Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #11729 > unrolled thread

Covid test regime algorithm

Back to article view | Back to comp.programming

Contents

  Covid test regime algorithm Malcolm McLean <malcolm.arthur.mclean@gmail.com> - 2021-01-08 04:54 -0800
    Re: Covid test regime algorithm David Brown <david.brown@hesbynett.no> - 2021-01-08 16:39 +0100
    Re: Covid test regime algorithm Richard Heathfield <rjh@cpax.org.uk> - 2021-01-08 16:03 +0000
      Re: Covid test regime algorithm David Brown <david.brown@hesbynett.no> - 2021-01-08 18:41 +0100
        Re: Covid test regime algorithm Richard Heathfield <rjh@cpax.org.uk> - 2021-01-08 19:01 +0000
          Re: Covid test regime algorithm David Brown <david.brown@hesbynett.no> - 2021-01-08 20:14 +0100

#11729 — Covid test regime algorithm

When testing large numbers of people for Covid, and expected cases are fairly low, you can reduce the number of test required by pooling samples. You then take the pools which test positive and do further tests to identify the posiitve patients. Pools which test negative are known to have no positive cases.

So the challenge is to devise an optimal pooling / testing regime.

Parameters to the model are the number of patients, the expected number of cases (actual cases will be Poisson distributed), the cost of a swab (one swb needs for each patient for each test), the cost of a test (incurred for each pool), the number of tests you can carry out in each "stage", and the cost of a "stage". (To explain, if you can carry out 500 tests in a stage, and your strategy is to test a pool of 1000, then test all individually, you will need three stages and 1001 tests. 1 test in the first stage, then you have to wait for the results, then a thousand tests, but you don't have capacity for that, only for 500, so you have to wait for 2 more stages until you get the result).

[toc] | [next] | [standalone]

#11730

On 08/01/2021 13:54, Malcolm McLean wrote:
> When testing large numbers of people for Covid, and expected cases
> are fairly low, you can reduce the number of test required by pooling
> samples. You then take the pools which test positive and do further
> tests to identify the posiitve patients. Pools which test negative
> are known to have no positive cases.
> 
> So the challenge is to devise an optimal pooling / testing regime.
> 
> Parameters to the model are the number of patients, the expected
> number of cases (actual cases will be Poisson distributed), the cost
> of a swab (one swb needs for each patient for each test), the cost of
> a test (incurred for each pool), the number of tests you can carry
> out in each "stage", and the cost of a "stage". (To explain, if you
> can carry out 500 tests in a stage, and your strategy is to test a
> pool of 1000, then test all individually, you will need three stages
> and 1001 tests. 1 test in the first stage, then you have to wait for
> the results, then a thousand tests, but you don't have capacity for
> that, only for 500, so you have to wait for 2 more stages until you
> get the result).
> 

With those numbers, it would surely then make sense to pool 500 in a
test, not 1000.

[toc] | [prev] | [next] | [standalone]

#11731

On 08/01/2021 12:54, Malcolm McLean wrote:
> When testing large numbers of people for Covid, and expected cases are fairly low, you can reduce the number of test required by pooling samples. You then take the pools which test positive and do further tests to identify the posiitve patients. Pools which test negative are known to have no positive cases.

With false negative rates of anything up to 25% (ref: 
<https://www.bmj.com/content/371/bmj.m4851/rr-0>), that's a somewhat 
optimistic claim.

-- 
Richard Heathfield
Email: rjh at cpax dot org dot uk
"Usenet is a strange place" - dmr 29 July 1999
Sig line 4 vacant - apply within

[toc] | [prev] | [next] | [standalone]

#11732

On 08/01/2021 17:03, Richard Heathfield wrote:
> On 08/01/2021 12:54, Malcolm McLean wrote:
>> When testing large numbers of people for Covid, and expected cases are
>> fairly low, you can reduce the number of test required by pooling
>> samples. You then take the pools which test positive and do further
>> tests to identify the posiitve patients. Pools which test negative are
>> known to have no positive cases.
> 
> With false negative rates of anything up to 25% (ref:
> <https://www.bmj.com/content/371/bmj.m4851/rr-0>), that's a somewhat
> optimistic claim.
> 

Aren't those false negative rates are for antibody tests (to see if you
have had it), rather than the tests for the virus proteins (to see if
you are carrying it now) ?  Several countries in Africa have been doing
combined samples since they started testing, with good effect when the
case rate is low.  I don't know if they are combining as many samples as
Malcolm suggested.

[toc] | [prev] | [next] | [standalone]

#11734

On 08/01/2021 17:41, David Brown wrote:
> On 08/01/2021 17:03, Richard Heathfield wrote:
>> On 08/01/2021 12:54, Malcolm McLean wrote:
>>> When testing large numbers of people for Covid, and expected cases are
>>> fairly low, you can reduce the number of test required by pooling
>>> samples. You then take the pools which test positive and do further
>>> tests to identify the posiitve patients. Pools which test negative are
>>> known to have no positive cases.
>>
>> With false negative rates of anything up to 25% (ref:
>> <https://www.bmj.com/content/371/bmj.m4851/rr-0>), that's a somewhat
>> optimistic claim.
>>
> 
> Aren't those false negative rates are for antibody tests (to see if you
> have had it), rather than the tests for the virus proteins (to see if
> you are carrying it now) ?

No.

"A false-negative case of severe acute respiratory syndrome coronavirus 
2 (SARS-CoV-2) infection is defined as a person with suspected infection 
and an initial negative result by reverse transcription-polymerase chain 
reaction (RT-PCR) test, with a positive result on a subsequent test."

<https://www.medrxiv.org/content/10.1101/2020.04.16.20066787v2>

-- 
Richard Heathfield
Email: rjh at cpax dot org dot uk
"Usenet is a strange place" - dmr 29 July 1999
Sig line 4 vacant - apply within

[toc] | [prev] | [next] | [standalone]

#11735

On 08/01/2021 20:01, Richard Heathfield wrote:
> On 08/01/2021 17:41, David Brown wrote:
>> On 08/01/2021 17:03, Richard Heathfield wrote:
>>> On 08/01/2021 12:54, Malcolm McLean wrote:
>>>> When testing large numbers of people for Covid, and expected cases are
>>>> fairly low, you can reduce the number of test required by pooling
>>>> samples. You then take the pools which test positive and do further
>>>> tests to identify the posiitve patients. Pools which test negative are
>>>> known to have no positive cases.
>>>
>>> With false negative rates of anything up to 25% (ref:
>>> <https://www.bmj.com/content/371/bmj.m4851/rr-0>), that's a somewhat
>>> optimistic claim.
>>>
>>
>> Aren't those false negative rates are for antibody tests (to see if you
>> have had it), rather than the tests for the virus proteins (to see if
>> you are carrying it now) ?
> 
> No.
> 
> "A false-negative case of severe acute respiratory syndrome coronavirus
> 2 (SARS-CoV-2) infection is defined as a person with suspected infection
> and an initial negative result by reverse transcription-polymerase chain
> reaction (RT-PCR) test, with a positive result on a subsequent test."
> 
> <https://www.medrxiv.org/content/10.1101/2020.04.16.20066787v2>
> 

Sorry - I was mixing up with the false positive rates.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.programming

csiph-web