/sci/ - Science & Math » Thread #11989192

522KiB, 2048x1957, 650C4741-9F6E-4AC4-8760-F47EE5E2704F.jpg

View Same Google iqdb SauceNAO

Sample size

Anonymous Sun 09 Aug 04:35:15 2020 No.11989192 View Reply Original Report

Quoted By: >>11989250

30 minute covid test with no lab equipment out of gynecology, Columbia university. Sample size: 30. Result: 97% sensitivity.

Would that 97% reliably hold with a larger sample size or does 30 tested people tell us nothing?

Anonymous

Anonymous Sun 09 Aug 2020 04:40:51 No.11989205 Report

Quoted By: >>11989233 >>11989237

Good job posting the methodology or any link you fucking mong

Anonymous

Anonymous Sun 09 Aug 2020 04:40:56 No.11989206 Report

Quoted By:

I don't fucking know, test more. Make a 100 people test. There's so much fucking people infected you can grab 50 infected assholes outta any home or recovery center

Anonymous

Anonymous Sun 09 Aug 2020 04:50:48 No.11989233 Report

Quoted By:

>>11989205
This is the most specific I could find...does it look valid or too small?

Anonymous

Anonymous Sun 09 Aug 2020 04:52:49 No.11989237 Report

Quoted By: >>11989292 >>11989311

>>11989205
>https://www.medrxiv.org/content/10.1101/2020.06.13.20129841v1

There’s the article

Anonymous

Anonymous Sun 09 Aug 2020 04:57:16 No.11989250 Report

Quoted By:

>>11989192
Does nobody know whether small sample sizes are useful?

Anonymous

Anonymous Sun 09 Aug 2020 05:13:51 No.11989292 Report

Quoted By: >>11989305

>>11989237
Preprints are for jews.

LAMP is already a well-established method. It doesn't solve the actual problem with testing right now, which is reagent supply. They also tested primarily with contrived samples. My concern is that they didn't do very strong concordance studies against the reference method, instead just taking a handful of samples, less than 20. That's not really useful.

It seems like a good proof of concept, but I highly doubt it's going to hold up over several hundred samples. Missing a positive contrived sample is pretty worrying, when we validated both of our current tests we used 50 contrived positives and 100 negatives with 100% accuracy.

Anonymous

Anonymous Sun 09 Aug 2020 05:18:21 No.11989305 Report

Quoted By:

>>11989292
And didn’t those tests end up having less than 80% accuracy when scaled up and put in use?

Anonymous

Anonymous Sun 09 Aug 2020 05:19:26 No.11989311 Report

Quoted By:

>>11989237
> https://www.cuimc.columbia.edu/news/one-two-three-spit-covid-19-test-born-columbia-fertility-clinic

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

Your latest searches