Covid : Infection rates v death rates and a possible second wave

[Domh245]

In the context of the overall pandemic it certainly is. From here it seems you have carefully chosen a start date that suits your agenda, rather than allowing the data to show the true picture.

I fail to see how the context of the overall pandemic is important here? Like I said previously, the request was to show evidence of an increase in hospital admissions - not "hospital admissions relative to the peak in April". All that including the earlier data does is change the vertical scaling of the graph making it much harder to pick out the small but definite increase*.

I'm not disputing that hospital admissions are nowhere near where they were during April, I'm very glad that we aren't. However hospital admissions are undeniably on the rise, and to try and ignore that by arguing it's small compared to April is to miss the point entirely.

*a problem solved by use of a log scale, but which google sheets won't let you do for that (combo) style of graph.

Edit: Oh and you've omitted the fact that as hospitals have started to catch-up on a huge backlog of cases, there will be an noticeable increase admissions & with that comes the possibility of more people with, but not necessarily suffering from covid.

An entirely fair point. There is that risk that some of the number of people are "people with COVID" as opposed to "people ill as a result of COVID", but that data isn't available. The closest analogue would be the "patients in mechanical ventilator" beds, but annoyingly the data portal is having issues and isn't showing that right now. My suspicion, and it is only that, however is the number of people in the statistics who are "with, not ill because of" is small. My understanding is that most people going for delayed elective surgery are tested at home several days before the surgery, meaning positive cases through that route shouldn't make it into the hospital admissions statistic as they'd be instructed to isolate at home if a positive test returns. That's not to rule out people turning up at A&E with a complaint and then discovering that they have COVID, but I really can't imagine that there are all too many cases of that compared to people actually ill with COVID and admitted because of it.

 

[AdamWW]

I fail to see how the context of the overall pandemic is important here? Like I said previously, the request was to show evidence of an increase in hospital admissions - not "hospital admissions relative to the peak in April". All that including the earlier data does is change the vertical scaling of the graph making it much harder to pick out the small but definite increase*.

Absolutely!

Squishing the recent data into the bottom right corner of the graph and then saying "Nothing to see here".... that looks more like picking the data to suit an agenda.

Although I'd say that looking back at points after the peak would be instructive, to see if there have been similar rises that then dropped away.

I don't think anyone here is claiming that this in itself is a huge cause for alarm, but given that it's coinciding with an increase in infections it may be the first signs of something that is a worry.

An entirely fair point. There is that risk that some of the number of people are "people with COVID" as opposed to "people ill as a result of COVID", but that data isn't available.

That is something I'd like to know. Though your argument as to why the "with COVID" numbers are likely to be small seems quite convincing.

I think it's very important, because if we're including a significant number of "with COVID" people it will automatically tend to track infections.

 

[Class 33]

2,621 new cases and 9 new deaths today. Not bad at all.

Deaths STILL very low.

As for the new cases, yes that is quite a substantial drop from yesterday's 3,300 something. And the THIRD day of cases dropping. Hopefully these cases will continue to fall week by week now. And that huge rise up was just a nasty spike and not a "Terrifying second wave". Let's get these cases further dropping so the government doesn't bring in yet more bloody restrictions and laws, AND by about November we can get this ridiculous social distancing and face mask wearing laws finally scrapped and the nation can start getting back to normal. Like many millions of people in the UK, this year from March onwards has been a complete and utter write-off due to the lockdown and these ongoing restrictions ruining all our plans and hopes for the year. I hope 2021 can be a full year of back to normal again, and we've moved on from all this.

 

[AdamWW]

Deaths STILL very low.

As for the new cases, yes that is quite a substantial drop from yesterday's 3,300 something. And the THIRD day of cases dropping. Hopefully these cases will continue to fall week by week now. And that huge rise up was just a nasty spike and not a "Terrifying second wave". Let's get these cases further dropping so the government doesn't bring in yet more bloody restrictions and laws, AND by about November we can get this ridiculous social distancing and face mask wearing laws finally scrapped and the nation can start getting back to normal. Like many millions of people in the UK, this year from March onwards has been a complete and utter write-off due to the lockdown and these ongoing restrictions ruining all our plans and hopes for the year. I hope 2021 can be a full year of back to normal again, and we've moved on from all this.

It seems that Sweden has had spikes in infections that dropped away again, so let's hope.

But if that's the case, it may just mean that the current level of precautions is indeed working.

I can't see how that indicates that we can just drop all the restrictions though.

 

[ainsworth74]

Back when we were having thousands of positives per day in March/April/May weren't Monday's figures always lower?

 

[adc82140]

Deaths were, yes, but not new cases, certainly not by this amount anyway.

 

[AdamWW]

Deaths were, yes, but not new cases, certainly not by this amount anyway.

But we also seem to have hit recent problems with testing capacity, so we may now be under-reporting infections...

 

[adc82140]

We were always under reporting infections. Probably by a factor of 1000% in April.

 

[AdamWW]

We were always under reporting infections. Probably by a factor of 1000% in April.

Yes of course.

But given that recently it seems to have become a lot harder to get tests, that suggests that under-reporting is becoming a bigger problem then it has been over the last month or so.

 

[kristiang85]

Deaths STILL very low.

As for the new cases, yes that is quite a substantial drop from yesterday's 3,300 something. And the THIRD day of cases dropping. Hopefully these cases will continue to fall week by week now. And that huge rise up was just a nasty spike and not a "Terrifying second wave". Let's get these cases further dropping so the government doesn't bring in yet more bloody restrictions and laws, AND by about November we can get this ridiculous social distancing and face mask wearing laws finally scrapped and the nation can start getting back to normal. Like many millions of people in the UK, this year from March onwards has been a complete and utter write-off due to the lockdown and these ongoing restrictions ruining all our plans and hopes for the year. I hope 2021 can be a full year of back to normal again, and we've moved on from all this.

Look at Spain and Belgium - they had a surge, which now is quickly petering out, with only a tiny increase in deaths. And the surges seemed to be in places that didn't have it yet. A similar-ish pattern seems to be here. London doesn't seem to be that affected at all, despite so many crowds/protests/pub gatherings/etc. in the last few weeks. So much evidence points to a low threshold for local population immunity, and that gives me a lot of hope that this could burn out soon. Obviously we need to wait another couple of months for it to become really apparent, though. Which might be too late to undo the damage to the economy.

 

[Jamesrob637]

Back when we were having thousands of positives per day in March/April/May weren't Monday's figures always lower?

Sunday and Monday are Saturday and Sunday for the purpose of statistics. No more bank holiday weekends now for the rest of the year, Christmas notwithstanding.

 

[Bantamzen]

I fail to see how the context of the overall pandemic is important here? Like I said previously, the request was to show evidence of an increase in hospital admissions - not "hospital admissions relative to the peak in April". All that including the earlier data does is change the vertical scaling of the graph making it much harder to pick out the small but definite increase*.

I'm not disputing that hospital admissions are nowhere near where they were during April, I'm very glad that we aren't. However hospital admissions are undeniably on the rise, and to try and ignore that by arguing it's small compared to April is to miss the point entirely.

*a problem solved by use of a log scale, but which google sheets won't let you do for that (combo) style of graph.



An entirely fair point. There is that risk that some of the number of people are "people with COVID" as opposed to "people ill as a result of COVID", but that data isn't available. The closest analogue would be the "patients in mechanical ventilator" beds, but annoyingly the data portal is having issues and isn't showing that right now. My suspicion, and it is only that, however is the number of people in the statistics who are "with, not ill because of" is small. My understanding is that most people going for delayed elective surgery are tested at home several days before the surgery, meaning positive cases through that route shouldn't make it into the hospital admissions statistic as they'd be instructed to isolate at home if a positive test returns. That's not to rule out people turning up at A&E with a complaint and then discovering that they have COVID, but I really can't imagine that there are all too many cases of that compared to people actually ill with COVID and admitted because of it.

Given that the official NHS waiting list at the end of last month was reported to be in the region of 3.9 million patients, it is possible that patients will be entering the system even if they have tested positive. Many people's treatments cannot be put on hold because of the virus. So the "with covid" figures could be higher than you might imagine.

Yes of course.

But given that recently it seems to have become a lot harder to get tests, that suggests that under-reporting is becoming a bigger problem then it has been over the last month or so.

You say bigger problem, but if those new cases being discovered are largely people who do not require medical attention or who do not display symptoms then surely the reverse is true? That is to say the more cases like these that are found, the less dangerous the virus actually is compared to what was believed at the beginning of the year, because it is generally passing through a larger proportion of people without causing issues.

 

[talldave]

Yes of course.

But given that recently it seems to have become a lot harder to get tests, that suggests that under-reporting is becoming a bigger problem then it has been over the last month or so.

Why is it a bigger problem? Why is it a problem at all?

 

[AdamWW]

Why is it a bigger problem? Why is it a problem at all?

Eh?

Are you saying you don't believe that difficulty in getting tests will result in under-reporting actual infections, or you think that we don't actually need to know how many people are infected?

Look at Spain and Belgium - they had a surge, which now is quickly petering out, with only a tiny increase in deaths. And the surges seemed to be in places that didn't have it yet. A similar-ish pattern seems to be here. London doesn't seem to be that affected at all, despite so many crowds/protests/pub gatherings/etc. in the last few weeks. So much evidence points to a low threshold for local population immunity, and that gives me a lot of hope that this could burn out soon. Obviously we need to wait another couple of months for it to become really apparent, though. Which might be too late to undo the damage to the economy.

I'm not sure Belgium is a good example given that their 'cases' and hospitalisations are now going up according to this. (But not deaths, attributed here to better protection of high risk groups than before). Reducing deaths is of course good, but in itself has no bearing on the issue of whether there is a risk of overloading the health system or not if infections continue to increase.

The reproduction rate of the virus, which reflects how contagious it is, was estimated at 1.36 for the September 6-12 period. On Saturday it was still 1.27. When the reproduction rate is over 1, that means the pandemic is spreading. A rate of less than 1 means it is on the decrease.

...

The number of COVID-19 infections since the start of the pandemic some six months ago is now 92,478. Daily admissions to hospitals continue to go up, averaging 26.6 from September 6 to 12, a 58% increase. All told, 19,273 persons have been hospitalised for COVID-19 in Belgium.

According to Sciensano, 260 hospital beds were occupied on Saturday, 68 of them in intensive care units.

However, the number of deaths has been going down and now averages 2.6 per day, an 18% drop compared to the previous one-week period.

“This is good news. It means we are managing to protect high-risk groups,” Van Gucht commented, although he did not rule out a deterioration of the situation, as occurred in Spain.
[\QUOTE]

 

[Bantamzen]

Eh?

Are you saying you don't believe that difficulty in getting tests will result in under-reporting actual infections, or you think that we don't actually need to know how many people are infected?

We will never fully know how many people have had it, largely because many people who have had it will not have known to even get a test. And we don't need to know exactly how many, any more than we need to know how many people have had the flu. What we need to understand what proportion of people might need medical help so that we can plan accordingly.

 

[AdamWW]

We will never fully know how many people have had it, largely because many people who have had it will not have known to even get a test. And we don't need to know exactly how many, any more than we need to know how many people have had the flu. What we need to understand what proportion of people might need medical help so that we can plan accordingly.

Which depends on how many people become infected, which depends on what the infection rates are.

Have I misunderstood or are you actually saying that we don't need to know what rate infections are rising or falling?

And, by the way, the government does put effort into tracking influenza spread:

From this government document: Sources of UK flu data: influenza surveillance in the UK

We use several sources of data to understand influenza activity in the UK. During the influenza season we include data from these sources in the weekly national influenza reports and graphs. Many of the data sources have been used for several years. We developed some new systems during the pandemic in 2009.

 

[Bletchleyite]

If you can be reinfected, you don't need a vaccine because it obviously didn't kill you.

This obsession with vaccines for something that primarily only kills people who aren't in full health is so tiring.

Which would be fine if it didn't kill anybody, of course, but it does.

Deaths STILL very low.

Deaths lag cases by about 3 weeks, so we won't know if the uptick in cases has affected them for about 2.5 weeks.

 

[Bantamzen]

Which depends on how many people become infected, which depends on what the infection rates are.

Have I misunderstood or are you actually saying that we don't need to know what rate infections are rising or falling?

No, we don't need to know exact numbers, and we will never know as many people will never be tested even though they may well have had the virus. What we need to know is an idea of how many people might need actual medical help as opposed to just being poorly at home, or even not feeling much at all.

And, by the way, the government does put effort into tracking influenza spread:

From this government document: Sources of UK flu data: influenza surveillance in the UK

Hmmm, I can see a rather large possible flaw in that. How many people get flu, ring in sick then just dive under the duvet until they feel better?

 

[AdamWW]

If you can be reinfected, you don't need a vaccine because it obviously didn't kill you.

No because vaccines aren't just about protecting the individual being vaccinated. Vaccination programmes are generally about producing herd immunity so that those who won't or can't be vaccinated are also protected.

Also assuming you won't die or develop long term illness second time round.

This obsession with vaccines for something that primarily only kills people who aren't in full health is so tiring.

Do explain.

Once someone is no longer in full health it's fine if they die?

I don't suppose that's what you're saying but I don't know what you do mean.

No, we don't need to know exact numbers, and we will never know as many people will never be tested even though they may well have had the virus.

Yes I do actually understand that.

But we do have figures from positive tests and that gives an indication of how infections are rising or falling. The more the criteria for getting a test change, the harder these results are to interpret.

Hmmm, I can see a rather large possible flaw in that. How many people get flu, ring in sick then just dive under the duvet until they feel better?

Well yes, if all they were doing was counting how many people phoned in sick. But - and this may surprise you - what the government does is actually a bit more sophisticated than that.

 

[Bantamzen]

Yes I do actually understand that.

But we do have figures from positive tests and that gives an indication of how infections are rising or falling. The more the criteria for getting a test change, the harder these results are to interpret.

Its not infections we are most interested in, it is the rate at which it causes people to require medical attention.

Well yes, if all they were doing was counting how many people phoned in sick. But - and this may surprise you - what the government does is actually a bit more sophisticated than that.

Well I work in the public sector, developing statistical applications, so yeah I kind of know that. But thanks for the condescending remark!

 

[talldave]

Eh?

Are you saying you don't believe that difficulty in getting tests will result in under-reporting actual infections, or you think that we don't actually need to know how many people are infected?

Neither. I'm saying it's not a problem that the figures aren't accurate. We got through lockdown without accurate figures. And if it's meant to be a bigger problem, bigger than what?

Masks are "the problem", they've killed antisocial distancing.

Which would be fine if it didn't kill anybody, of course, but it does.

It kills people for sure, in "bad flu" proportions. But this country is behaving as if it is Ebola we're dealing with.

 

[AdamWW]

Its not infections we are most interested in, it is the rate at which it causes people to require medical attention.

Yes, agreed.

But since hospitalisations lag infections and infections have the potential to rise rapidly, it seems to me and the consensus appears to be that monitoring infections is a useful thing to do.

Well I work in the public sector, developing statistical applications, so yeah I kind of know that. But thanks for the condescending remark!

OK maybe that was a bit low, but your comment did rather suggest you thought they were doing something rather simplistic.

I can't tell unless someone says what their background is - people have a habit of making definitive statements as if they are experts in what they are talking about. Some no doubt are, and others aren't.
(And while I try to avoid doing that myself, I know I don't always succeed).

Neither. I'm saying it's not a problem that the figures aren't accurate. We got through lockdown without accurate figures. And if it's meant to be a bigger problem, bigger than what?

Er...bigger than it was when if you wanted a test you could actually get one.

As for getting through lockdown - I'd say that was the easy bit.

Working out how and at what rate to release restrictions is the difficult part, and that's where you need data.

 

[jtuk]

Look at Spain and Belgium - they had a surge, which now is quickly petering out, with only a tiny increase in deaths.

Spain R reported to already be under 0.5, the lowest rate it's ever been

London doesn't seem to be that affected at all, despite so many crowds/protests/pub gatherings/etc. in the last few weeks. So much evidence points to a low threshold for local population immunity, and that gives me a lot of hope that this could burn out soon. Obviously we need to wait another couple of months for it to become really apparent, though. Which might be too late to undo the damage to the economy.

Well duh, London got its protection before Boris bottled it, the rest of us have to wait, one rule for London and one rule for everyone else as usual

 

[nlogax]

Spain R reported to already be under 0.5, the lowest rate it's ever been

Where was that reported?

 

[WelshBluebird]

This obsession with vaccines for something that primarily only kills people who aren't in full health is so tiring.

Do you include flu with that comment then?
We spend a considerable amount of time and money on the annual flu vaccine. Are you suggested we shouldn't bother with that either?

 

[AdamWW]

Do you include flu with that comment then?
We spend a considerable amount of time and money on the annual flu vaccine. Are you suggested we shouldn't bother with that either?

Yes - I find it interesting that a lot of the "why are we treating this as any different from flu" comments seem to me to overlook not only various good reasons why we should treat it differently, but also the time and effort that does go into mitigating the effects of flu.

It's true that we accept a considerable number of deaths each winter due to flu. It's clearly not true that this arises from just letting flu take its course without any mitigation.

 

[Domh245]

Given that the official NHS waiting list at the end of last month was reported to be in the region of 3.9 million patients, it is possible that patients will be entering the system even if they have tested positive. Many people's treatments cannot be put on hold because of the virus. So the "with covid" figures could be higher than you might imagine.

Granted, but I'd be surprised if they did choose to admit somebody for elective surgery even if they tested positive, not only is that putting staff at risk, but also increases the risk of complications arising from the surgery. I can't find any definitive source* for "if they test positive in their pre-admission test, then you must refuse admission" - indeed they tend to only caution that "surgery may be postponed if you test positive" but my suspicion is that once again the numbers of people who fall under this criteria are tiny, certainly not enough to distort the trend of COVID hospitalisations.


* Operating framework for urgent and planned services in hospital settings during COVID-19 (14.5.20)
NICE new COVID-19 rapid guideline on arranging planned care in hospitals and diagnostic services
Preparing for Elective Surgical Admissions - Advice for Patients (Northern Care Alliance)

 

[Huntergreed]

Yes - I find it interesting that a lot of the "why are we treating this as any different from flu" comments seem to me to overlook not only various good reasons why we should treat it differently, but also the time and effort that does go into mitigating the effects of flu.

It's true that we accept a considerable number of deaths each winter due to flu. It's clearly not true that this arises from just letting flu take its course without any mitigation.

There’s a different level of mitigation used for flu which is several orders of magnitude smaller and less damaging on the economy than the current COVID mitigations.

How long do you propose we continue with these restrictions for?

 

[AdamWW]

There’s a different level of mitigation used for flu which is several orders of magnitude smaller and less damaging on the economy than the current COVID mitigations.

Agreed fully.

How long do you propose we continue with these restrictions for?

I dont. Fortunately I'm not running the country. And if I was I would be spending a lot more time than I do looking into the evidence and I'd be able to question those providing the advice.

However given what I do know, I wouldn't be inclined to drop all restrictions now and hope for the best because while where we now isn't great, I think a country with coronavirus spreading unchecked would be even less pleasant.

 

[Bantamzen]

*

Granted, but I'd be surprised if they did choose to admit somebody for elective surgery even if they tested positive, not only is that putting staff at risk, but also increases the risk of complications arising from the surgery. I can't find any definitive source* for "if they test positive in their pre-admission test, then you must refuse admission" - indeed they tend to only caution that "surgery may be postponed if you test positive" but my suspicion is that once again the numbers of people who fall under this criteria are tiny, certainly not enough to distort the trend of COVID hospitalisations.


* Operating framework for urgent and planned services in hospital settings during COVID-19 (14.5.20)
NICE new COVID-19 rapid guideline on arranging planned care in hospitals and diagnostic services
Preparing for Elective Surgical Admissions - Advice for Patients (Northern Care Alliance)

Maybe, but remember that the NHS basically shut down a lot of treatments, ops etc. So in a growing backlog of cases, some routine cases may have moved into the urgent stage, meaning that the possibility of more covid-positive cases entering the system.