R "Ray" Wang
R "Ray" Wang
Data In Early Case Fatality Rates Are Naturally Biased To Show Massive Fatality
While there's nothing wrong with an abundance of caution for high case fatality rates we are seeing for the COVID-19 coronavirus, the data is not accurate. We keep hearing 3.5% or 4% of the population is going to die. Why is the rate so high? The denominator is inaccurate. Most countries have not done broad testing to know how many cases are prevalent in the general population. So, let's start with the definition of Case Fatality Rate.
Case fatality rate is calculated by dividing the number of deaths from a specified disease over a defined period of time by the number of individuals diagnosed with the disease during that time; the resulting ratio is then multiplied by 100 to yield a percentage.
1. Case Fatality Rate or Mortality Rate = Number of Deaths / by Total Number of Cases X 100
2. Total Number of Cases = Prevalence
3. Prevalence is all the reported cases AND the estimated cases in the environment
The denominator here is very important. What makes up the total number or cases is all the reported cases that we know of in the hospital and the broad sample of what’s in the environment.
A good example of why the rates look so scary at first, can be shown in South Korea in early reporting. The early cases were only the sick ones or those who fell ill. After broad testing in South Korea, the case fatality results were 0.6%, much lower than earlier results of 3 or 4% of case fatality rates in early reporting.
Public Response To Data Fails To Account For Accurate Prevalence In Case Fatality Rates
After broader testing, you could see how fast the virus had spread and how much lower the number of deaths were. Don’t get me wrong, this virus is very contagious. However, the virus is not as deadly as some may have first believed. Moreover, it's not from watching the media and folks on social media going nuts, screaming, "Oh my god, this is the Bill Gates 100 year Spanish Flu pandemic!"
Understand how case fatality rates are studied, then we can figure out the appropriate proportionality of response.
IN THE US, WE HAVE NOT DONE BROAD TESTING. WE COULD ALL BE CARRIERS AND NOT SHOW IT.
The Bottom Line: Understand Proportionality Of Response Before We Do More Self-Inflicted Damage To The Economy
Let's take another way to look at our response to this outbreak:
In the US, prevalence of a specific type of flu was 15M as of Jan 2020. We had:
- 140k hospitalizations
- 8200 deaths,
- 54 pediatric deaths
What would you do in that situation?
- quarantine everyone?
- cancel events?
- stop sports?
- hunker down?
- close schools?
That's Influenza B. A known flu which we even have vaccines for, albeit they don't always work so well and we don't all take them. The CDC reports less than half of American adults got a flu shot last season. Even more interesting, only 62 percent of children got the vaccine, despite being vulnerable to respiratory illnesses.
We don’t go crazy on the flu because we’re accustomed to the risk and have factored for it. Right now we’re going ape $sh!t because of imperfect data and taking a massive abundance of caution (nothing wrong with that).
However, the response to this crisis is 10X of what we do for the normal flu. Either we step up when the regular flu shows up in the same manner, shut down everything, and self-inflict wounds to bring down 0.5% to 1.0% of global GDP, or let's get a grip on the panic.