April 17 , 2020
How Deadly is COVID-19?
A very important question for doctors and for governments to understand is exactly how deadly the pandemic virus we are facing actually is. This would seem to be a simple question to answer. Just take the number of deaths and divide by the number of infections, and that tells you how deadly the virus is. But it is actually far more complicated.
There are real problems in trying to figure out both of those numbers, and in how well the data can be transferred from one country to another based on demographics and health status. On the front end, how many people are being killed by COVID-19?
There are the obvious cases, deaths in patients with fever and cough and who test positive for COVID by PCR DNA testing. But DNA testing is falsely negative about 10% of the time, even though it is fairly evident to the doctors the patient does in fact have COVID. Those deaths should also count. There are also people who die without being tested, either because the health care system is unable to do testing, or because the patients die at home and are never tested. Both in Italy and in New York City there have been several hundred extra deaths at home daily during this outbreak, those are almost certainly COVID infections in frail people who rapidly progress from mildly ill to dead without calling for help.
Finally, there are excess deaths caused by a healthcare system that is overloaded and unable to provide adequate care to patients with other life-threatening conditions. This has been a problem in Italy and perhaps Spain but not yet in the US.
The second number, the total number of infections, is also very hard to pin down. We have not done widespread testing of the population so we cannot get a good sense of how many total infections there are. Because of a shortage of test kits, hospitals are only testing seriously ill people, everyone else is told to go home and rest. We also know from testing that has been done widely in some locations that many patients infected with COVID-19 have little or no symptoms. But do they make up 50% of infections or 95% of infections? That we don’t know. The higher the number the better, actually, because it means the virus is less deadly than the worst case scenarios.
As of this writing the US had 615,000 diagnosed cases and 26,000 deaths. But that death total is going to rise even if we completely stopped all new infections. It takes about two weeks from exposure to get sick enough to end up in the hospital and another 3-7 days to end up in ICU. So there are several hundred thousand people who are currently infected that are still at risk of getting much worse and dying. If we make allowances for that, and for the undercounting of COVID deaths at home, it would be reasonable to assume that we will end up with 40,000 deaths roughly. With 615,000 infections, that gives a fatality rate of almost 7%. But no serious observer thinks that is correct.
We also need to get a true number for the total infected in the US. Is that number 2 million or 4 million or 8 million? Or more? The only way to really find out is to do wide population sampling to test for previous infection. We can detect the fact that someone had COVID by the presence of anti-COVID antibodies in their blood, using the same technology that is used in home pregnancy tests. These tests are being rolled out at present, and we should get some data soon. The numbers in New York City for example will be quite revealing.
There have already been 10,000 deaths in New York, in a city with 9 million people. If 10% of New York residents have been infected that would mean the fatality rate is about 1%. But we know that the total number of deaths will be higher than that.
While more data will give us a much firmer estimate, my best guess is that the fatality rate of COVID in the US is about 1.5%. That may not seem like a lot, but that is catastrophic for a novel virus to which no one has immunity. If allowed to spread in the US unchecked, it could easily infect 100 million people, and kill 1.5 million. An average flu season results in 15,000 deaths. There is just no comparison.
Will the fatality rate for COVID be higher or lower in other nations? That depends on a number of factors. The US has a very sophisticated health system with plenty of ICU beds and ability to treat people on the edge of death. Third World countries do not have this. On the other hand, the lack of universal health insurance, and the limited nature of healthcare in rural America, will lead to excess deaths.
Another big factor is the demographics of each country. While the death rate is perhaps 1.5% in a large US city, most of those deaths are in people over the age of 65. The US, Europe, and East Asia have very large shares of elderly in their total population. For the US, 16% of the population is over 65, while in the Democratic Republic of the Congo, it is 2%. Italy has one of the oldest populations in the world, which contributed to its horrific death toll. We may find that the fatality rate of this virus is actually lower in poor countries, simply because their share of vulnerable elderly is much lower.
