Respect The Denominator

These are pressing days for medical people. Among them, the most challenged may well be the epidemiologists. These are the people who determine how bad the this latest novel coronavirus really is, how many people it is infecting, and what to do about it. Not easy.

Perhaps the major difficulty in this outbreak is the fact that our grasp of how many people are infected remains hazy at best. Until only recently, we have had almost no reliable way to even test for the COVID-19 infection. Remember, this thing is totally new. Coming up with ways to test for it takes a minute.

But knowing the number of infected people, as accurately as possible, is the fundamental value driving the entire response to this virus. This number forms the denominator in the fractions that help us determine just how serious this situation is. Fifth grade throwback: the numerator is the top number in a fraction, the denominator is the bottom number. Divide the two numbers, multiply by 100 and you have a percentage.

Let’s consider the death rate calculation: it is a relationship between the number of people who have died from the infection divided by the number of people who are infected with it. If the denominator in that calculation is wrong, our estimation of the deadliness of the virus is wrong too.

Until very recently we’ve been able to fairly accurately determine the numerator in the equation: how many people have died. But we’ve had almost no idea how many people had the infection but didn’t die, because testing has been so limited. Thus, we almost certainly have been underestimating our bottom number in the fraction. When this happens, mathematically it appears that the virus is MUCH more deadly than it really is.

Here’s a simple example: Say we identify 100 people who contract COVID-19. Of those, 5 die. We thus calculate the death rate as 5/100, which equals 5%. This is a scary number. If half the U.S. contracts the virus (about 165 million people), and 5% of them die, that’s more than 8 million people!

But wait! What if we got that bottom number wrong? Oops, it’s not 100 people infected, it’s 100 THOUSAND people. Let’s recalculate: 5/100,000 = 0.005%. This is a much better number.

This little example above bears out in real life, too. The U.S. has been woefully unprepared to test for COVID-19 so far. Between January 18th and March 10th, the U.S. tested only about 11,000 people for COVID-19, which is approximately 7 people per million. Based on that limited number (a very low denominator), the mortality rate in the U.S. was recently calculated to be above 5%. By contrast, South Korea has done MUCH more testing, approximately 1,100 per million people (over 100,000 tests) and their mortality rate is calculated to be 0.6%.

So which is it? 5%, or 0.6%? Speaking at a media briefing on March 3rd, the World Health Organisation’s director-general, Tedros Adhanom Ghebreyesus recently asserted that the mortality rate globally for the COVID-19 outbreak was 3.4%.

Respectfully, I assert that it was irresponsible to make such a claim this early in the outbreak. The mortality rate of the Spanish flu of 1918, estimated to have killed 17-50 million people, was just over 2%. And that epidemic had malnourishment, overcrowded medical camps, poor hygiene and bacterial superinfection to help it along. Today, garden-variety influenza has a mortality rate of 0.1% and has killed more than 20,000 people in the U.S. this year alone. Is the coronavirus that much more deadly?

The answer, which the WHO should have clearly stated, is that it isn’t likely but we just don’t know yet. Our denominator is a mess. Despite the pressure on health officials to provide information and conclusions, it’s important to explain when the data is still insufficient. And while it was said clumsily by President Trump, (“I think it’s probably less than 1%”), he had a point.

The good news is, the “lights are coming on” in the testing world. We are rationing tests less and less, and supplies to administer the tests are increasing daily. This is great, of course, as it will give us a much more accurate idea of just how deadly this infection is to the population.

But there’s another wrinkle with additional testing: It’s virtually certain that we’ll find more cases of coronavirus. Commensurate with how much more we test, we will find many more positive cases too. A sudden surge in positive tests gives the appearance of a sudden surge in infections overall. In reality, those infections were probably already there. But suddenly knowing about them creates quite a spectacle.

So brace yourselves. Often even the well-meaning use of data can lead to inaccurate appearances. In time, we will almost certainly find that the virus is less deadly than it currently appears and that it’s much more prevalent in the population than it currently appears. But that’s just a guess at this point. With a better denominator, we’ll know for sure. Until then, be patient, don’t panic, and stay tuned. It’s probably not as bad as it looks. More to follow.

One thought on “Respect The Denominator

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s