COVID-19 is a respiratory virus that is spread by droplet spread and body fluids. It has an R0 (defined as the number of other people an infected person will subsequently infect) of at least 2–3 (This number was published early on by the WHO based on data from China which, IMHO is extremely flawed given the experience now unfolding around the world and especially in Europe). My personal opinion is that the R0 is much higher…

Much has been made in the media of the apparently low mortality with this virus… Well, that is false! Again the false data from China gave that impression and the experience elsewhere belies that!

Part of the problem is the use of the case fatality rate as a metric of mortality… Doctors and epidemiologists, because we know exactly what its definition is know what the case fatality rate means – and, also what it doesn’t mean! Clearly, the media, politicians, and the general public have little comprehension here and therefore use the data well out of context…

The first misconception that needs to be dealt with is that the case fatality rate is written in stone… It is NOT! As soon as either the number of cases reported or the number of deaths increases then the previous calculation of the case fatality rate becomes invalid and the calculation needs to be with the updated data. In fact, this is also true for other measures of mortality (and we will discuss these measures shortly)…

To understand what the case fatality rate is, and is NOT, telling one, one needs to understand how the case fatality rate is defined:

Now this definition is fixed and can be found in any publication dealing with public health or epidemiology.

It provides a running rate of mortality as a percentage of total cases. Nothing more and nothing less… It is an EXCEPTIONALLY poor metric of real mortality especially early on in an epidemic for reasons that will become clear…

The big problem with the case fatality rate as defined above is that it mixes active cases with closed cases to provide a fatality rate. A closed case is a case with an outcome – either death or recovery. It really makes no sense to calculate a mortality rate and include cases that are still active! After all, one does not know the outcome of those cases yet! The result of this is that the real mortality rate is vastly underestimated by this calculation because all the active cases are included as part of the denominator…

However, if one calculates mortality using ONLY the closed cases then a very different picture emerges…

Case fatality rate (at 17/03/2020) = 7173 deaths / 182683 as a percentage is 3.93%

Mortality rate calculated using

ONLY closed cases (at 17/02/2020) = 7173 / (7173 + 79883) as a percentage is 8.24%

Calculating mortality using only closed cases, especially early on in an epidemic, where the proportion of active cases to total cases is high, is a much more accurate measure of true mortality…

Astute readers may also have realized a couple of things about the case fatality rate and mortality rate according to closed cases:

Firstly, as the epidemic progresses the two numbers trend toward each other; and

Secondly, once the epidemic is over (all cases now closed), both calculations will yield EXACTLY the same result!

Now, I believe that the data from China is very flawed! The reason I say this is that the apparent pattern of this epidemic is SO DIFFERENT comparing China with the rest of the world that is difficult to believe that one is looking at data from the SAME epidemic!

Now, in order to stay on topic with the question (and I will shortly make the connection between all this stuff about different ways of calculating mortality with the essence of the question) I will not go off on a tangent to explain all the reasons why I believe the data from China is fundamentally flawed. Rather I will demonstrate, now that the rest of the world has both more cases and more deaths than China, the extraordinary differences between China and the rest of the world with respect to this epidemic…

## Answer ( 1 )

It is very dangerous!

COVID-19 is a respiratory virus that is spread by droplet spread and body fluids. It has an R0 (defined as the number of other people an infected person will subsequently infect) of at least 2–3 (This number was published early on by the WHO based on data from China which, IMHO is extremely flawed given the experience now unfolding around the world and especially in Europe). My personal opinion is that the R0 is much higher…

Much has been made in the media of the apparently low mortality with this virus… Well, that is false! Again the false data from China gave that impression and the experience elsewhere belies that!

Part of the problem is the use of the case fatality rate as a metric of mortality… Doctors and epidemiologists, because we know exactly what its definition is know what the case fatality rate means – and, also what it doesn’t mean! Clearly, the media, politicians, and the general public have little comprehension here and therefore use the data well out of context…

The first misconception that needs to be dealt with is that the case fatality rate is written in stone… It is NOT! As soon as either the number of cases reported or the number of deaths increases then the previous calculation of the case fatality rate becomes invalid and the calculation needs to be with the updated data. In fact, this is also true for other measures of mortality (and we will discuss these measures shortly)…

To understand what the case fatality rate is, and is NOT, telling one, one needs to understand how the case fatality rate is defined:

case fatality rate = reported deaths (to date) / reported cases (to date)

Now this definition is fixed and can be found in any publication dealing with public health or epidemiology.

It provides a running rate of mortality as a percentage of total cases. Nothing more and nothing less… It is an EXCEPTIONALLY poor metric of real mortality especially early on in an epidemic for reasons that will become clear…

The big problem with the case fatality rate as defined above is that it mixes active cases with closed cases to provide a fatality rate. A closed case is a case with an outcome – either death or recovery. It really makes no sense to calculate a mortality rate and include cases that are still active! After all, one does not know the outcome of those cases yet! The result of this is that the real mortality rate is vastly underestimated by this calculation because all the active cases are included as part of the denominator…

However, if one calculates mortality using ONLY the closed cases then a very different picture emerges…

Case fatality rate (at 17/03/2020) = 7173 deaths / 182683 as a percentage is 3.93%

Mortality rate calculated using

ONLY closed cases (at 17/02/2020) = 7173 / (7173 + 79883) as a percentage is 8.24%

Calculating mortality using only closed cases, especially early on in an epidemic, where the proportion of active cases to total cases is high, is a much more accurate measure of true mortality…

Astute readers may also have realized a couple of things about the case fatality rate and mortality rate according to closed cases:

Firstly, as the epidemic progresses the two numbers trend toward each other; and

Secondly, once the epidemic is over (all cases now closed), both calculations will yield EXACTLY the same result!

Now, I believe that the data from China is very flawed! The reason I say this is that the apparent pattern of this epidemic is SO DIFFERENT comparing China with the rest of the world that is difficult to believe that one is looking at data from the SAME epidemic!

Now, in order to stay on topic with the question (and I will shortly make the connection between all this stuff about different ways of calculating mortality with the essence of the question) I will not go off on a tangent to explain all the reasons why I believe the data from China is fundamentally flawed. Rather I will demonstrate, now that the rest of the world has both more cases and more deaths than China, the extraordinary differences between China and the rest of the world with respect to this epidemic…