Number of Deaths

[vegster]

@HWT. Any thoughts on how obesity is affecting mortality rates among Covid 19 cases? Not very scientific (actually not scientific at all) but a large number of the younger individuals who have died with no apparent underlying health issues appear from photos to be on the large size?


Sent from my iPad using Tapatalk

I read somewhere that factors that increase the risk of being hospitalised with serious symptoms are...

Being Male, of an Asian ethnicity, 50 + years old, underlying health problems, Smoking and obesity.

 

[Harry Wilson's tackle]

@HWT. Any thoughts on how obesity is affecting mortality rates among Covid 19 cases? Not very scientific (actually not scientific at all) but a large number of the younger individuals who have died with no apparent underlying health issues appear from photos to be on the large size?


Sent from my iPad using Tapatalk

Hi - just spotted your Q by chance - you need to type my silly user name in full to ensure my attention.

That's a very good spot. It had crossed my mind there may be something like this in play. Unfortunately I can't make any definitive comment. That said, it is recognised that obesity can cause shortness of breath due to the pressure and extra weight exerted by the blubber. Given this is a respiratory infection there is a possibility that obesity may increase risk of death from COVA as a consequence.

This is one of the possible co-morbidities I wrote about in another post. We need to know what increases the probability of death so we can protect the vulnerable.

The way to look at it is like this. We all have a probability (P) of dying if we get COVA. If your P is zero you will live. If your P is 1 you will die. If your P is somewhere in between (as it will be for all of us) you have a risk and you may live or die. If the P is 0.5 you have a 50:50 chance.

Bear in mind when reading what I'm going to say next, we are not sure what constitutes a comorbidity yet, apart from old age and immunosuppression, and we certainly don't know how much the P value is affected by each....

So....if you are aged 30, say, and have no comorbidities (fit, healthy, nonsmoker) let's say you P value is a very low, 0.01. This means you have only a 1% chance of dying if you get the virus. It also means that if there are 100 people like you, and all 100 get the virus, statistically one of you will die. If you sample 100 people like you with the virus, you may find no deaths. Do it with another hundred there may be 3. Do it again and again and on average there will be one death in each 100 with the virus sampled.

Let's imagine you are a heavy smoker. This may increase your P value to 0.05. So you now have a 5% chance of dying. And on average, 5 in every hundred people like you will die if they get the virus. I don't know what the actual P is if you are a fit 30 year old smoker, I am just using this as an illustration.

So maybe if you are 30 and obese your P is 0.05 too. If you are 60, obese and a smoker you P value may be much higher. Maybe 0.5. A 50% chance of dying.

That's how it works - your risk profile will determine the liklihood of a good and a bad outcome if you get the virus.

It also determines your risk of other outcomes as well. And in the long run, we are all dead.

 

[Harry Wilson's tackle]

And there's me hoping that you would say 20%.
I got stats o level once, it's not much good here is it?

Well it is....but we are all in the same boat in that we don't yet have enough data to be able to work out how much date we need

But we do have enough data to see that some of the data cannot be correct

 

[Deportivo Seagull]

Well it is....but we are all in the same boat in that we don't yet have enough data to be able to work out how much date we need

But we do have enough data to see that some of the data cannot be correct

And that’s problem with modelling at the moment. If your data set is wrong then the output of the model will be wrong, rather than the model itself...
Crap data in will ultimately lead to crap set of results.


Sent from my iPad using Tapatalk

 

[Weststander]

So we finally have an answer to this from the ONS:

"The provisional number of deaths registered in England and Wales in the week ending 3 April 2020 (Week 14) was 16,387; this represents an increase of 5,246 deaths registered compared with the previous week (Week 13) and 6,082 more than the five-year average"

"The 16,387 deaths that were registered in England and Wales during the week ending 3 April is the highest weekly total since we started compiling weekly deaths data in 2005.”

Out of the 6,082 deaths, 4,120 involved COVID-19 and 90.2% (3,716) occurred in hospitals and 404 in care homes, private homes and hospices.

To the 4,120, deaths involving COVID-19 in Scotland and NI need to be added.

 

[Swansman]

I just checked the BBC pages and they are reporting the numbers of cases and deaths in the main table for April 9, values with are of course lower than the ones they were showing yesterday and which I posted on here. Don't know what that's all about. And I did click 'refresh'.

In Sweden today when asked about a lot of technicalities of how they determine reason of death and who goes on the covid-19 death list and who doesnt they said that there's been a few cases where the attending physician had stated "covid-19" as the reason of death (based on symtoms) despite no test being done and that they would remove a very small number of deaths from the statistics because of this. Maybe something similar going on?

 

[Guinness Boy]

Hi - just spotted your Q by chance - you need to type my silly user name in full to ensure my attention.

That's a very good spot. It had crossed my mind there may be something like this in play. Unfortunately I can't make any definitive comment. That said, it is recognised that obesity can cause shortness of breath due to the pressure and extra weight exerted by the blubber. Given this is a respiratory infection there is a possibility that obesity may increase risk of death from COVA as a consequence.

This is one of the possible co-morbidities I wrote about in another post. We need to know what increases the probability of death so we can protect the vulnerable.

The way to look at it is like this. We all have a probability (P) of dying if we get COVA. If your P is zero you will live. If your P is 1 you will die. If your P is somewhere in between (as it will be for all of us) you have a risk and you may live or die. If the P is 0.5 you have a 50:50 chance.

Bear in mind when reading what I'm going to say next, we are not sure what constitutes a comorbidity yet, apart from old age and immunosuppression, and we certainly don't know how much the P value is affected by each....

So....if you are aged 30, say, and have no comorbidities (fit, healthy, nonsmoker) let's say you P value is a very low, 0.01. This means you have only a 1% chance of dying if you get the virus. It also means that if there are 100 people like you, and all 100 get the virus, statistically one of you will die. If you sample 100 people like you with the virus, you may find no deaths. Do it with another hundred there may be 3. Do it again and again and on average there will be one death in each 100 with the virus sampled.

Let's imagine you are a heavy smoker. This may increase your P value to 0.05. So you now have a 5% chance of dying. And on average, 5 in every hundred people like you will die if they get the virus. I don't know what the actual P is if you are a fit 30 year old smoker, I am just using this as an illustration.

So maybe if you are 30 and obese your P is 0.05 too. If you are 60, obese and a smoker you P value may be much higher. Maybe 0.5. A 50% chance of dying.

That's how it works - your risk profile will determine the liklihood of a good and a bad outcome if you get the virus.

It also determines your risk of other outcomes as well. And in the long run, we are all dead.

Having just replied to you on another thread, I was reminded of this thread and a thought that occurred today after an interview I saw on the death stats.

*Disclaimer - this thought is in purely statistical terms and in no way a comment on the health or likelihood of survival of individuals*

The interview stated that one in eleven deaths had no obvious underlying health conditions but that most deaths had two or more, with the most linked being heart disease and the next, pre-existing respiratory illness. And here's the bit where I felt the need to add the disclaimer, because it's a grim question.

As the number of people with multiple health issues in the "at risk" age and gender category decreases (because CV19 is killing them) does the ratio number increase from 8 (because there are fewer people in the at risk group and those in it should be shielding) or is the b'stard thing simply going to come after the rest of us?

If it stays at 8 while the number of tests carried out increases and the at risk group decreases (or shields) I assume that means a fairly devastating mutation?

Sorry, I must be in a morbid frame of mind today.

 

[rogersix]

Having just replied to you on another thread, I was reminded of this thread and a thought that occurred today after an interview I saw on the death stats.

*Disclaimer - this thought is in purely statistical terms and in no way a comment on the health or likelihood of survival of individuals*

The interview stated that one in eleven deaths had no obvious underlying health conditions but that most deaths had two or more, with the most linked being heart disease and the next, pre-existing respiratory illness. And here's the bit where I felt the need to add the disclaimer, because it's a grim question.

As the number of people with multiple health issues in the "at risk" age and gender category decreases (because CV19 is killing them) does the ratio number increase from 8 (because there are fewer people in the at risk group and those in it should be shielding) or is the b'stard thing simply going to come after the rest of us?

If it stays at 8 while the number of tests carried out increases and the at risk group decreases (or shields) I assume that means a fairly devastating mutation?

Sorry, I must be in a morbid frame of mind today.

must be "lowest hanging fruit" initialy, and then it becomes less virulent, assuming no mutation

 

[Harry Wilson's tackle]

Having just replied to you on another thread, I was reminded of this thread and a thought that occurred today after an interview I saw on the death stats.

*Disclaimer - this thought is in purely statistical terms and in no way a comment on the health or likelihood of survival of individuals*

The interview stated that one in eleven deaths had no obvious underlying health conditions but that most deaths had two or more, with the most linked being heart disease and the next, pre-existing respiratory illness. And here's the bit where I felt the need to add the disclaimer, because it's a grim question.

As the number of people with multiple health issues in the "at risk" age and gender category decreases (because CV19 is killing them) does the ratio number increase from 8 (because there are fewer people in the at risk group and those in it should be shielding) or is the b'stard thing simply going to come after the rest of us?

If it stays at 8 while the number of tests carried out increases and the at risk group decreases (or shields) I assume that means a fairly devastating mutation?

Sorry, I must be in a morbid frame of mind today.

Good question.

I don't think the number of tests will affect the ration of cases to deaths. These are independent variables (which is why I have been interested in them). If the high risk groups shield the number of cases per deaths should decrease. I don't think the number of deaths is making inroads in the total number of those at risk yet (curiously the number of cases is still not very high - and I don't buy the idea that with 100,000 recorded cases there are a million or more other folk who have had the virus and been asymptomatic). So if our cases per deaths staus the same I don't think this means the virus has mutated. What happens is the infection needs to almost disappear (as ordinary flu does in May/June) and then the tiny numbers of mutated viruses lurk around and then spread about when conditions (Autumn) triggers them to behave differently. So there is no reason to fret yet. Or at all, with respect to your concerns, I don't think.

 

[Harry Wilson's tackle]

Here are todays calculations and 'death rates' (total deaths per total number of cases) continue to increase in country after country which, as I explained earlier mean that new cases falling in most countries, especially those with the most cases, meaning all these countries are at or over the (first) plateau. More comments and analysis later....

 

[Kalimantan Gull]

So we finally have an answer to this from the ONS:

"The provisional number of deaths registered in England and Wales in the week ending 3 April 2020 (Week 14) was 16,387; this represents an increase of 5,246 deaths registered compared with the previous week (Week 13) and 6,082 more than the five-year average"

"The 16,387 deaths that were registered in England and Wales during the week ending 3 April is the highest weekly total since we started compiling weekly deaths data in 2005.”

Thank you, so despite many protestations from a single poster, we were correct in assuming that the number would increase but some of the deaths, albeit extremely sad and regrettable as any death is, would be within the Govts numbers.

What a very stubborn way to stick to your original hypothesis despite it being shown to be totally wrong by the poster you're quoting

Deaths for the week in question increased by 6000 on the 5-yr average even though only about 3700 were confirmed covid related, with the obvious inference that the remainder of the increase were covid caused as well.

In other countries we have seen funeral services in Jakarta increase from a monthly range of 2300-2800 for the past two years to 4500 in March this year, and 6700 people dying in Ecuador's Guayas province during the first two weeks of April as compared to a normal average of 1000 deaths for the region in this period.

So quite categorically these covid deaths are not just the same people dying who would have died anyway.

 

[Harry Wilson's tackle]

Latest cases per deaths calculation. Once again the difference (in yellow) betyween cases per deaths today versus 2 days ago (too few countries update daily at the same time so doing this over 2 days makes better sense) reveals nearly all values are positive, meaning that the plateauing appears to be continuing. I can't be arsed to go down this list any further as I am having to move the countries around as the source data keeps being shuffled by John Hopkins

 

[Weststander]

Latest cases per deaths calculation. Once again the difference (in yellow) betyween cases per deaths today versus 2 days ago (too few countries update daily at the same time so doing this over 2 days makes better sense) reveals nearly all values are positive, meaning that the plateauing appears to be continuing. I can't be arsed to go down this list any further as I am having to move the countries around as the source data keeps being shuffled by John Hopkins

View attachment 122389

Just looked at JHU online, now 34,566 confirmed US deaths. I realise you’re using data at an earlier snapshot.

 

[Harry Wilson's tackle]

Just looked at JHU online, now 34,566 confirmed US deaths. I realise you’re using data at an earlier snapshot.

Yep. And so far it doesn't matter. Every peek I have taken is saying the same thing. If the trend stops in a week or so, as the increase in deaths slows down to match the slowdown in increase of cases I'll know I've cracked it.

Unfortunately what goes down can come back up again.

 

[Titanic]

Not sure if this is helpful... I guess it will vary depending on where you are on the curve, but this is based on the John Hopkins data... countries with >500 deaths by deaths per 100k population:

 

[Titanic]

Not sure if this is helpful... I guess it will vary depending on where you are on the curve, but this is based on the John Hopkins data... countries with >500 deaths by deaths per 100k population:

Is there any reliable data on the "caused by Covid19" rather than "Covid19 positive but was very ill / old / would have probably died anyway" numbers of deaths in each country?

If not, is there reliable data on the current death rates compared to average rates over previous years?

 

[Tim Over Whelmed]

What a very stubborn way to stick to your original hypothesis despite it being shown to be totally wrong by the poster you're quoting

Deaths for the week in question increased by 6000 on the 5-yr average even though only about 3700 were confirmed covid related, with the obvious inference that the remainder of the increase were covid caused as well.

In other countries we have seen funeral services in Jakarta increase from a monthly range of 2300-2800 for the past two years to 4500 in March this year, and 6700 people dying in Ecuador's Guayas province during the first two weeks of April as compared to a normal average of 1000 deaths for the region in this period.

So quite categorically these covid deaths are not just the same people dying who would have died anyway.

If you re-read the original post you'll read that I wasn't disputing that I was saying that unfortunately some people do die at this time of year and not ALL deaths could be attributed to COVID-19.

 

[Machiavelli]

Is there any reliable data on the "caused by Covid19" rather than "Covid19 positive but was very ill / old / would have probably died anyway" numbers of deaths in each country?

If not, is there reliable data on the current death rates compared to average rates over previous years?

Prof Neil Ferguson said that 2/3 of deaths would have happened soon anyway.
As to Q2, the ONS figures released this week will point to that. IIRC, it was quite a pronounced increase.

 

[WATFORD zero]

I'm still not convinced that we are getting accurate, comparable figures across Europe. Belgium are insisting that their fatality statistics include all Covid 19 deaths across Hospitals, Care Homes and Private residences, including deaths where the patient had coronavirus symptoms but had not been tested. They are saying that this is why their figures are so much higher.

https://www.dailymail.co.uk/news/article-8229275/Is-Europes-coronavirus-death-toll-DOUBLE-official-figures.html

There is also data coming from Italy, Spain, Ireland, Belgium and France that show Hospital deaths only account for about half of the fatality figures with 42-57% coming from care homes.

https://www.independent.co.uk/news/health/coronavirus-deaths-care-homes-cases-uk-eu-italy-spain-ireland-a9463846.html

Meanwhile, we have the UK currently saying that they are only counting hospital deaths because that is 'how everyone else is doing it'. And today come reports of between 6,000 and 7,500 fatalities (depending on source) in Care Homes alone with no indications whatsoever of home deaths.

https://www.ft.com/content/9d6b46e2-55f4-4de1-ba21-f1ab9f14bcbf

https://www.itv.com/news/2020-04-18/7-500-feared-to-have-died-with-coronavirus-in-care-homes/

Now I know I am naturally quite cynical, but I can't help starting to get the feeling these statistics released are being 'managed'

 

[Green Cross Code Man]

I'm still not convinced that we are getting accurate, comparable figures across Europe. Belgium are insisting that their fatality statistics include all Covid 19 deaths across Hospitals, Care Homes and Private residences, including deaths where the patient had coronavirus symptoms but had not been tested. They are saying that this is why their figures are so much higher.

https://www.dailymail.co.uk/news/article-8229275/Is-Europes-coronavirus-death-toll-DOUBLE-official-figures.html

There is also data coming from Italy, Spain, Ireland, Belgium and France that show Hospital deaths only account for about half of the fatality figures with 42-57% coming from care homes.

https://www.independent.co.uk/news/health/coronavirus-deaths-care-homes-cases-uk-eu-italy-spain-ireland-a9463846.html

Meanwhile, we have the UK currently saying that they are only counting hospital deaths because that is 'how everyone else is doing it'. And today come reports of between 6,000 and 7,500 fatalities (depending on source) in Care Homes alone with no indications whatsoever of home deaths.

https://www.ft.com/content/9d6b46e2-55f4-4de1-ba21-f1ab9f14bcbf

https://www.itv.com/news/2020-04-18/7-500-feared-to-have-died-with-coronavirus-in-care-homes/

Now I know I am naturally quite cynical, but I can't help starting to get the feeling these statistics released are being 'managed'

I do think the majority of countries only count hospital deaths in their running total. It reflects the haphazard nature that the deaths are recorded in. The Spanish figure quoted in that article for instance does not make it clear whether their care home figures had been added to the total or whether the 57% figure was equivalent to 57% of their current (as of 8th April) number of hospital deaths. For quicker analysis, I can see why governments use the data this way, and for comparison, it is helpful that we all use a similar method. However as Harry Wilson has stated, the important thing is that each country's recording process remains the same to ensure consistency.