Deaths from COVID -19 as a proportion of total deaths

[jimmydasaint]

I am putting this topic here as it does not fit anywhere else.

I have been puzzled over the few weeks, as a concerned citizen, why the number of deaths by COVID were not given as a proportion of total deaths.

I went to the Official Government website Gov.UK and downloaded figures to compare the number of deaths attributed to coronavirus in relation to normal deaths for this period.

This is not something, in the UK, that is mentioned by news outlets very often.  Sky News can be renamed  as  Coronavirus Updates now because they have gone into the reporting of this awful viral infection 24/7.

I wanted to check out relative figures of coronavirus infection in relation to total numbers of deaths from all causes at this time of year. 

After downloading the Excel files from the Government website, I got the following figures:

Total number of deaths from all causes in 2018 in England and Wales

    Jan-18    Feb-18    Mar-18    Apr-18    May-18    Jun-18    Jul-18    Aug-18      Sep-18     Oct-18     Nov-18     Dec-18
    64,020    49,087    51,131    46,383    42,685    39,679    40,621    40,071    37,013    44,311    43,834    41,430

Total number of deaths from all causes in 2019  in England and Wales

Jan-19	Feb-19	Mar-19	Apr-19	May-19	Jun-19	Jul-19	Aug-19	Sep-19	Oct-19	Nov-19	Dec-19
53,772	45,696	43,815	44,003	44,290	38,511	42,183	38,719	39,915	46,133	45,111	47,384

 

Total number of deaths from all causes in 2020 in England and Wales

 Jan-20               Feb-20          First 3 weeks of Mar-20

 62770                43,587          32559

Deaths from respiratory diseases in 2020 in England and Wales included in figures for total deaths

Jan-20                 Feb-20           First 3 weeks of Mar-20

10446                  6,300              4533

Deaths from COVID-19 mentioned on death certificates in England and Wales

Jan-20                 Feb-20           First 3 weeks of Mar-20

0                           0                      108

Because the figures are not complete for March 2020, when I accessed the website on 4 April 2020, I will take an average of all deaths from all caused in March 2018 and March 2019 as a representative figure for March 2020.  I calculate this to be 47,413  

If we take the current figures for COVID-19 deaths to be 4313 (as at 7.36 a.m. 5 April 2020), the percentage of deaths attributed to COVID-20 as an approximate percentage of all deaths is 9.097%. 

This is a fairly high estimate. but, in the interests of accurate reporting:

1.  I would wish our news agencies to drop the absolute numbers and give us the number of COVID-19 deaths as a percentage of deaths by all causes in the UK.

2. The Government should be duty bound to report the deaths by COVID-19 as a percentage of those (large pool of) people randomly tested and positive for COVID-19 infection.

3. We should continue social distancing and self-isolation in order to reduce viral transmission rates.

4. We should acknowledge, as scientists, that our world has changed profoundly, socially and economically, and that we are unlikely to go back to the "normality" of 2018 and 2019. 

 

 

Number of deaths 2018-2020.xls

Edited April 5, 2020 by jimmydasaint

[Dagl1]

Some things to note, but it would probably be more fair to take a longer time period for each month (multiple years), to calculate an estimated amount of deaths in March, there is a 7-8k difference between 2018 and 2019. A major risk factor for death is accumulated age (not only for corona virus), and ages are not uniformly distributed over all the years (there are not equally as many 40 as 60 years old etc.. 
Those things don't really change the focus of your post though, but I thought they would be good to note anyway.

I do question why you feel that we should drop absolute numbers in favour of percentage of deaths, other than that it represents population and life expectancy (sort of)?
Point two I don't really understand, should governments report deaths as a percentage of people tested, why? Assuming the minority of people is tested, this will lead to a much higher current morbidity than the actual numbers that we will be able to get after this is 'over' and do antibody testing.

 

[Danijel Gorupec]

Very good, jimmydasaint. I was trying to obtain total death figures for various countries (because covid death reports seems to be heavily affected by different methodologies used in different countries), but I was unsuccessful.

As I understood, in your estimated calculation you assumed that covid deaths will be just added to the 'usual' number of deaths. Am I right? This does not have to be. Some of the people who die, might die within months even without covid - in this case covid only shifts the death statistics some months earlier. On the other hand, if hospitals become overloaded, mortality form non-covid causes might also increase. So, the total death figure might be either less or more than the 'usual'+covid.

(Off topic: I am puzzled now, what happens in UK in January?)

[Robittybob1]

It is not the deaths to COVID 19 that are the problem.  It appears to be the large number of hospital cases requiring ventilation for weeks and then possibly months to recover.

[Strange]

Note that Italy and Spain (who now appear to be passing the peak) had something like 800 deaths a day; say 24,000 a month. The UK (and US) will soon exceed that.

People who die of because of Covid-19 are mainly those who end up in hospital and often in intensive care. The vast majority of the other deaths take place at home, in a normal hospital bed, at the scene of the accident, etc. Therefore Covid-19 puts a much greater strain on the health system.

1 hour ago, Danijel Gorupec said:

As I understood, in your estimated calculation you assumed that covid deaths will be just added to the 'usual' number of deaths. Am I right? This does not have to be. Some of the people who die, might die within months even without covid - in this case covid only shifts the death statistics some months earlier.

There is some interesting analysis of the age effects of both "normal" deaths and covid-19 here: https://medium.com/wintoncentre/how-much-normal-risk-does-covid-represent-4539118e1196

 

[jimmydasaint]

1 hour ago, Dagl1 said:

Some things to note, but it would probably be more fair to take a longer time period for each month (multiple years), to calculate an estimated amount of deaths in March, there is a 7-8k difference between 2018 and 2019. A major risk factor for death is accumulated age (not only for corona virus), and ages are not uniformly distributed over all the years (there are not equally as many 40 as 60 years old etc.. 
Those things don't really change the focus of your post though, but I thought they would be good to note anyway.

I do question why you feel that we should drop absolute numbers in favour of percentage of deaths, other than that it represents population and life expectancy (sort of)?
Point two I don't really understand, should governments report deaths as a percentage of people tested, why? Assuming the minority of people is tested, this will lead to a much higher current morbidity than the actual numbers that we will be able to get after this is 'over' and do antibody testing.

You are right that I should have taken a larger number of years for the average.  I considered a short term trend in numbers of deaths. Accumulated age is a large risk factor, I agree. 

IMHO, fear is being created (perhaps rightly so, but I am not a social scientist)  by broadcasting cumulative numbers of COVID-19 attributable deaths but we should have a news item which says something along the lines of:

"The cumulative total of COVID-19 deaths in the UK has now reached 4313 tragic deaths, in relation to the total number of deaths for March 2020 which was (estimated) at 47, 413 from all causes...."  IMO this would make people fearful, but careful, and not just fearful and ignorant. 

As far as the point about widespread testing, how can the Government talk about disease spread accurately without numbers?  The Prime Minister's speech with his two medical advisers was almost all guesswork. Educated guesswork but all guesswork because there was not enough testing throughout the population to predict a rate of spread of the viral infection.  Even now the percentages are based, for the most part, on patients who have reported in to hospital and have been then tested for coronavirus. 

1 hour ago, Danijel Gorupec said:

Very good, jimmydasaint. I was trying to obtain total death figures for various countries (because covid death reports seems to be heavily affected by different methodologies used in different countries), but I was unsuccessful.

As I understood, in your estimated calculation you assumed that covid deaths will be just added to the 'usual' number of deaths. Am I right? This does not have to be. Some of the people who die, might die within months even without covid - in this case covid only shifts the death statistics some months earlier. On the other hand, if hospitals become overloaded, mortality form non-covid causes might also increase. So, the total death figure might be either less or more than the 'usual'+covid.

(Off topic: I am puzzled now, what happens in UK in January?)

The Government are publishing COVID-19 figures in relation to the total, including the virus (IIRC). You are absolutely correct with the scenario you presented.  I think that hospitals are being overloaded and there is a possibility of increasing mortality from other causes.  Agreed! However, all I am looking for is a month to month comparison (whilst taking in the points you made) to give me a reference point as a citizen.

In January, we are in the height of winter, and if you look at the number of deaths from respiratory diseases in winter, it is quite high.  However, I did not check the numbers of people who had respiratory diseases in January in previous years.

26 minutes ago, Robittybob1 said:

It is not the deaths to COVID 19 that are the problem.  It appears to be the large number of hospital cases requiring ventilation for weeks and then possibly months to recover.

I thin COVI-19 is a problem for people of a certain age group (60 and above) or for those individuals who have underlying health conditions.  SARS is a nasty virus which can flatten a person with underlying health conditions. However, yes, you are correct, the National Health Service are overwhelmed and so people who have different life-threatening conditions may have less priority, than those with suspected, or actual COVID--19,.

29 minutes ago, Strange said:

Note that Italy and Spain (who now appear to be passing the peak) had something like 800 deaths a day; say 24,000 a month. The UK (and US) will soon exceed that.

People who die of because of Covid-19 are mainly those who end up in hospital and often in intensive care. The vast majority of the other deaths take place at home, in a normal hospital bed, at the scene of the accident, etc. Therefore Covid-19 puts a much greater strain on the health system.

I cannot disagree with you on most points. However as far as the death statistics, and we must remember that these are human tragedies and not just numbers, The Lancet published a model which was based on flu infection and for Singapore, but had important implications for the UK. The Lancet 23 Mar 2020

I am quoting the effects of self-isolation and lockdown in their article, assuming R0 (reproductive rate of viral spread, I assume) reducing daily numbers by 93%: (the emphasis in the last sentence of the quote is my own).

Quote

 

When R₀ was 2·0, the proportion of the Singaporean population infected under the baseline scenario increased to 19·3% (IQR 17·8–20·6), with a median cumulative number of infections on day 80 of 727 000 (IQR 670 000–776 000; figure 1, table). On day 80, quarantine resulted in a cumulative median of 130 000 cases (38 000–244 000), school closure 97 000 cases (14 000–219 000), workplace distancing 67 000 cases (11 000–145 000), and the combined intervention 50 000 cases (2000–143 000; figure 2, table). The maximum number of daily infections was 27 800 (IQR 27 300–28 000) for the baseline scenario, 11 000 (4100–18 600) for the quarantine scenario, 8400 (1900–17 000) for the school-closure scenario, 6100 (1500–12 000) for the workplace-distancing scenario, and 4900 (100–11 700) for the combined intervention (Figure 1, Figure 2). The combined approach resulted in the largest reduction in cases from baseline (93·0% reduction [IQR 81·5–99·7]).

 

 

For a more serious scenario, wit R0=2.5, the following conclusions are reached by th emodellers:

Quote

For the baseline scenario, when R₀ was 2·5, 32% (IQR 30·9–33·1) of the Singaporean population were infected, with a cumulative median of 1 207 000 cases (IQR 1 164 000–1 249 000) on day 80 (figure 1, table). At this level of transmission, quarantine resulted in a median of 520 000 cases (268 000–754 000), school closure 466 000 cases (175 000–728 000), workplace distancing 320 000 cases (116 000–558 000), and the combined intervention 258 000 cases (65 000–508 000; table; appendix p 10). The maximum number of daily infections was 42 800 (IQR 41 400–43 500) for the baseline scenario, 37 900 (25 900–41 800) for the quarantine scenario, 36 400 (18 100–41 400) for the school-closure scenario, 29 800 (12 500–40 500) for the workplace-distancing scenario, and 25 200 (7700–39 200) for the combined intervention (figure 1; appendix p 10). Compared with baseline, the combined approach resulted in the greatest reduction in cases (78·2% [IQR 59·0–94·4]).

The Lancet 23 March 2020

 

[Strange]

27 minutes ago, jimmydasaint said:

I cannot disagree with you on most points. However as far as the death statistics, and we must remember that these are human tragedies and not just numbers, The Lancet published a model which was based on flu infection and for Singapore, but had important implications for the UK. The Lancet 23 Mar 2020

I am quoting the effects of self-isolation and lockdown in their article, assuming R0 (reproductive rate of viral spread, I assume) reducing daily numbers by 93%: (the emphasis in the last sentence of the quote is my own).

I'm not quite sure what the disagreement is. To add a bit more detail, the UK is currently at about 500 deaths a day. Even if it had already reached the peak (it hasn't the death rate is still increasing exponentially) then that would be 15,000 deaths for the month, which is about 1/3rd of the average death rate. And, as noted, each of these deaths puts a much greater strain on the heath service.

Yes, isolation and quarantining will help reduce the height of the peak (but possibly not the total number of deaths). But the UK was late to do this and, initially, rather half-hearted.

We had scenes here of the army being called in to take coffins away because the funeral services were overloaded. So the UK could see what was coming and did nothing. Countries have had months to prepare and plan but still seem to be reacting as if it was all completely unexpected.

[StringJunky]

44 minutes ago, Strange said:

I'm not quite sure what the disagreement is. To add a bit more detail, the UK is currently at about 500 deaths a day. Even if it had already reached the peak (it hasn't the death rate is still increasing exponentially) then that would be 15,000 deaths for the month, which is about 1/3rd of the average death rate. And, as noted, each of these deaths puts a much greater strain on the heath service.

Yes, isolation and quarantining will help reduce the height of the peak (but possibly not the total number of deaths). But the UK was late to do this and, initially, rather half-hearted.

We had scenes here of the army being called in to take coffins away because the funeral services were overloaded. So the UK could see what was coming and did nothing. Countries have had months to prepare and plan but still seem to be reacting as if it was all completely unexpected.

But what has always been "over there" has never been, to the extent it is now, in our respective Western world backyards. If anything, this virus should inform us that no country is an island.

Edited April 5, 2020 by StringJunky

[Strange]

1 minute ago, StringJunky said:

But what has always been "over there" has never been, to the extent it is now, in our respective Western world backyards.

Agreed. I was in south-east Asia when one of the previous outbreaks happened. People were dying and countries were taking drastic action but it was treated as a bit of a joke in Europe.

The UK is about 2 or 3 weeks behind Italy. So they don't even seem to have learned from the experiences in their own backyard.

[iNow]

Lots of thoughtful feedback above. One additional point we ought to consider when framing the discussion in terms of deaths:

The social isolation and change in behavior were currently experiencing will lead to far higher drops in the long term death rate, and this is true for a great many things largely unrelated to covid.

Pollution is down right now. Air and water is cleaner... Manufacturing facilities are quit and their smoke stacks dormant... Traffic accidents are down right now. Crime and murder and shootings are down right now. War zones are quieter and fewer bullets are flying....

Countless other variables contributing to accelerated human death are all collectively down right now... We’re going to see a dramatic decrease in the deaths caused by these things... on the order of several hundreds of thousands of lives saved and extended... and that will eclipse deaths caused by this virus preoccupying us. 

Just food for thought. If Covid takes a million lives, but saves 10+ million bc of how we respond to it, then maybe that’s another reason to remind our minds to let go of and release the anxiety these daily infection and death count numbers often create within us.

[Strange]

58 minutes ago, iNow said:

Pollution is down right now. Air and water is cleaner... Manufacturing facilities are quit and their smoke stacks dormant... Traffic accidents are down right now. Crime and murder and shootings are down right now. War zones are quieter and fewer bullets are flying....

I have seen some anecdotal reports that ER/A&E attendances are much lower (fewer people out and about breaking legs and things).

[jimmydasaint]

1 hour ago, iNow said:

Lots of thoughtful feedback above. One additional point we ought to consider when framing the discussion in terms of deaths:

The social isolation and change in behavior were currently experiencing will lead to far higher drops in the long term death rate, and this is true for a great many things largely unrelated to covid.

Pollution is down right now. Air and water is cleaner... Manufacturing facilities are quit and their smoke stacks dormant... Traffic accidents are down right now. Crime and murder and shootings are down right now. War zones are quieter and fewer bullets are flying....

Countless other variables contributing to accelerated human death are all collectively down right now... We’re going to see a dramatic decrease in the deaths caused by these things... on the order of several hundreds of thousands of lives saved and extended... and that will eclipse deaths caused by this virus preoccupying us. 

Just food for thought. If Covid takes a million lives, but saves 10+ million bc of how we respond to it, then maybe that’s another reason to remind our minds to let go of and release the anxiety these daily infection and death count numbers often create within us.

That was the main point I tried to posit, suing the Singapore model from the Lancet.  Hope you and yours are OK iNow. 

[iNow]

10 minutes ago, jimmydasaint said:

Hope you and yours are OK iNow. 

Likewise ✌️

[Strange]

Nice animation showing how deaths from covid-19 compare with other causes (in the US) over time

https://public.flourish.studio/visualisation/1712761/

Another good source of data: https://ourworldindata.org/

[CharonY]

With regard to OP, I think it the different measures are useful for different purposes. The one outlined in OP would provide something like that relative health burden (as a count of deaths) though one would probably also want to see or at least try to account for secondary effects(e.g. whether there is also an increase in morbidity and mortality due to overburdening the health system).

In addition, as iNow pointed other causes of mortality are way down. OTOH this year's influenza seasons is shaping up to be pretty bad (estimated at ~60k in the US and it is not over yet). So it is probably quite difficult.

The crude number, on the other hand, are easy to generate but may not be easy for most folks to contextualize. Also, some countries have for some time not published all data (e.g. number of folks tested total) or unclear data (e.g. number f samples vs number of people tested). Often, because the reporting system was not centralized and clear.

However, it is convenient for epi-folks to grab those numbers that come out and plug them into their models (say, playing around with Rt). It is one of the rare reasons where they can make predictions and see them unfold in real time. Wuhan data was used quite a bit to this end and as StringJunky noted, it is sad that much of the World outside of Asia has not taken steps once it was clear what was happening. I also do see the of some news articles indicating that China was fudging their numbers as if that would excuse inactivity (which I find worrying, as it clearly indicates that folks in the US are not trying to address their own shortcomings). So at least in that sense rapid release of numbers could help to at least appear more transparent.

[MigL]

8 hours ago, Strange said:

I have seen some anecdotal reports that ER/A&E attendances are much lower

Most likely because that's where all the infected people are.

People used to go to the ER to get a band-aid applied.
( one of the dis-advantages of free health care )
These days, I wouldn't go near a hospital unless I was seriously injured.

[Strange]

18 minutes ago, MigL said:

Most likely because that's where all the infected people are.

People used to go to the ER to get a band-aid applied.
( one of the dis-advantages of free health care )
These days, I wouldn't go near a hospital unless I was seriously injured.

That was one reason given (by someone who thought they might have broken something but was too scared to go to the ER)