HN Explorer

Ask HN: Why can't corona testing be done in bulk?

2 points by sriram_malhar ↗ HN
I have often wondered why it isn't possible to combine swabs from large numbers of people and PCR the soup to check for the presence of a virus. In the best case, one can eliminate large numbers (dorms, hospital janitorial and nursing staff) at one shot, and makes frequent (daily) testing possible for people at risk). If the sample tests positive, then perhaps (perhaps..) it is possible to figure out the affected subset in a log number of further tests. Am I missing something?

3 comments

[ 5.0 ms ] story [ 26.5 ms ] thread
rolph
more samples together increases offchance of crossover contamination producing a false indication.

under the assumption that a corona virus case displays symptoms around day 5 it doesnt make any sense to test if the results come through at the same time as symptoms would arise in positive case.

in my area there is a claim of 3 days shipping of a sample before tests are performed and results submitted.

if you have a lot of samples, abunch of 96 , and 48 well sample trays then bulk testing is efficient if you are careful. If you have to wait, to accumulate samples for a batch, that is not soon enough.

sriram_malhar
Not sure i understand the issue of crossover contamination. If you mix together say 500 swabs and if no one has the virus, then will there be a false positive? My understanding is that there won't be. And you have saved 500 people-hours of testing time in one go.

I was thinking about this when our university was closed and upwards of ten thousand people were told to go home. It is good for them and for everyone else on public transport to get a clean chit.

I could be wrong, but as far as I know it doesn't take 3 days to do a test. It takes that much time because of shipping time and demand on the machines and crew. Doing bulk testing should alleviate the latter.

rolph
ahh you are talking about taking all samples and combining them, this would destroy the ability to trace transmission across reported contacts. i takes ~3hours to run the test if your very familiar with the procedures the problem with centralized testing is the delay if samples are shipped and handled from distributed locations.

the process involves a proceedure that is very sensative to initial sample frequencies. if there is a lot of negative samples mixed with a small number of positive samples there is a pretty good chance that when you remove an aliquot you will sample a population of samples and have artefact.

the chance of amplifying non targeted RNA [human] is increased in this way thus increasing chance of false negatives.

cross over is about handling series of samples. the more samples on the bench at once the more chance to incidentally combine them, from things like "tip drip" or spatter.