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

[Domh245]

I had to read the above several times before it sunk in, as it's so utterly counterintuitive, but once you 'get it', you realise that that PCR test results are a terrible basis (on their own) for deciding where/when to introduce local restrictions/lockdowns.


Unfortunately, the vast majority of the public, including senior government figures, seem to think the test is infallible.

That logic holds when the sampling is random or of asymptomatic people. When the sampling is of people with symptoms (Dido Harding's evidence to a select committee yesterday is that around 75% of people taking the tests are symptomatic), the argument doesn't work so well.

 

[trebor79]

That logic holds when the sampling is random or of asymptomatic people. When the sampling is of people with symptoms (Dido Harding's evidence to a select committee yesterday is that around 75% of people taking the tests are symptomatic), the argument doesn't work so well.

It does, because only about 1.5% to 2% of those tested are testing positive. If the specificity of the test is 99.5%, that means between a quarter and a third of the positives are false positives at present.

 

[Domh245]

It does, because only about 1.5% to 2% of those tested are testing positive. If the specificity of the test is 99.5%, that means between a quarter and a third of the positives are false positives at present.

I'm not saying that there aren't false positives, only that these are likely far less than the actual number of actual positives. In the quoted post, it was estimated that 91% of test results were false positives which is obviously ludicrous. Because of the self selecting nature of the people testing, the assumed 50 in the 100,000 is far too low for the number of actual cases (especially as last week the ONS estimate for the general population had gone up to 1 in 1400 for the week ending 5th September, and will no doubt increase again when the results of last week's study are published) even though the rest of the argument holds.

Going through the testing and cases data, our 'worst case' specificity is 99.61% if all of the 512 cases announced on the 10th of July (from 131848 tests) were false positives, meaning that we'd be currently sitting at about 25% false positives. Even taking an average (99.46) of the specificity of tests calculated by PHE, we're sitting at a third of cases being false positives. But that's still a far stretch from the 90% false positive rate claimed in the first post.

You've also got to remember that public health officials will have more than just the test results to look at - things like the ZOE and ONS studies will also be used to build a complete picture off the back of which they might decide to go into lockdown (poor idea as it is IMO)

 

[takno]

I'm not saying that there aren't false positives, only that these are likely far less than the actual number of actual positives. In the quoted post, it was estimated that 91% of test results were false positives which is obviously ludicrous. Because of the self selecting nature of the people testing, the assumed 50 in the 100,000 is far too low for the number of actual cases (especially as last week the ONS estimate for the general population had gone up to 1 in 1400 for the week ending 5th September, and will no doubt increase again when the results of last week's study are published) even though the rest of the argument holds.

Going through the testing and cases data, our 'worst case' specificity is 99.61% if all of the 512 cases announced on the 10th of July (from 131848 tests) were false positives, meaning that we'd be currently sitting at about 25% false positives. Even taking an average (99.46) of the specificity of tests calculated by PHE, we're sitting at a third of cases being false positives. But that's still a far stretch from the 90% false positive rate claimed in the first post.

You've also got to remember that public health officials will have more than just the test results to look at - things like the ZOE and ONS studies will also be used to build a complete picture off the back of which they might decide to go into lockdown (poor idea as it is IMO)

Agreed that a third is probably the right number just now, and it's always worth bearing that in mind when doing international comparisons or comparisons to earlier in the year where the numbers of tests being performed were much lower.

The ONS studies, being less targeted, are even more subject to false positives than the tier 2 figures. I haven't read the ONS methodology for a couple of months, so they make take that into account now, but there were few signs that it was being taken into account in July and August. Hopefully specialist public health officials are able to take account of these things, but my impression is that very few of the non-academics can really get their head round the meaning of the numbers they are looking at all, and a surprisingly large number of the specialist academics aren't fantastic at it either.

 

[Jozhua]

A two week lockdown? What the hell is that supposed to achieve?
What's so ridiculous is that there is no certainty. We have no idea when measures will be added or removed, what to expect.

The government says they will not do something, just to do it two weeks later. Generally, the rule is, whatever the government denies they will do now, is what will happen in the next couple of weeks.

A significant amount of people are now straight up ignoring the rules, as we see a confusing mess of massively varying rules from region to region. Not only that, but each country seems to be taking a differing approach too.

The idea seems to be that rules act like a "brake", with the more restrictions leading to an associated drop in transmission. In reality, some changes are significantly more effective than others and some have a much larger cost than others.

For now, I'm avoiding large gatherings, parties, etc and wearing a face covering in indoor shared spaces. I was incredibly careful when visiting my grandad to make sure I washed my hands, etc.

But I am still meeting friends, using public transport, shopping, etc and otherwise living my life as nornal. Doing so has helped keep me sane and healthy.

 

[Yew]

A two week lockdown? What the hell is that supposed to achieve?

Make it seem like the government is doing something? Appease the screaming deluded hordes who think this is some sort of enemy to be defeated? Make Boris seem "tough" on COVID?

 

[bramling]

I see on the Jeremy Vine show this morning the debate is about whether we should have another two week national lockdown to try and get the virus under control. A mix of support and opposition. Some people saying they believe that a second lockdown is needed to keep us safe and they would support a harsher lockdown than the first one (ie everything closes, if you haven’t got enough food, tough.) It’s scary how authoritarian some people are.



I believe this would be pointless, because if we did have one it would simply delay what’s already happening and the virus would spread again when we come out. We cannot go on like this, we can’t eliminate it, the only realistic option is to let it spread. Otherwise, pending a vaccine, we’re facing these measures for quite literally decades.

Yes I can’t see what on earth a two week lockdown will achieve, apart from cause disruption and then we start again afterwards. Also a lot of people thought the last one was going to be fairly short, yet it ended up lasting for months.

Personally, I will not abide by it. Hancock can do one AFAIAC!

Having said that, the vibe I’m getting it there’s certainly people who are all for it. No doubt they find the idea of more paid time off work appealing.

== Doublepost prevention - post automatically merged: 18 Sep 2020 ==

Another lockdown will surely mean the Bank of England go into negative interest rates? So savers, especiall the retired banking on savings, be prepared to pay banks to keep your money.

That’s lethal, as it could mean people decide to simply withdraw their savings. There’s a million and one reasons why that’s a *really* *bad* thing.

I have to say I am becoming increasingly bothered by the government’s handling of all this now. There is the risk of some really catastrophic damage being done to livelihoods if this nonsense carries on. At the moment we’ve already stored up a massive debt for the future, and not really got a lot in return for that, but it’s starting to go beyond that now. A run on the banks would be utterly catastrophic, yet it’s exactly the sort of thing this shower of a government would cause as they seem utterly unable to consider the consequences of their actions.

Thankfully I don’t think banks would pass negative rates onto households as they surely know it would lead to mass withdrawals. And surely the government would know people stashing notes in the bottom of wardrobes is a terrible thing?

Hancock is completely out of depth, as it Johnson, and in some ways Sunak is worse - let’s pack all the restaurants through August and undo everything they’d tried to do through the preceding months.

 

[DB]

in some ways Sunak is worse - let’s pack all the restaurants through August and undo everything they’d tried to do through the preceding months.

... and then three weeks later Sunak's colleagues start threatening to make them close again.

 

[bramling]

... and then three weeks later Sunak's colleagues start threatening to make them close again.

It’s complete and utter madness. Did any other country go for a run on restaurants at the point where they’re still in a supposedly bad position?

Whether or not restaurants really did contribute or not, it sends out (yet another) conflicting message.

 

[MikeWM]

Thankfully I don’t think banks would pass negative rates onto households as they surely know it would lead to mass withdrawals. And surely the government would know people stashing notes in the bottom of wardrobes is a terrible thing?

Today, perhaps. If the people pushing the cashless society agenda succeed, you won't have that option anymore.

 

[Peter Mugridge]

As per Yorkie's reply, there is no 'data' as such, but the nature of the PCR test
means that can never be 100% reliable, and because of the way that errors are
compounded when prevalence of the virus in the population is extremely low,
as it is at the moment, even if the test is is 99% reliable that can still lead
to up to 90% false positives!

I realise that the bit in bold above sounds absolutely bonkers,
but it was explained to me by a statistician on another forum
like this:

I had to read the above several times before it sunk in, as it's so utterly counterintuitive,
but once you 'get it', you realise that that PCR test results are a terrible basis (on their own)
for deciding where/when to introduce local restrictions/lockdowns.

Unfortunately, the vast majority of the public, including senior government figures,
seem to think the test is infallible.

I just asked a highly experienced professional statistician at work to have a look, and he confirms that the rate of false positives is very high. What he said is below:

They are confusing specificity with sensitivity, but otherwise the reasoning is sound.

Specificity is a type of accuracy measure, it tells how well a test identifies patients who do not have a disease (ie covid). It is the proportion of people without the disease who will have a negative result.

Sensitivity is a type of accuracy measure, it measures the proportion of positives that are correctly identified (e.g., the percentage of people with the disease who are correctly identified as having the disease).

Say we test 1 000 000 people with no symptoms. 2% of those tested are testing positive. If the sensitivity of the test is 99.5%, then:

1) 2% of 1M is 20000 people who are tested positive

2) if covid prevalence is 1.5%, than 15000 got it.

3) this means 20000-15000=5000 are false positives so 0.5% -> 99.5% sensitivity of the test (how many of the people with the condition are tested positive)

4) 5000/20000=25% of the positive are false positive.

This means a quarter of the positives are false positives at present.

 

[bramling]

Today, perhaps. If the people pushing the cashless society agenda succeed, you won't have that option anymore.

Shudder!

To be honest I find it irresponsible that the B of E hasn’t actually ruled out negative interest rates. A run on the banks would be catastrophic, and could easily happen if there was even a whiff of such rates being passed on.

 

[trebor79]

I just asked a highly experienced professional statistician at work to have a look, and he confirms that the rate of false positives is very high. What he said is below:

Ironically, your correspondent appears himself to have muddled up specificity and sensitivity!

 

[yorksrob]

Yes I can’t see what on earth a two week lockdown will achieve, apart from cause disruption and then we start again afterwards. Also a lot of people thought the last one was going to be fairly short, yet it ended up lasting for months.

Personally, I will not abide by it. Hancock can do one AFAIAC!

Having said that, the vibe I’m getting it there’s certainly people who are all for it. No doubt they find the idea of more paid time off work appealing.

== Doublepost prevention - post automatically merged: 18 Sep 2020 ==



That’s lethal, as it could mean people decide to simply withdraw their savings. There’s a million and one reasons why that’s a *really* *bad* thing.

I have to say I am becoming increasingly bothered by the government’s handling of all this now. There is the risk of some really catastrophic damage being done to livelihoods if this nonsense carries on. At the moment we’ve already stored up a massive debt for the future, and not really got a lot in return for that, but it’s starting to go beyond that now. A run on the banks would be utterly catastrophic, yet it’s exactly the sort of thing this shower of a government would cause as they seem utterly unable to consider the consequences of their actions.

Thankfully I don’t think banks would pass negative rates onto households as they surely know it would lead to mass withdrawals. And surely the government would know people stashing notes in the bottom of wardrobes is a terrible thing?

Hancock is completely out of depth, as it Johnson, and in some ways Sunak is worse - let’s pack all the restaurants through August and undo everything they’d tried to do through the preceding months.

I don't agree regarding eat out to help out. The idea was to get people out, helping the economy in a covid secure way. In pretty much everywhere I've eaten out to help out, regulations have been followed and distances kept, so I don't see what else could have been done, short of letting the hospitality industry croak it.

 

[bramling]

I don't agree regarding eat out to help out. The idea was to get people out, helping the economy in a covid secure way. In pretty much everywhere I've eaten out to help out, regulations have been followed and distances kept, so I don't see what else could have been done, short of letting the hospitality industry croak it.

Yet that’s what has essentially been done with other attractions like museums and the arts. Any reintroduction of invasive measures now will be utterly terminal.

Restaurants would have seen some natural recovery in demand over the school holidays anyway, but could have done without being mobbed. It’s all very well saying it was all Covid secure, but if that was really the case why did we see places pulling out of the scheme or having to close because it was causing problems?

 

[yorksrob]

Yet that’s what has essentially been done with other attractions like museums and the arts. Any reintroduction of invasive measures now will be utterly terminal.

Restaurants would have seen some natural recovery in demand over the school holidays anyway, but could have done without being mobbed. It’s all very well saying it was all Covid secure, but if that was really the case why did we see places pulling out of the scheme or having to close because it was causing problems?

I think that's wrong.

The restaurants needed the trade. I'm highly sceptical that the extra business has contributed to the rise in infections in any meaningful way.

 

[Bletchleyite]

The restaurants needed the trade. I'm highly sceptical that the extra business has contributed to the rise in infections in any meaningful way.

If sit-down restaurants were, the local lockdowns would all have closed them. Only one (Bolton) did.

 

[DB]

I don't agree regarding eat out to help out. The idea was to get people out, helping the economy in a covid secure way. In pretty much everywhere I've eaten out to help out, regulations have been followed and distances kept, so I don't see what else could have been done, short of letting the hospitality industry croak it.

In which case, talk of another lockdown just indicates how shambolic this government is! Anyone could have predicted that an increase in cases around the time when all respiratory infections increase was highly likely - so what point in encouraging people to use restaurants, then threatening to close them a few weeks later because of a highly predictable situation with case numbers?

 

[bramling]

In which case, talk of another lockdown just indicates how shambolic this government is! Anyone could have predicted that an increase in cases around the time when all respiratory infections increase was highly likely - so what point in encouraging people to use restaurants, then threatening to close them a few weeks later because of a highly predictable situation with case numbers?

I’m wouldn’t be surprised if there was any thinking it went like “cases are going down so let’s be popular and give people a bit off their restaurant bills, keep them happy before it all kicks off again”.

It certainly won’t keep me happy if my September holiday gets disrupted.

 

[big_rig]

I see the Independent Sage have released their '10 point plan' (one page of text in case you thought it was a report) to prevent a national lockdown. I hope nobody who matters is taking this lot seriously because it really sounds like nonsense - especially their IMMEDIATE MEASURES (!!). At least they have left transport alone for the most part but this just doesn't seem like serious stuff - is it just an excuse to pivot to the lockdown call?

The absence of mask rules is notable too. I imagine now that push has come to shove they don't actually trusts masks hence the desire just to close everything.

IMMEDIATE MEASURES (to be reviewed, as appropriate, when testing is functional and infections are back under control)
1. Close indoor service in pubs and restaurants [kill hospitality, ok.]
2. Work or study at home if possible. If impossible; don’t return to the workplace/University/College until certified COVID safe and provide financial support to employees in the interim [I guess this removes some travel but largely redundant as all of these locations are currently are Covid secure if open currently - if you want to close everything be honest]
3. Limit indoor social interaction to a bubble of three households [can't imagine this is really high impact and intent seems redundant in most part due to rule of 6]
4. Fund schools to allow smaller socially distanced classes and provide digital resources for all students and teachers, should they have to study and work at home [redundant due to point 2, has already happened anyway]
5. Create an emergency fund to support all those needing to shield or self-isolate (including those unable to get tested and their contacts) [SSP is bad so this has some merit]
6. Re-establish the 2 metre distance rule [why bother, you have closed all hospitality in point 1, a little more space between people in shops will not kill transmission]

https://www.independentsage.org/wp-content/uploads/2020/09/SR-emergency-plan-integrated-170920-amended-ES.pdf

 

[adc82140]

Looking at the figures for the last week or so, based on the swab date, not the reported date:

7/9- 3846
8/9- 3296
9/9- 3324
10/9- 3601
11/9- 3281
12/9- 2650
13/9- 2082
14/9- 3102

and for completeness, though these will increase a bit when results get returned

15/9- 2634
16/9- 2103
17/9- 167 (ignore this figure, too soon for most swabs to be tested)

Hardly running away exponentially. In fact a slight dip.

Note also these swab date figures de-duplicate as well- one person can only ever have one positive swab with these figures, no matter how often they are tested. For the daily tally they don't specifically say this (can anyone clarify?)

 

[Bletchleyite]

The trouble with the test figures is that at present they have effectively "hit the ceiling" so cases are "escaping" like they did in March.

 

[yorksrob]

I see the Independent Sage have released their '10 point plan' (one page of text in case you thought it was a report) to prevent a national lockdown. I hope nobody who matters is taking this lot seriously because it really sounds like nonsense - especially their IMMEDIATE MEASURES (!!). At least they have left transport alone for the most part but this just doesn't seem like serious stuff - is it just an excuse to pivot to the lockdown call?

The absence of mask rules is notable too. I imagine now that push has come to shove they don't actually trusts masks hence the desire just to close everything.

IMMEDIATE MEASURES (to be reviewed, as appropriate, when testing is functional and infections are back under control)
1. Close indoor service in pubs and restaurants [kill hospitality, ok.]
2. Work or study at home if possible. If impossible; don’t return to the workplace/University/College until certified COVID safe and provide financial support to employees in the interim [I guess this removes some travel but largely redundant as all of these locations are currently are Covid secure if open currently - if you want to close everything be honest]
3. Limit indoor social interaction to a bubble of three households [can't imagine this is really high impact and intent seems redundant in most part due to rule of 6]
4. Fund schools to allow smaller socially distanced classes and provide digital resources for all students and teachers, should they have to study and work at home [redundant due to point 2, has already happened anyway]
5. Create an emergency fund to support all those needing to shield or self-isolate (including those unable to get tested and their contacts) [SSP is bad so this has some merit]
6. Re-establish the 2 metre distance rule [why bother, you have closed all hospitality in point 1, a little more space between people in shops will not kill transmission]

https://www.independentsage.org/wp-content/uploads/2020/09/SR-emergency-plan-integrated-170920-amended-ES.pdf

I think I have more faith in ONION or STUFFING at the moment.

 

[Howardh]

Wonder if there are any statistics on this - I'm on my own at home with no-one coming or going, yet going out and about as safely as possible. Unless I have had Covid and been unaware, so far so good.
But in families, that were locked down with no visitors, now things are partly back to normal one member can go out to work/shop/pub and if they become infected, then they are simply around the other family members all the time, so tose members continually pick up the particles - thus getting inough to infect them badly.
So a single person going about their daily life comes into contact with covid out and about, but is never around long enough to get a dose enough to infect?
In the last six months I must have touched probably no more than three others. I've played hockey though and been in pyhsical contact with the opposition.
Meaning; is covid actually quite hard to catch, and you have to have a significant dose to become ill?
As I asked, I've no idea and never seen any stats on the domestic arrangements of those with covid and those (so far) who have stayed clear, can anyone help?

 

[Bletchleyite]

The "bubble of 3 households" thing probably makes sense and I think always did - basically merge households to be considered as one, with no interaction indoors in any setting outside of that. That does mean if it gets in it works its way through, but then it stops. And it allows things like support bubbles and childcare. In many ways better than the "rule of 6" because that allows one set of 6 now, then a different set of 6 in an hour - it's fully exclusive. So for instance my parents could "bubble" up with my sister's family and one more if there was one that made sense, but no other family could be included. 3 is a bit of a "magic number" as it allows a family with kids plus both sets of grandparents, or some other relevant combination.

Schools - the bubble thing is clearly not working. I'd go to half-time plus paper-based project work for now, perhaps except years 10 and 11. Half sized classes should be able to 1m distance with masks as mitigation.

Emergency fund - yes, that, definitely that. I read that only one in 5 is isolating properly if they tested positive (let alone close contacts) - this is absolutely key. So...
1. Pay anyone who is required to isolate either through a positive test or through T&T 80% of salary, bottoming out at minimum wage, with no cap, for those two weeks. Consider even going to 100% as there is no way you can choose to be isolated.
2. Enforce that they do isolate - random door knocks and phone calls to ensure they have. Fine of at least £1000 if they do not. If they live alone, food parcels to be supplied at zero cost.

 

[Tom B]

Indeed.

Proper invasive contact tracing is basically private investigation - you interview the person in detail, and then you go further - for instance, if you have record of them being in a supermarket you get the CCTV and see where they went and who they came close to, getting peoples' details via payment records etc if necessary, and so on. You need to find out everyone that person walked past, basically.

That type of contact tracing does stop spread. A "phone your mates" service doesn't.

It is possible for them to do this level of contact tracing if they desired. About a decade ago, the police rang me asking about a visit I'd made to a bank six months ago. I vaguely remembered it, purely because it wasn't a branch I often visited. It transpired that, in the lobby, I'd held the door open for a guy coming the other way - a couple of hours later, the guy had been murdered. The investigation had hit a dead end and they were tracing anybody with any slight contact with him before he died, and having viewed the CCTV got my details from the bank.

People who spout nonsense like that have no right to criticise Hancock as incompetent

Unfortunately, the court of public opinion is a strong force and a lot of government policy seems to try to pacify Karen offve Facebook.

Meaning; is covid actually quite hard to catch, and you have to have a significant dose to become ill?
As I asked, I've no idea and never seen any stats on the domestic arrangements of those with covid and those (so far) who have stayed clear, can anyone help?

I can provide only anecdotal evidence.

I had covid back in May. In the last few days prior to my confinement, there were probably half a dozen people who I had close contact with for >15mins, these being work colleagues. None took ill.

 

[Bletchleyite]

It is possible for them to do this level of contact tracing if they desired. About a decade ago, the police rang me asking about a visit I'd made to a bank six months ago. I vaguely remembered it, purely because it wasn't a branch I often visited. It transpired that, in the lobby, I'd held the door open for a guy coming the other way - a couple of hours later, the guy had been murdered. The investigation had hit a dead end and they were tracing anybody with any slight contact with him before he died, and having viewed the CCTV got my details from the bank.

Oh, I know it's possible, and it's how it would be done if someone got in with another nasty infectious disease that isn't endemic in the UK. Local authorities generally do it, and they do it quite well.

The problem is having the resources to do it for 4000 people a day and growing - that just isn't realistic. There aren't many murders in the UK (though surprisingly more than I thought) - about 650 last year, or roughly two per day. So if you had the daily caseload in low double figures you probably could viably contact trace that way. Indeed, that's what New Zealand has been doing when odd cases sprung up - they only jumped back to lockdown when they had an unexplained case that they couldn't trace.

As an aside, one benefit of the phone app thing is that it eases that sort of tracing.

 

[Megafuss]

What do people think is the reason for latest spike in cases. Since June and July face coverings have been mandatory of transport and in shops and we've had the 1 meter plus rule. We've went through August relatively steady and then bang!

My parents local authority claims that the data is telling them the rise in cases is due to more people mixing in homes, at grassroots sports events and social events. Why would they suddenly cause this high level of transmission now? Folk becoming careless?

 

[Bletchleyite]

My parents local authority claims that the data is telling them the rise in cases is due to more people mixing in homes, at grassroots sports events and social events. Why would they suddenly cause this high level of transmission now? Folk becoming careless?

I would imagine schools have to do with it - kids bringing asymptomatic but infectious cases home, and then them being spread around extended families. But generally people have been becoming lax at distancing, so that too.

The "rule of 30" was the case for several months, but came with guidance stricter than the "rule of 6", which said indoors only 2 households. That guidance was followed by most at the start, but had fallen wholly by the wayside in most people by the end. Even in pubs/restaurants "we will do you 2 tables of 6 right next to each other" had become common.

 

[Richard Scott]

What do people think is the reason for latest spike in cases. Since June and July face coverings have been mandatory of transport and in shops and we've had the 1 meter plus rule. We've went through August relatively steady and then bang!

My parents local authority claims that the data is telling them the rise in cases is due to more people mixing in homes, at grassroots sports events and social events. Why would they suddenly cause this high level of transmission now? Folk becoming careless?

It was likely to happen until enough people have had it. This doesn't seem to be the simple problem politicians don't understand.