The problem with that 3.4 per cent mortality rate

World Health Organization’s widely quoted stat requires some context

Posted March 9, 2020By Dr. Peter Nord, Chief Medical Officer, Medcan

People walk through the crowded core of a major city

Recently, the World Health Organization’s director-general told media that the COVID-19 coronavirus has a mortality rate of 3.4 per cent. That’s a scary number—and correctly interpreting it requires a lot more context. 

In fact, most experts believe that the novel coronavirus will ultimately have a mortality rate of much less than 2 per cent. That’s a fraction of SARS’ 10 per cent mortality rate, and smaller than ailments like smallpox (3 per cent even in vaccinated people) or the Spanish flu of 1918 (more than 2.5 per cent). 

So how to account for the difference between the 3.4 per cent that the director-general cited, and the less-than-2 per cent that comes from most other experts? 

Here is what the director-general actually said: 

“Globally, about 3.4 per cent of reported COVID-19 cases have died. By comparison, seasonal flu generally kills far fewer than 1 per cent of those infected.”

The crucial part of that statement is the “reported” qualifier.

Many jurisdictions aren’t yet able to accurately report the actual number of people who are afflicted with the virus. Consider what happens at the beginning of an outbreak. Authorities at first count the most serious cases—the ones who go to hospital with obvious symptoms. But because screening procedures haven’t fully been implemented, officials vastly undercount the number of people who actually have the virus.

A mortality rate amounts to a fraction.

The numerator, or top numeral, is the number of deaths. The denominator, or bottom numeral, is the number of people who have tested positive for the virus. In the case of the coronavirus, the numerator is the easier figure to calculate, because a death is such a definitive and traumatic event. 

What is much more difficult to define accurately in the early days of a virus outbreak is that denominator. Probably what’s happening in places where the mortality rate looks higher is that health officials don’t yet have a good grasp of the number of people infected. They’re undercounting the sample size. They may not be able to obtain the number of test kits required to confirm whether sick people actually have the coronavirus, or get those kits to people who are infected. It’s not a numerator problem, but a denominator problem. 

Consider the situation in Italy. As I write this, the Johns Hopkins coronavirus tracker reveals that 197 people have died in Italy, out of a pool of 4,636 infections. Dividing deaths by reported cases in Italy reveals a mortality rate of about 4 per cent—but it’s too early in that outbreak to use that stat.  

Envision an infectious disease sample size as an iceberg.

If you picture the iceberg, the very top is the death rate, which is very small, and then you have individuals with severe illness, and then you have people with mild but verified cases of illness. All that’s above the waterline. And below the water line we have people who have contracted the virus. They actually are infected with COVID-19, but have such mild symptoms, or even no symptoms at all, that they don’t actually ever get tested. It’s not even on their radar that they might have COVID.

At this point, we just don’t know how much of that iceberg is under the water. 

If we were to add all those “below the waterline” cases into the denominator, the mortality rate would be much less than the director-general’s 3.4 per cent.

At this point, the most accurate assessment of COVID-19 mortality rates likely comes from the Diamond Princess cruise ship so famously quarantined off Japan. We’re able to get accurate numbers there because the people were stuck on the boat, making it easy to assess who was infected, and what happened to them. As recounted in this great Slate article by Jeremy Samuel Faust, a Boston emergency medicine physician and Harvard Medical School instructor, six people died on the boat out of 705 people who tested positive for the virus. That yields a mortality rate of 0.85 per cent.

Many experts predict that the actual, scientifically assessed mortality rate may even approach the 0.1% to 0.2% numbers that we see from variants of influenza A and B.

So yes, the WHO director’s stat was concerning—but these are early days, and we expect that mortality rate statistic to fall. COVID-19 is a lot less deadly than most of the numbers suggest.  

Dr. Peter Nord is Medcan’s Chief Medical Officer.