm writing this article a few days before posting. At the time of my write-up, Worldometer had the number of deaths in the United States at 140,000. After a brief drop in deaths to less than 500 a day, it looks like daily deaths are picking up. I think it would be reasonable to say that we'll get to 150,000 deaths by July 27th.

Worldometer also had US case count at 3,617,000. We've been averaging over 65,000 cases recently. I think by July 27th, it might be reasonable to assume we'll be at 4,400,000 confirmed cases.

If we take 150,000 and divide by 4,400,000, we get to a CFR % of 3.4%.

Now there are a number of caveats to this, as well. A few that I can think of include:

1. I am hearing complaints that coronavirus deaths are being over-estimated and that just because someone died had coronavirus doesn't mean that they actually died because of coronavirus. I've heard the argument about how someone who had coronavirus and died due to a car accident might be counted as having died from coronavirus vs a car accident.

2. On the other hand, I hear complaints that coronavirus deaths are being under-estimated.

3. Sure, maybe when we get to 150,000 deaths people can reasonably argue that deaths aren't really 150,000. Yet, can't the same be said about the average flu deaths of 37,462 - that this number isn't accurate? Maybe 1,000 people are being falsely put down as coronavirus deaths. It doesn't really impact this analysis.

4. The CFR % of 3.4% is based on total deaths divided by total symptomatic cases. Can we really saying that only 4,400,000 (again, an estimate amount that we'll be at by July 27th) were symptomatic for coronavirus? I would agree that my CFR % of 3.4% is probably too high as a lot of people especially early on just weren't tested; however, to get to the CDC target we'd have to assume that 37,500,000 Americans were symptomatic for the virus. That would be over 10% of the population and would imply that we only confirmed 13% of those who are/were symptomatic.

5. Of my 4,400,000 estimate, there are still many of those people who will die, which would increase my 3.4% figure.

As a summary:

1. I don't have all the data points that the CDC has.

2. Due to lack of data, I don't feel like I can come up with the accurate CFR %.

3. Yet, I'm fairly confident that the CDC will have to revise their estimates up significantly.