Coronavirus

[Witton Park]

So, as we've said before in the current situation, false positives aren't as much of an issue as false negatives.

However, if we get to a situation of very low infection rates, but very high levels of screening, false positives become (relatively) more important.

NNNNNOOOOOOOOOOOOOOO!!

Take the Liverpool LFT which has showed 0.61% incidence. Just assume this is accurate and reflects the level of infection in the community.

With Stolly's googled figures you go out and test 10,000 people with the PCR test.

You should find 61 cases.

But you will get 56 that will be positive ie the 61 less 8% false negatives.

Then you will get another 80 dues to the 0.8% rate of False Positives on 9939 at 0.8%.

So instead of 61, you get 136. More than double.

If you look at the actual PCR figures, assuming LFT is accurate, the PCR test is running at more like 2.5% False Positive in the Liverpool area.
Maybe higher if the False Negative rate is more towards Mike's figure.

Of course that could be because the 0.8% refers to "in lab" but there are further errors creep in due to operator errors.

[Mike T]

False negatives 30%? Where do you get that from?

Anyway, you need context.

The 100,000 tests seem to be split.

Attachment 8865

New LFT Test (claimed to be 99.7% accurate) shows 80% fewer cases than standard PCR test (as used in Sedbergh and rest of UK)

So maybe this will help our decision makers get their act together, although I doubt it. They are too invested in the current track of lockdown until vaccine.

Those who are asymptomatic are having the "lateral flow" test. I have had a look around to try and find its false negative rate, and the commonest figure seemed to be 30%.

There is this, from the BMJ: " A preliminary evaluation from Porton Down and Oxford University9 throws little light on its performance in asymptomatic people or in the field. It suggests the test misses between one in two and one in four cases."

[Witton Park]

Those who are asymptomatic are having the "lateral flow" test. I have had a look around to try and find its false negative rate, and the commonest figure seemed to be 30%.

There is this, from the BMJ: " A preliminary evaluation from Porton Down and Oxford University9 throws little light on its performance in asymptomatic people or in the field. It suggests the test misses between one in two and one in four cases."

The theory that the PCR test is attended only by symptomatic folk is just wrong. It's certainly wrong in Blackburn as I know people who have booked on without symptoms over a period of 3 months.
I've asked some why they've done it. Just to know seems to be the main reason, although I suspect one or two of them fancied a couple of weeks off work.
One bloke at my wife's work had a test because his daughter had a runny nose.

Consider this Mike - you talk of the False Negative rate, with the False Positive rate of 0.8% you can go out and test 100,000 people.
You will get 800 positive tests.

Do you know what the top rate is at the moment?

It's Kingston upon Hull. The rate is 754 per 100,000.

A Covid free area could top the National Chart.

We can never escape this until they wake up from their trance.

[noel]

NNNNNOOOOOOOOOOOOOOO!!
...

You seem to be looking at this purely from the point of numbers. Having to self-isolate for two weeks is a minor inconvenience when compared with contracting a disease that kills 1% of people.

It's not solely about how large the numbers are, it's about how large the numbers are multiplied by the effect of those numbers.

To put it another way:
In situations/areas with high covid levels, mass screening leads to:
False negatives (relatively common) multiplied by a substantial number of people dying = very serious
False positives (more common than false negatives) multiplied by a substantial number needing to self isolate = significant hassle, but not as much of an issue as the impact of false negatives

Whereas:
In situations/areas with very low covid levels:
false negatives (less common, because fewer people have it) = less serious than above
false positives same as above and [relatively] more of a problem when viewed against very low numbers of false negatives.

Do you see what I mean? The conclusion I would draw is that mass testing is a good idea in areas of high covid infection.

[Wetherby whaler]

I understand that the infection mortality rate is around 0.05% among healthy people under 70.

[Witton Park]

You seem to be looking at this purely from the point of numbers. Having to self-isolate for two weeks is a minor inconvenience when compared with contracting a disease that kills 1% of people.

It's not solely about how large the numbers are, it's about how large the numbers are multiplied by the effect of those numbers.

To put it another way:
In situations/areas with high covid levels, mass screening leads to:
False negatives (relatively common) multiplied by a substantial number of people dying = very serious
False positives (more common than false negatives) multiplied by a substantial number needing to self isolate = significant hassle, but not as much of an issue as the impact of false negatives

Whereas:
In situations/areas with very low covid levels:
false negatives (less common, because fewer people have it) = less serious than above
false positives same as above and [relatively] more of a problem when viewed against very low numbers of false negatives.

Do you see what I mean? The conclusion I would draw is that mass testing is a good idea in areas of high covid infection.

I'm looking at it from what I perceive to be common sense.

But looking at your points.

It doesn't kill 1%.

The highest covid levels based on the current Government stats are in Kingston upon Hull as set out earlier. But it's a house built on sand.
How do we know?
Even in that area, they put it as 754 per 100,000 per week.
0.7% incidence according to the Govt figures based on PCR tests.

That is your area with high covid level. Less than 1% per week. It is a lot, but the dangers of mass screening with the PCR test are clear.

Hull might be clear. How will we know?

With the PCR test there is really no way of knowing with any degree of certainty which areas are high, moderate, low, or free of it.

[noel]

With the PCR test there is really no way of knowing with any degree of certainty which areas are high, moderate, low, or free of it.

I know you like a good argument WP, but I'm unsure what your overall point is. Are you against testing? against mass testing? or just pointing out that we need better tests? If the latter - I think we all agree.

[Fellbeast]

I know you like a good argument WP, but I'm unsure what your overall point is. Are you against testing? against mass testing? or just pointing out that we need better tests? If the latter - I think we all agree.

I was wondering that too?

My rough rule of thumb though is that Witton is typically for things I’m strongly against and against things I’m strongly for

[Mike T]

The theory that the PCR test is attended only by symptomatic folk is just wrong. It's certainly wrong in Blackburn as I know people who have booked on without symptoms over a period of 3 months.
I've asked some why they've done it. Just to know seems to be the main reason, although I suspect one or two of them fancied a couple of weeks off work.
One bloke at my wife's work had a test because his daughter had a runny nose.

Consider this Mike - you talk of the False Negative rate, with the False Positive rate of 0.8% you can go out and test 100,000 people.
You will get 800 positive tests.

Do you know what the top rate is at the moment?

It's Kingston upon Hull. The rate is 754 per 100,000.

A Covid free area could top the National Chart.

We can never escape this until they wake up from their trance.

We can discuss the false results/types of tests until the cows come home - this is the result - from yesterday - that really matters:

- Cases: 1,410,732 (+20,051)

- Deaths: 52,745 (+598)

That so little fuss is being made of these daily deaths is extraordinary.

[Witton Park]

We can discuss the false results/types of tests until the cows come home - this is the result - from yesterday - that really matters:

- Cases: 1,410,732 (+20,051)

- Deaths: 52,745 (+598)

That so little fuss is being made of these daily deaths is extraordinary.

You started it with the 30% False negative

So the cases were the lowest for 2 weeks.

As for the deaths, people die, 11,812 in England and Wales last week declared, week 46.

It's extraordinary that when it happened in 2015 and 2018 from flu, we didn't hear a dicky-bird.

https://www.pulsetoday.co.uk/news/ur...ths-last-year/