Yes, you can do this.zsn wrote: ↑Fri Apr 10, 2020 4:24 pmI don't know if there are any epidemiologists on this board but we have some smart people who work in somewhat overlapping fields or may have had coffee with (or slept with) one. I was wondering about the lack of available test kits and had a thought about something the blood banks used to do during the early days of AIDS. Would "pooled testing" work in this case? For those who are not familiar with this concept (or know of it by another name) this is the way it works. It's not unlike the puzzle where you have a 1000 coins and one is defective and all you have is a two-pan balance; you are asked to find the defective one with a minimum number of weighings.
You combine a certain number of samples, let's say 25, and subject it to the test. If it tests negative then you haven't wasted 25 kits testing each one individually. If it tests positive you use this combinatorial matrix (some kid back in the day wrote an alogrithm for this) and combine subset of samples in a particular way and then test those. Typically you aliquot the samples initially so that you don't have to go back to the original samples. You then test these and based on the results of which subsamples come out positive you then run the deconvolution program and test a few final samples to determine the actual positives. These can be turned around quickly and will use a lot fewer test kits. The statistics of how many to pool etc are quite straightforward (for epidemiologists) and there can be a very high throughput. We use this in many other areas where quick yes/no answers are needed in short period and efficiently.
Granted that you will have to have a large sample size available simultaneously for this to work, but I believe that we are already there.
I saw one place was pooling tests. I forget where or what the source was.
They were doing pools of 10.
