Possible 2nd lockdown for Leicester?

[Chrysalis]

On 17/07/2020 at 01:24, WigstonWanderer said:

TBH I was just confirming for my own benefit that I had understood what @Fktf had written earlier as prior to that I had been rather confused as to exactly what metric was being quoted. I wasn’t trying to add to the base of information about the stats themselves, so apologies if I have introduced any confusion.

 

Nevertheless, the argument still holds. If (as you say) the hit rate (positive tests per 100k) is still 20% higher than elsewhere despite doing many more tests, then you do have a more significant problem. You would expect the wider sample in Leicester, which is diluted with a load of asymptomatic people, to have a lower hit rate.
 

I agree that in these circumstances (many more tests) the positive rate per 100k population cannot be compared with elsewhere. In any case, any statistician worth his salt should be able to correct for different rates I should have thought. Whether the government are playing any political games or not I have no idea.

 

I think the hit rate has reduced.

 

The hit per 100k population is not a rate its a count, obviously if you test more people you find more cases.  That number would only ever go up if you test more people, unless the actual hit rate goes down enough to cancel it out.

Edited 18 July 2020 by Chrysalis

[Chrysalis]

This is a wtf

 

https://www.leicestermercury.co.uk/news/leicester-news/leicester-mayor-ask-again-most-4338269

 

Quote

Health Secretary Matt Hancock told Parliament that he had offered Sir Peter the chance to define part of the lockdown zone within the city on Thursday afternoon but the mayor had declined.

In response, Sir Peter said: "He was seeking in a Skype conversation to draw lines on the map.

 

"That's not the way to do businesses.

 

[st albans fox]

2 hours ago, Chrysalis said:

 

I think the hit rate has reduced.

 

The hit per 100k population is not a rate its a count, obviously if you test more people you find more cases.  That number would only ever go up if you test more people, unless the actual hit rate goes down enough to cancel it out.

But the rate per 100k adjusts with the number of tests you undertake ....... surely if you test 10k people and find 20 positives then that 200 per 100k ??  Is that not how it works ??? 

[Paninistickers]

2 hours ago, Chrysalis said:

What is people's opinion of these new powers coming into law next week, do you think they will be used in the Leicester lockdown?

 

If they are applied then ironically the media portraying the lockdown as been eased could be very misleading.

 

As last week one could travel, no real enforcement, potentially next week one inside the lockdown might be under a stay in home order, or unable to travel to places like Glenfield as that would now be counted as out of Leicester.

It's grotesque. 

 

I am worried how meekly the public are accepting a totalitarian state under the pre-text of 'for the greater good' 

 

I posted a while back that govt strategy should perhaps begin to shift towards normalising covid and taking away some of the fear factor (as, let's remember, most people are extremely unlikely to die from this) ....which would have a direct positive effect on the economy, mental health and basic freedom. 

 

Instead, it looks like govt have got drunk.on power. 

 

To think the average Joe was terrified of Jeremy Corbyn being PM (I couldn't stand Corbyn btw) ....yet this has all got out of control

 

 

 

 

 

 

[stix]

19 minutes ago, st albans fox said:

But the rate per 100k adjusts with the number of tests you undertake ....... surely if you test 10k people and find 20 positives then that 200 per 100k ??  Is that not how it works ??? 

Yeah, but it’s guessing abit though isn’t it. Like I’ve said previously, if other major cities in England had been tested at the levels that Leicester has over the last three weeks, I think it’d be a different matter. What the government don’t want is for many parts of the country going back into the localised lockdown that Leicester is having to endure. Devastating for an already crippled economy. 

[st albans fox]

8 minutes ago, stix said:

Yeah, but it’s guessing abit though isn’t it. Like I’ve said previously, if other major cities in England had been tested at the levels that Leicester has over the last three weeks, I think it’d be a different matter. What the government don’t want is for many parts of the country going back into the localised lockdown that Leicester is having to endure. Devastating for an already crippled economy. 

The number of infected people per 100k is going to reflect the density of infected people ! 
 

in cities where there aren’t so many infected, the rate per 100k will be low. It doesn’t matter how many you test, the rate will reflect the infections ..... the only difference I can see in Leicester is the encouragement for people without symptoms to get tested but even then the percentage rate isn’t being skewed. 

 

 

 

[Chrysalis]

2 hours ago, st albans fox said:

But the rate per 100k adjusts with the number of tests you undertake ....... surely if you test 10k people and find 20 positives then that 200 per 100k ??  Is that not how it works ??? 

its actually the number of currently confirmed active cases, so once you confirmed you stay on there until you test negative.  If you dont get a repeat test, I dont know if they assume you negative two weeks after a positive, as logically there would be some kind of timeout.

 

So lets say you do 1000 tests and 30% are positive, then that is 300 cases for 2 or so weeks.  If the population is 300k (leics is 330k), then I think the number would be divided by 33% so those 300 positives would be changed to 100 per 100k.  They would stay until the person either dies, tests negative (recovers), and maybe also if it just expires (no further test taken).

 

As far as I know there is no adjustments to account for amount of testing, what makes me think that is the data in the PHE document here.  The author even states, the numbers are high due to the testing done, which also suggests no adjustment is made.  If we actually had a spread rate of 3x the 2nd highest city, the R rating for Leicester would be way higher than it is in the document.

 

https://www.gov.uk/government/publications/covid-19-exceedances-in-leicester

 

Also there is a grading colour system, green. amber and red.  On July 14th, Leicester was Green.  I am not sure how that system works, but I think its graded vs a some kind of predicted case count, Green is when you doing better than expectations I think.

Edited 18 July 2020 by Chrysalis

[Chrysalis]

2 hours ago, st albans fox said:

The number of infected people per 100k is going to reflect the density of infected people ! 
 

in cities where there aren’t so many infected, the rate per 100k will be low. It doesn’t matter how many you test, the rate will reflect the infections ..... the only difference I can see in Leicester is the encouragement for people without symptoms to get tested but even then the percentage rate isn’t being skewed. 

 

 

 

What would the rate per 100k be if you only test one person in a heavily infected city?  I struggle to understand how you dont think the amount of testing manipulates the figures.

 

So here is an example (30% infection, 300k population)

Test 1000 people, 300 positive tests, figure is 100 per 100k.

Test 100 people, 30 positive tests, figure is 10 per 100k.

Edited 18 July 2020 by Chrysalis

[st albans fox]

3 minutes ago, Chrysalis said:

What would the rate per 100k be if you only test one person in a heavily infected city?  I struggle to understand how you dont think the amount of testing manipulates the figures.

You can’t measure a rate from a small sample ........you can only extrapolate when the sample is large enough to smooth out any oddities  ....

 

I appreciate that the more you test the more positive cases you find .....But you also find many more negative cases too .....

[Chrysalis]

I accept the accuracy argument, but ultimately you will find less positives.  Even if everyone was positive, the maximum case load would be the amount of tests carried out.

 

So going back to my example.  Even if somehow you fluked an unbalanced result, and on the smaller sample you had double infection rate, the per 100k population figure would still be one fifth of the larger sample. 

 

So here is an example (30% infection, 300k population)

Test 1000 people, 300 positive tests, figure is 100 per 100k.

Test 100 people, 60 positive tests, figure is 20 per 100k.

 

The negative test results are seemingly ignored by the criteria they using, so they are moot.  They mainly using active cases per 100k population criteria.

Edited 18 July 2020 by Chrysalis

[WigstonWanderer]

2 hours ago, Chrysalis said:

its actually the number of currently confirmed active cases, so once you confirmed you stay on there until you test negative.  If you dont get a repeat test, I dont know if they assume you negative two weeks after a positive, as logically there would be some kind of timeout.

 

So lets say you do 1000 tests and 30% are positive, then that is 300 cases for 2 or so weeks.  If the population is 300k (leics is 330k), then I think the number would be divided by 33% so those 300 positives would be changed to 100 per 100k.  They would stay until the person either dies, tests negative (recovers), and maybe also if it just expires (no further test taken).

 

As far as I know there is no adjustments to account for amount of testing, what makes me think that is the data in the PHE document here.  The author even states, the numbers are high due to the testing done, which also suggests no adjustment is made.  If we actually had a spread rate of 3x the 2nd highest city, the R rating for Leicester would be way higher than it is in the document.

 

https://www.gov.uk/government/publications/covid-19-exceedances-in-leicester

 

Also there is a grading colour system, green. amber and red.  On July 14th, Leicester was Green.  I am not sure how that system works, but I think its graded vs a some kind of predicted case count, Green is when you doing better than expectations I think.

Perhaps I’m misunderstanding something here, but on page 7 of the document you link, figure 1.3, the graphs shown are cases per 100 tests. Leicester shows a significantly higher percentage of positives than Barnsley and Bradford and is rising whereas the other 2 are falling.

 

Just to add, 30 or 40 positives per 100 tests is huge. We’re going nuts over here based on a case rate of less than 1 positive per 100 tests in the current Melbourne outbreak.

Edited 19 July 2020 by WigstonWanderer

[Guest Fktf]

3 hours ago, Chrysalis said:

its actually the number of currently confirmed active cases, so once you confirmed you stay on there until you test negative.  If you dont get a repeat test, I dont know if they assume you negative two weeks after a positive, as logically there would be some kind of timeout.

 

So lets say you do 1000 tests and 30% are positive, then that is 300 cases for 2 or so weeks.  If the population is 300k (leics is 330k), then I think the number would be divided by 33% so those 300 positives would be changed to 100 per 100k.  They would stay until the person either dies, tests negative (recovers), and maybe also if it just expires (no further test taken).

 

As far as I know there is no adjustments to account for amount of testing, what makes me think that is the data in the PHE document here.  The author even states, the numbers are high due to the testing done, which also suggests no adjustment is made.  If we actually had a spread rate of 3x the 2nd highest city, the R rating for Leicester would be way higher than it is in the document.

 

https://www.gov.uk/government/publications/covid-19-exceedances-in-leicester

 

Also there is a grading colour system, green. amber and red.  On July 14th, Leicester was Green.  I am not sure how that system works, but I think its graded vs a some kind of predicted case count, Green is when you doing better than expectations I think.

You are cherry picking. The author also states: "The proportion of positive PCR tests (as a proportion of all test) is rising. This is suggestive of a genuine increase in numbers of new infections, not simply an artefact of increasing test rates".

Edited 19 July 2020 by Fktf

[Guest Fktf]

6 hours ago, st albans fox said:

The number of infected people per 100k is going to reflect the density of infected people ! 
 

in cities where there aren’t so many infected, the rate per 100k will be low. It doesn’t matter how many you test, the rate will reflect the infections ..... the only difference I can see in Leicester is the encouragement for people without symptoms to get tested but even then the percentage rate isn’t being skewed. 

 

 

 

Technically it is - but in a way that works against finding 'bad news'. Or you could say, in a way that would prove there isn't a problem here.

 

Test a bunch of people with no symptoms - given the rarity of the disease - you should get a bunch of negative returns. If you test indiscriminately, but keep recording a high rate of infection - it is about the strongest evidence you can find of their being.... a high rate of infection

 

 But I know you already know this..

[Guest Fktf]

6 hours ago, Paninistickers said:

It's grotesque. 

 

I am worried how meekly the public are accepting a totalitarian state under the pre-text of 'for the greater good' 

 

I posted a while back that govt strategy should perhaps begin to shift towards normalising covid and taking away some of the fear factor (as, let's remember, most people are extremely unlikely to die from this) ....which would have a direct positive effect on the economy, mental health and basic freedom. 

 

Instead, it looks like govt have got drunk.on power. 

 

To think the average Joe was terrified of Jeremy Corbyn being PM (I couldn't stand Corbyn btw) ....yet this has all got out of control

 

 

 

 

 

 

You say this knowing that it is only true in the sense of 'the greater good' - right? Your argument seems to be 'get things functioning again, because for the people that don't die it will be better'.

 

 

[Chrysalis]

2 hours ago, Fktf said:

You are cherry picking. The author also states: "The proportion of positive PCR tests (as a proportion of all test) is rising. This is suggestive of a genuine increase in numbers of new infections, not simply an artefact of increasing test rates".

I am not, I already mentioned that.

 

I didnt mention that in the newer posts simply because the government are not mentioning it either.

 

The government are concentrating on the per 100k figures, of which that comment has nothing to do with, that comment is referring to the cases per test figures.

[Chrysalis]

3 hours ago, WigstonWanderer said:

Perhaps I’m misunderstanding something here, but on page 7 of the document you link, figure 1.3, the graphs shown are cases per 100 tests. Leicester shows a significantly higher percentage of positives than Barnsley and Bradford and is rising whereas the other 2 are falling.

 

Just to add, 30 or 40 positives per 100 tests is huge. We’re going nuts over here based on a case rate of less than 1 positive per 100 tests in the current Melbourne outbreak.

Are they 300% higher though? no.

 

I never said Leicester wasnt worse than other areas, I am merely stating the case rate is not 300% as the media are reporting.  Which this report shows.

[Chrysalis]

3 hours ago, WigstonWanderer said:

Perhaps I’m misunderstanding something here, but on page 7 of the document you link, figure 1.3, the graphs shown are cases per 100 tests. Leicester shows a significantly higher percentage of positives than Barnsley and Bradford and is rising whereas the other 2 are falling.

 

Just to add, 30 or 40 positives per 100 tests is huge. We’re going nuts over here based on a case rate of less than 1 positive per 100 tests in the current Melbourne outbreak.

The problem is the test count or same data is not available for other cities, so there is no comparison, but if I am wrong and you find the data for one of the other cities been compared to Leicester then please provide the data, without a comparison for the per 100 tests we just dont know how bad or good those figures are.  Bradford however is shown in the document and has a circa 20% average vs a circa 30% average for Leicester, that's a 33% difference, not a 300% difference.

 

Boris said he wont lock down cities because of other cities, in other words he is saying he wont use a sledgehammer to kill a fly, yet wards are been locked that have less cases than Glenfield.  So he is been inconsistent.

[Chrysalis]

8 hours ago, Paninistickers said:

It's grotesque. 

 

I am worried how meekly the public are accepting a totalitarian state under the pre-text of 'for the greater good' 

 

I posted a while back that govt strategy should perhaps begin to shift towards normalising covid and taking away some of the fear factor (as, let's remember, most people are extremely unlikely to die from this) ....which would have a direct positive effect on the economy, mental health and basic freedom. 

 

Instead, it looks like govt have got drunk.on power. 

 

To think the average Joe was terrified of Jeremy Corbyn being PM (I couldn't stand Corbyn btw) ....yet this has all got out of control

 

 

 

 

 

 

 

The reason is because most people in this country only worry about #1, a lot of the posts in this thread are questions like "am I affected", "what does it mean for me" etc.

 

If the national lockdown lasted 5 months and powers were been introduced to enforce no travel and house arrest people, the public would be up in arms.

 

I dont agree with you about downplaying covid, its a horrible disease, my issue really is the handling of the problem, the misleading press reports, and the lack of a coherent plan.  There has been no well thought out plan announced by government or the council, the mayor is just taking pot shots at the government, the government taking some back, the only actions I see is a huge jump in testing (which by itself doesnt do anything, its data collecting), and some plans to prevent people from leaving the city and maybe their homes.  There is clear political games been played by both sides, the lines are clearly political.  I still think its a social experiment.

 

I am also left wondering what the mayor has to gain by refusing the invitation to redraw the lines.  I expect he just wants the city centre open, and wasnt interested in removing some suburbs only, as looking at his track record he only cares about the city centre.

[WigstonWanderer]

39 minutes ago, Chrysalis said:

The problem is the test count or same data is not available for other cities, so there is no comparison, but if I am wrong and you find the data for one of the other cities been compared to Leicester then please provide the data, without a comparison for the per 100 tests we just dont know how bad or good those figures are.  Bradford however is shown in the document and has a circa 20% average vs a circa 30% average for Leicester, that's a 33% difference, not a 300% difference.

 

Boris said he wont lock down cities because of other cities, in other words he is saying he wont use a sledgehammer to kill a fly, yet wards are been locked that have less cases than Glenfield.  So he is been inconsistent.

But those graphs are per 100 tests?
 

They are all over the place, but the last black dots show Leicester at about 30%, and Bradford at about 10%, so you could certainly make a case for Leicester being 3 times the rate of Bradford. The Barnsley rate is even less. Probably more important is the trend which is up for Leicester but down for the others.

 

Must say it’s nice taking time off from being an armchair football manager to be an armchair epidemiologist 

Edited 19 July 2020 by WigstonWanderer

[WigstonWanderer]

Here's the page in the paper you posted that I'm referring to:

 

 

The time-series plots for Leicester based on cases per 100 tests and number of tests per day is presented below. In contrast to a decreasing trend noted in the vast majority of other UTLAs in East Midlands and England, a gradual increasing trend has been noted in Leicester in the last few weeks. 

Figure 1.3. Time series plot of cases per 100 tests and number of tests per day, Leicester and other UTLs

 

The red line is the 99% threshold, the blue line is the expected or average number of cases per 100 tests each day which is predicted from the regression model. The grey line and points are the observed number of cases per 100 tests in the baseline period, i.e. the 6 weeks of data that are used in the exceedance algorithms regression model. The black line and points are the observed number of cases per 100 tests in the 10-day investigation period. The yellow points are the observed number of cases per 100 tests in the last 4 which are subject reporting delay so ignored. The green dashed line is the total number of tests (pillar 1 and pillar 2 tests combined) each day in the LTLA. A rough estimate of the number of cases each day can be obtained by multiplying the observed cases per 100 tests each day by the number of tests each day, and dividing by 100.

 

[CosbehFox]

Too late now for Leicester but Hancock has made an u turn on providing councils with precise data. Previously councils received only numbers of infections to post code. Which could be misleading in the extreme. Now they will receive it to the precise address. 

[Lionator]

9 minutes ago, Cardiff_Fox said:

Too late now for Leicester but Hancock has made an u turn on providing councils with precise data. Previously councils received only numbers of infections to post code. Which could be misleading in the extreme. Now they will receive it to the precise address. 

So it was Hancock and Dorries who were lying all along regarding data and not Soulsby, why should we believe them on anything?

[Lionator]

8 hours ago, Chrysalis said:

What would the rate per 100k be if you only test one person in a heavily infected city?  I struggle to understand how you dont think the amount of testing manipulates the figures.

 

So here is an example (30% infection, 300k population)

Test 1000 people, 300 positive tests, figure is 100 per 100k.

Test 100 people, 30 positive tests, figure is 10 per 100k.

Test 100 people, 30 positives, that would actually be 30000 per 100k because you're multiplying by 1000.

Test 1000 people, 300 positives is also 30000 per 100k because you're only multiplying by 100?

 

This is why confidence intervals exist. In most science it's a 95% confidence interval ie. you can be 95% certain that the true number is within that range.

 

Now, the larger the sample, the smaller the confidence interval which means that 95% certainty has a much smaller range. For example;

 

Test 100 people with 30 positives, your 95% CI for the true number might be 5,000 to 55,000 per 100k because it's such a low sample size. You can practically discount this data, it tells us nothing.

Test 10000 people with 3000 positives and it's likely to be 28,000 to 32,000 per 100k because your sample is larger therefore much more representative of the true population. 

[CosbehFox]

12 minutes ago, Lionator said:

So it was Hancock and Dorries who were lying all along regarding data and not Soulsby, why should we believe them on anything?

Councils were receiving data but not to the level they wished. Apparently even now the data is essentially a week old when they receive it 

[Guest Fktf]

5 hours ago, Chrysalis said:

I am not, I already mentioned that.

 

I didnt mention that in the newer posts simply because the government are not mentioning it either.

 

The government are concentrating on the per 100k figures, of which that comment has nothing to do with, that comment is referring to the cases per test figures.

So you're saying there is a problem in Leicester because cases are rising, but the stats that are being reported to the public to demonstrate this could be improved?