Editors Note: This post is reshared from Quora because it helps us understand how we calculate CFR (case fatality rates) from epidemics like #COVID-19. Link to the original post is here: https://qr.ae/ppY992
Post Author: C Stuart Hardwick, Award-Winning Scifi Author, Analog regular
Guess what? Epidemiologists aren’t dumb.
Mortality rate never includes undiagnosed cases—of any disease, at least not in the way you mean.
The mortality rate is a way of comparing the severity of known diseases. Roughly speaking, it measures the percentage of known or presumed cases that lead to death. Known or presumed cases is a number based on the consensus judgment of trained epidemiologists. There is no other way to get it because we do not have any way to routinely test even a random sampling of the population at large for any disease.
The Spanish Flu of 1918 is estimated to have had a mortality rate of 2.5%, but not one single diagnosis was made based on a lab test looking for viral DNA or circulating antigens—those technologies did not exist.
Today, they do, but they are still only one tool in the epidemiological toolkit—for the simple reason that such tests are necessarily virus-specific, and can never be made ready fast enough or in large enough numbers to test an entire population during an emerging outbreak.
Instead, known or presumed cases are identified first by looking at the clinical presentation—symptoms, then by performing tests. When available, the two are combined with sampling parts of the population and with estimating disease burden throughout the community as a whole.
But the numbers are always estimates, and can never be anything else, and the methods used by epidemiologists are often confounded by local conditions, or even by happenstance.
As I write this, numbers released by the Iranian government imply a terrifyingly high 11% mortality rate for SARS-CoV-2. However, we have good reason to believe the Iranian government is attempting to hide the extent of the outbreak there, and deaths are harder to hide than infections. This explains the anomalously high mortality rate, as reported.
On the other hand, data from China tells an interesting story. As of February 20th, China is reporting mortality rates of 5.8% in Wuhan and only 0.7% other areas.
Does that imply the COVID-19 is not a big deal? No. Definitely not.
First, .7% mortality is seven times worse than the average for seasonal flu, and the novel coronavirus is about twice as contagious. So going from that number alone, you can expect an uncontrolled COVID-19 outbreak in the US to kill three-quarters of a million people—except contagion doesn’t scale linearly, so the number would actually be higher than that.
Second, that low rate outside Hubei province is in a population already on high alert, taking stentorian measures to contain the outbreak, and in which cases can be quickly identified and effectively treated, including with anti-viral medications. The initial fatality rate in Wuhan was a whopping 17.3%! The kind of intensive treatment needed to keep mortality rates that low just cannot be maintained if the virus is allowed to sicken millions all at the same time.
“Coronavirus probably isn’t very serious” is precisely the WRONG conclusion to draw from this data. The correct conclusion is, “Coronavirus is very, very serious, but if we can slow it down enough, we can prevent it from becoming a catastrophe.”
How do we do that?
Wash hands correctly. Learn how to wear a mask properly. Comply with control orders from health officials. And just in case, designate a room in your home as an isolated sick ward, and give some thought to how to provide care while minimizing the spread of virus particles through physical contact and exhaled droplets.
It’s looking increasingly unlikely that this virus can be contained, but the more we can slow it down, the better the care we can provide to those who catch it. That’s as true in our individual homes as it is in nations.
And if we slow it down enough, we can stop it with a vaccine.
