I've read several reports, in supposedly reliable sources of information, about the numbers of death to infection ratio and the methods of collection. There seems to be little agreement about the figures. Many of the reports are contradictory mainly in the deaths numbers. According to which report one believes some countries are attributing death to the virus only if the deceased was being treated for in hospital Covid 19. Meanwhile, a neighbouring country is reported as being meticulous about recording the true cause of death but the overworked doctors don't have time to be as meticulous as the politicians claim. Many deaths where Cov 19 is merely suspected are recorded as due to the virus. All that, of course, is on top of discrepancies caused by false positives and negatives in testing and governments deliberately distorting figures for political purposes. "We don't want to be blamed for allowing so many deaths so announce a reduced figure." Alternatively, "Make our deaths rate as high as possible and we can argue for more aid when everything calms down." Yes, I am an old cynic.
Can anyone tell me where I could find figures that are truly accurate and can be compared with some chance of getting near reality?
Coronavirus, what happens next
Re: Coronavirus, what happens next
Probably not.
As Guy said the only truly reliable figure is the number of excess deaths from all causes compared to the average of say the last five years. It then doesn't matter what is recorded on the death certificate. I think most countries keep quite good records of the fact that somebody has died and when.
The only thing that is different this year from the last five years is COVID-19 so there is a pretty high certainty that any excess deaths this year are due to COVID-19 one way or another. It could be people who died of heart attacks or strokes because they weren't admitted to hospital fast enough. That would happen to approximately the same number of people every year but this year it is probably happening more because people are frightened to go to hospitals or ambulances are too busy to respond. Some cancer patients die each year but this year much cancer treatment has been suspended because of COVID-19 so again it is likely that the number who die will be higher. Many will have died of COVID-19 itself whether it is on the death certificate or not. Excess deaths agregates everything so it really is the best measure.
Having said that I don't think other measures are as bad as people make out. There are two main figures when it comes to any form of testing, specificity and sensitivity. The latter relates to how many false negatives you get in a batch of tests and in the case of the swab test used to detect if you have active virus in your body it depends on a lot of factors. There probably are quite a few false negatives, perhaps as high as 30 %, but they don't really change the death reporting. They are more a concern if it is used to tell people that they are safe to go back to work when they aren't. The former relates to how many false positives you get in a batch of tests. The swab test is highly specific since it is analysed in a lab looking for the virus genome. I don't know if there are accurate figures yet but certainly less than 5 % false positives. If the test says you have got it you are over 95 % certain to have it. Couple that with diagnosis based on symptoms and I think it is likely that records of people dying in hospital from COVID-19 are virtually 100 % accurate. In care homes and the community there has been much less testing so for people who die there and have COVID-19 on their death certificates there is a greater uncertainty. I suspect that in Europe, North America and places like South Korea, Singapore and Taiwan doctors certifying death have sufficient experience to produce pretty accurate results even without testing since the way COVID-19 progresses is quite specific.
The UK is an example of a country that was initially only reporting hospital deaths. They did that for a very good reason. They needed rapid accurate results to be able to understand how the virus was progressing. The NHS is a single organisation with well established mechanisms for reporting a variety of information including deaths and initially tests were only being performed in hospitals. The care home sector is a vast collection of largely privately owned businesses with no real structure for rapidly reporting detailed information to the government. Together with deaths in the community the only way the government could collect this information was through death registrations which have an inevitable delay. They are now providing all three figures on a weekly basis from ONS although I think it is still hospital deaths that provide the daily figure. I suspect most developed countries have gone through a similar evolution with some getting complete figures faster than others.
If you look at the Worldometer data rather than John Hopkins you can see the number of cases, tests and deaths normalised to population size. While not perfect I think you can realistically use these figures to do coarse comparisons but probably not fine ones. As an example Italy currently shows 511 deaths per million population while the UK shows 482. To draw the conclusion that the UK has done better than Italy would be very foolish, especially when you also see that Italy is recording 172 new deaths per day while the UK is recording 627. It seems likely that the UK will overtake Italy and the numbers are so close that differences in the methods of recording could be affecting the results. If you now look at a country like Greece, however, it currently shows 15 deaths per million population. I cannot think of any factor of any sort that can account for such a big difference between the UK and Greece other than Greece's restrictions have worked better than those in the UK. As to the future nobody knows.
Warwick
As Guy said the only truly reliable figure is the number of excess deaths from all causes compared to the average of say the last five years. It then doesn't matter what is recorded on the death certificate. I think most countries keep quite good records of the fact that somebody has died and when.
The only thing that is different this year from the last five years is COVID-19 so there is a pretty high certainty that any excess deaths this year are due to COVID-19 one way or another. It could be people who died of heart attacks or strokes because they weren't admitted to hospital fast enough. That would happen to approximately the same number of people every year but this year it is probably happening more because people are frightened to go to hospitals or ambulances are too busy to respond. Some cancer patients die each year but this year much cancer treatment has been suspended because of COVID-19 so again it is likely that the number who die will be higher. Many will have died of COVID-19 itself whether it is on the death certificate or not. Excess deaths agregates everything so it really is the best measure.
Having said that I don't think other measures are as bad as people make out. There are two main figures when it comes to any form of testing, specificity and sensitivity. The latter relates to how many false negatives you get in a batch of tests and in the case of the swab test used to detect if you have active virus in your body it depends on a lot of factors. There probably are quite a few false negatives, perhaps as high as 30 %, but they don't really change the death reporting. They are more a concern if it is used to tell people that they are safe to go back to work when they aren't. The former relates to how many false positives you get in a batch of tests. The swab test is highly specific since it is analysed in a lab looking for the virus genome. I don't know if there are accurate figures yet but certainly less than 5 % false positives. If the test says you have got it you are over 95 % certain to have it. Couple that with diagnosis based on symptoms and I think it is likely that records of people dying in hospital from COVID-19 are virtually 100 % accurate. In care homes and the community there has been much less testing so for people who die there and have COVID-19 on their death certificates there is a greater uncertainty. I suspect that in Europe, North America and places like South Korea, Singapore and Taiwan doctors certifying death have sufficient experience to produce pretty accurate results even without testing since the way COVID-19 progresses is quite specific.
The UK is an example of a country that was initially only reporting hospital deaths. They did that for a very good reason. They needed rapid accurate results to be able to understand how the virus was progressing. The NHS is a single organisation with well established mechanisms for reporting a variety of information including deaths and initially tests were only being performed in hospitals. The care home sector is a vast collection of largely privately owned businesses with no real structure for rapidly reporting detailed information to the government. Together with deaths in the community the only way the government could collect this information was through death registrations which have an inevitable delay. They are now providing all three figures on a weekly basis from ONS although I think it is still hospital deaths that provide the daily figure. I suspect most developed countries have gone through a similar evolution with some getting complete figures faster than others.
If you look at the Worldometer data rather than John Hopkins you can see the number of cases, tests and deaths normalised to population size. While not perfect I think you can realistically use these figures to do coarse comparisons but probably not fine ones. As an example Italy currently shows 511 deaths per million population while the UK shows 482. To draw the conclusion that the UK has done better than Italy would be very foolish, especially when you also see that Italy is recording 172 new deaths per day while the UK is recording 627. It seems likely that the UK will overtake Italy and the numbers are so close that differences in the methods of recording could be affecting the results. If you now look at a country like Greece, however, it currently shows 15 deaths per million population. I cannot think of any factor of any sort that can account for such a big difference between the UK and Greece other than Greece's restrictions have worked better than those in the UK. As to the future nobody knows.
Warwick
Re: Coronavirus, what happens next
According to Crete Tip movement between the mainland and Crete will be allowed from Monday 18 May and Church services start again on Sunday 17 May. Movement between the mainland and other islands is under consideration from 25 May.
I must admit I had hoped that opening of cafés and tavernas would have been allowed before movement from the mainland. It would have been nice to go out for a meal and/or drink in safety before the virus is imported onto the island from the mainland.
Warwick
I must admit I had hoped that opening of cafés and tavernas would have been allowed before movement from the mainland. It would have been nice to go out for a meal and/or drink in safety before the virus is imported onto the island from the mainland.
Warwick
Re: Coronavirus, what happens next
It would have been nice to go out for a meal and/or drink in safety before the virus is imported onto the island from the mainland.
It's not exactly as if the mainland is overcome with Covid-19 so I would not worry too much.
Jean www.west-crete.com
Re: Coronavirus, what happens next
I like the number zero, Jean. Ancient Rome had no concept of zero at all and in ancient Greece it was pretty thin and fleeting.
Warwick
Warwick
Re: Coronavirus, what happens next
I think that zero in terms of Covid-19 is wishful thinking unless you decide to live in a bubble
Jean www.west-crete.com
Re: Coronavirus, what happens next
Kilkis wrote:According to Crete Tip movement between the mainland and Crete will be allowed from Monday 18 May and Church services start again on Sunday 17 May. Movement between the mainland and other islands is under consideration from 25 May.
I must admit I had hoped that opening of cafés and tavernas would have been allowed before movement from the mainland. It would have been nice to go out for a meal and/or drink in safety before the virus is imported onto the island from the mainland.
Warwick
I agree! I thought we might get a few safe weeks
Re: Coronavirus, what happens next
The U.K. government has now stopped showing the death count in the U.K. in comparison with other countries during the daily Number 10 Briefing. - There is a thought that this has happened because the U.K. death count is now an embarrassment for the government.
To be honest, the most important fact is that SO many deaths have occurred in the U.K. regardless of what has happened in other countries! The Government has not dealt well with this crisis, and many lives have been lost because of it. They were too slow to react to begin with. - There was talk of ‘herd immunity‘ and letting people build up a natural defence.....which would result in even more vulnerable and elderly people dying! The lock-down came too late, (after the Cheltenham Festival etc), and even when it was imposed, too much of it was ‘advisory’ rather than compulsory. Almost every day contradictory advice came form different government ministers......sometimes even the same minister!
This government is not fit for purpose. It was elected to ‘get brexit done’ without people thinking how good it would be at doing anything else! - In life, anyone can cope with a situation when all is going well. It is only when there are problems that we see the true capabilities of a person/business/government. I am afraid to say they the majority of the U.K. public voted for BJ and his band of followers, and they are now about to witness what they have put in place to govern the country and steer it through the biggest challenge to the the U.K. in a generation.
To be honest, the most important fact is that SO many deaths have occurred in the U.K. regardless of what has happened in other countries! The Government has not dealt well with this crisis, and many lives have been lost because of it. They were too slow to react to begin with. - There was talk of ‘herd immunity‘ and letting people build up a natural defence.....which would result in even more vulnerable and elderly people dying! The lock-down came too late, (after the Cheltenham Festival etc), and even when it was imposed, too much of it was ‘advisory’ rather than compulsory. Almost every day contradictory advice came form different government ministers......sometimes even the same minister!
This government is not fit for purpose. It was elected to ‘get brexit done’ without people thinking how good it would be at doing anything else! - In life, anyone can cope with a situation when all is going well. It is only when there are problems that we see the true capabilities of a person/business/government. I am afraid to say they the majority of the U.K. public voted for BJ and his band of followers, and they are now about to witness what they have put in place to govern the country and steer it through the biggest challenge to the the U.K. in a generation.
Re: Coronavirus, what happens next
filippos wrote: Can anyone tell me where I could find figures that are truly accurate and can be compared with some chance of getting near reality?
The Financial Times - their pandemic coverage is free to read. There are daily graphs comparing countries (and regions) by excess death percentages - this takes away country differences (because you are looking at excess deaths ie comparing movement in mortality numbers within a country over time) and population size differences (because you are looking at percentage changes so you can compare like with like).
What’s been proved beyond doubt in the last few weeks is that, when schools finally do go back, more time should be spent teaching children about how numbers and charts work, so the next generation is not as ill-informed on what goes on around them as the current one.
Re: Coronavirus, what happens next
Guy M wrote:
What’s been proved beyond doubt in the last few weeks is that, when schools finally do go back, more time should be spent teaching children about how numbers and charts work, so the next generation is not as ill-informed on what goes on around them as the current one.
Like it! Bob.
Yesterday today was tomorrow. Don't dilly dally!
Re: Coronavirus, what happens next
You can get ONS figures of weekly deaths registered in England and Wales here so you can easily see the excess death rate there. I'm not sure where you can get worldwide data for annual deaths so I am not sure that you can do your own comparison of excess death rate in different countries. You can, however, download the worldwide COVID-19 data here and do your own modelling using Excel if you are so inclined.
For example you can easily filter out a particular country, divide the daily new cases/deaths by the population in millions to get the daily new cases/deaths per million. A simple analysis would be to produce two new columns containing 7 day moving averages of new cases/deaths since that only requires addition and division. Plotting the moving averages gives you a smoother curve since it largely gets rid of random reporting fluctuations. That should allow you to see the delay between the peak in cases and the peak in deaths. Comparing these curves for different countries would show roughly how well each country is doing, e.g. height of peak, width of peak, rate of fall after the peak . Before plotting you could adjust the dates for each country backwards/forwards in time so their peaks coincide. You can accumulate daily deaths to produce the sigmoid curve for each country date adjusted by the same amount used to align the peaks. OK that isn't perfect because it still suffers from reporting differences but it gives you a broad-brush idea. I think the biggest anomalies will be in the new cases data because that does depend on the number of tests being carried out and that does vary markedly from country to country. I think that, for a large number of countries, however, the death data will give a reasonable comparison of deaths directly due to COVID-19 in each country but it won't estimate excess deaths from all causes.
If you want to go further you can easily do normal distribution modelling but it does need a bit more mathematical knowledge and a bit more expertise in using Excel. Fortunately Excel has a Solver function which means that you can put rough guesses for the parameters that define the normal distribution and get solver to adjust those parameters to get a best fit between the theoretical distribution and the real data. You can get the equations for the normal distribution and the Error Function, which models the cumulative data, from Wikipedia. Remember that you need to multiply everything by an amplitude factor because the equation are for probability and so integrate to 1. The amplitude factor would just be a guess for what the total number of cases/deaths will be when the disease has run its course.
Warwick
For example you can easily filter out a particular country, divide the daily new cases/deaths by the population in millions to get the daily new cases/deaths per million. A simple analysis would be to produce two new columns containing 7 day moving averages of new cases/deaths since that only requires addition and division. Plotting the moving averages gives you a smoother curve since it largely gets rid of random reporting fluctuations. That should allow you to see the delay between the peak in cases and the peak in deaths. Comparing these curves for different countries would show roughly how well each country is doing, e.g. height of peak, width of peak, rate of fall after the peak . Before plotting you could adjust the dates for each country backwards/forwards in time so their peaks coincide. You can accumulate daily deaths to produce the sigmoid curve for each country date adjusted by the same amount used to align the peaks. OK that isn't perfect because it still suffers from reporting differences but it gives you a broad-brush idea. I think the biggest anomalies will be in the new cases data because that does depend on the number of tests being carried out and that does vary markedly from country to country. I think that, for a large number of countries, however, the death data will give a reasonable comparison of deaths directly due to COVID-19 in each country but it won't estimate excess deaths from all causes.
If you want to go further you can easily do normal distribution modelling but it does need a bit more mathematical knowledge and a bit more expertise in using Excel. Fortunately Excel has a Solver function which means that you can put rough guesses for the parameters that define the normal distribution and get solver to adjust those parameters to get a best fit between the theoretical distribution and the real data. You can get the equations for the normal distribution and the Error Function, which models the cumulative data, from Wikipedia. Remember that you need to multiply everything by an amplitude factor because the equation are for probability and so integrate to 1. The amplitude factor would just be a guess for what the total number of cases/deaths will be when the disease has run its course.
Warwick
Re: Coronavirus, what happens next
Maud wrote: ......In life, anyone can cope with a situation when all is going well. It is only when there are problems that we see the true capabilities of a person/business/government. I am afraid to say they the majority of the U.K. public voted for BJ and his band of followers, and they are now about to witness what they have put in place to govern the country and steer it through the biggest challenge to the the U.K. in a generation.
Or as Mike Tyson once said 'Everyone has a plan until they're punched in the face'
Peter
Re: Coronavirus, what happens next
Peter W wrote:Maud wrote: ......In life, anyone can cope with a situation when all is going well. It is only when there are problems that we see the true capabilities of a person/business/government. I am afraid to say they the majority of the U.K. public voted for BJ and his band of followers, and they are now about to witness what they have put in place to govern the country and steer it through the biggest challenge to the the U.K. in a generation.
Or as Mike Tyson once said 'Everyone has a plan until they're punched in the face'
Peter
I think it is generally true to say that most governments, and the Civil Service are pretty useless in a crisis. Anything other than the status quo throws then into a complete panic. In the case of politicians, this is mostly because they are obsessed with popularity and public opinion rather than doing the right thing in a given situation. In the case of the Civil Service, it's because an emergency requires initiative and creative thinking.
Andrew
Re: Coronavirus, what happens next
YoMo2 wrote:I think it is generally true to say that most governments, and the Civil Service are pretty useless in a crisis. Anything other than the status quo throws then into a complete panic. In the case of politicians, this is mostly because they are obsessed with popularity and public opinion rather than doing the right thing in a given situation. In the case of the Civil Service, it's because an emergency requires initiative and creative thinking.
Andrew
As a retired civil servant, I refuse to rise to the bait! Believe what you like. Who cares? Bob.
Yesterday today was tomorrow. Don't dilly dally!
Re: Coronavirus, what happens next
I don't think the individual civil servants lack initiative or creativity but the civil service does tend to work on quite strict procedures which inhibit/stifle initiative and creativity.
As an example, in 2018 I needed DWP to send me a letter that said that my payments were a UK State Pension and the amount paid for the period 01/01/2016 to 31/12/2016. They couldn't do it. It was totally impossible. They were only allowed to send out standard format letters and there was no such standard format letter. In the end they managed to send two letters. The first, IPC PMTS 133 01/14, contained the statement "I am writing to confirm that your State Pension, awarded to you from ..., is now payable at the weekly amount of .... This identified that I was receiving a UK State Pension but the amount shown was the weekly amount in 2018. The second, IPC 147 12/13, showed the amount paid for the correct period but had the statement "The statement of benefit payments that you asked for is given below." i.e it did not say what the benefit was.
That type of procedural rigidity does not lend itself to rapid proactive response during an emergency. I suspect that rigid procedural thinking did cause some of the problems in PPE procurement for example. Several established UK suppliers of PPE to non-NHS users, both in the UK and abroad, who are approved by the NHS but did not normally supply to the NHS, were contacting the NHS to offer PPE and not getting any response. Having said that I think the main problem in the UK was slow political decision making.
Warwick
As an example, in 2018 I needed DWP to send me a letter that said that my payments were a UK State Pension and the amount paid for the period 01/01/2016 to 31/12/2016. They couldn't do it. It was totally impossible. They were only allowed to send out standard format letters and there was no such standard format letter. In the end they managed to send two letters. The first, IPC PMTS 133 01/14, contained the statement "I am writing to confirm that your State Pension, awarded to you from ..., is now payable at the weekly amount of .... This identified that I was receiving a UK State Pension but the amount shown was the weekly amount in 2018. The second, IPC 147 12/13, showed the amount paid for the correct period but had the statement "The statement of benefit payments that you asked for is given below." i.e it did not say what the benefit was.
That type of procedural rigidity does not lend itself to rapid proactive response during an emergency. I suspect that rigid procedural thinking did cause some of the problems in PPE procurement for example. Several established UK suppliers of PPE to non-NHS users, both in the UK and abroad, who are approved by the NHS but did not normally supply to the NHS, were contacting the NHS to offer PPE and not getting any response. Having said that I think the main problem in the UK was slow political decision making.
Warwick
Return to “General Discussion & News”
Who is online
Users browsing this forum: Bing [Bot], Google Feedfetcher and 45 guests