0

I have seen multiple sources claiming that the fatality rate of the 2019-nCoV virus is around 2% to 10%.

At the same time, as of February 3rd 2020, 362 patients were reported to have died from the virus, while only 523 have recovered. [1]

Simple maths show that this would result in a fatality rate of 40.9%.

So where is this great difference in values coming from?

Patches
  • 8,162
  • 8
  • 44
  • 58
Markus Appel
  • 141
  • 2
  • 3
    Better question for a site like Biology.SE – rjzii Feb 03 '20 at 13:24
  • 2
    There is a very similar question on MedicalSciences: https://medicalsciences.stackexchange.com/questions/20970/is-the-mortality-rate-of-2019-ncov-41 It's fine there IMO, but I would not recommend posting it to Biology. – Bryan Krause Feb 03 '20 at 23:47
  • 1
    A simple assumption that all who've contracted the illness have been put in either the 362 and 523 numbers. It's a false assumption. Only the very ill are counted in those numbers; only the ones shown up at the hospital. –  Mar 12 '20 at 20:06
  • Simply put, "died" and "recovered" aren't the only possible states for those infected; you've neglected to include "currently alive and infected", which covers far more people (currently) than the other two categories. – anaximander Mar 16 '20 at 15:57

1 Answers1

6

Take a look at this live map of the virus. The calculation is based on the number of deaths, divided by the number of confirmed cases (multiplied by 100 to get it as a percent%).

As of February 3rd 2020, that would be 362 / 17405 * 100 ~= 2.08% fatality rate.

Laurel
  • 30,040
  • 9
  • 132
  • 118
Entropy
  • 2,344
  • 2
  • 11
  • 21
  • 3
    If that's true, the fatality rate is a really bad measurement for how dangerous a disease is. If a disease (like 2019-nCov) takes longer to show symptoms (and possibly kill you) the fatality rate would be low because many more get infected before deaths occur, allthough the fatality *chance* can be almost 50%. Correct? – Markus Appel Feb 03 '20 at 10:03
  • 15
    @MarkusAppel The issue is that the 2019-nCoV has only been out for about a month, so the number of people who are either dead or who recovered from the virus are much lower than for the number of ongoing cases. The fatality rate is going to undergo pretty drastic changes until things settle down. With time, the number of ongoing cases is outshined by the combined number of deaths + recovered cases. For short: yes, it's a mediocre way to measure fatalities early on, but it becomes more accurate in the long term. – Entropy Feb 03 '20 at 10:40
  • 16
    There is also the problem with underreporting. Mild cases might get diagnosed as a regular seasonal flu or not get diagnosed at all. – Philipp Feb 03 '20 at 13:08
  • 2
    FWIW: it seems that people who are going to die, do so in the first 5-7 days, even though recovery takes longer [citation-needed desperately] – Oddthinking Mar 11 '20 at 07:17
  • 4
    @Oddthinking "From illness onset, the median time to discharge was 22 days, and the average time to death was 18.5 days." https://eurekalert.org/pub_releases/2020-03/tl-pss030920.php – DavePhD Mar 11 '20 at 14:30
  • @DavePhD: Thank you. I was under the misapprehension that median time to death was earlier. Your reference shows I was wrong. – Oddthinking Mar 11 '20 at 16:23