One simple way to estimate the mortality rate of COVID-19 is to divide total deaths by the number of confirmed cases. Here is what that has looked like over time:
I could go into all the reasons why this isn’t a perfect metric but just because it has flaws doesn’t mean it’s not useful. There is a lot of information in this time series.
Early on, the death rate was fairly low. That was the same case with SARS, which eventually rose to about a 10% mortality rate. So, the rise in the chart wasn’t a surprise. What is different here though is the current downtrend. That didn’t happen with SARS.
For simplicity of analysis, let’s assume that only one factor is driving the change in this time series (instead of real life where it is some complicated combination of many factors). For the ratio to fall, one of two things has to happen:
The relative rate of deaths decreases
The relative number of confirmed cases increases
That’s it. That’s math. There is a numerator and denominator. So, let’s consider each case.
If the relative rate of deaths is decreasing, it could potentially mean that we are getting better at treating this virus. That makes sense. Early on, doctors had only a vague sense of what things helped or not. Now, they know better - one example is having patients lay on their stomachs as opposed to their backs to get better oxygen intake. Take one tiny fix like that with a bunch of other little tips learned through trial and error and maybe it is enough to save a small portion of lives. Those individual lives add up!
On the less optimistic side of things, perhaps the number of deaths is now falling because the disease is spreading in countries that have less reporting transparency. While Europe and the United States aren’t perfect in their reporting of numbers, there may be other countries that are even more reluctant to release their true death rates. There is some evidence of this with the wide disparity of death rates amongst individual countries.
On the bottom of the ratio we have confirmed cases. This is definitely rising as the amount of testing worldwide has expanded. A lot of people that may not have gotten confirmed because they were not deathly ill are now being counted. I think back to January, February, and even early March in the United States. There were a lot of sick people and they couldn’t get tested! So, they didn’t count.
The less optimistic side of this one is that the virus outbreak is accelerating. Since death lags infection, the rise in cases relative to deaths may mean that there are just a lot more people getting sick now. The deaths could come later.
The truth of all this is that reality is some messy combination of all those things above and a bunch of other things not mentioned. I’m sure you can think of a few pet theories as well that are just as valid potential factors.
But, at the end of the day, I think a falling death rate is a good sign. While the drivers may not be all good, the upside to a falling death rate is more than the downsides here. At the very least, it signals a new chapter on this strange strange virus journey.