## The data you get when you go out and test people without symptoms is different from the data you get when you only test those who show up in a hospital

Saturday, May 9th, 2020

Testing residents of San Francisco’s Mission District for Covid-19 revealed that the data you get when you go out and test people without symptoms is different from the data you get when you only test those who show up in a hospital:

Slightly more than 2% of the people tested in the four square blocks currently have the coronavirus, for instance. Only 1 in 10 of them has a fever and most have no symptoms at all — which is to say that, absent the testing, they’d probably still be walking around and infecting people without knowing it.

There was more. Of the 981 white people tested, zero have the virus. Latinos are only 44% of the study but 95% of the positives. Men are a bit more than half of the sample but 75% of the positives. Just over half of the people tested say they’re unable to work from home — but that group registers 90% positive. Being in a crowded house appears to be a problem: nearly 30% of the positives come from households of more than five people, though those households make up only 15% of the population. Being poor is a bigger problem: people earning less than \$50,000 a year are 89% of the positives, though only 39% of the group studied. Only a quarter of the people with the virus have a primary care doctor. Six of the people with the virus still haven’t been located by the researchers, and informed that they’re ill.

1. Alistair says:

Big question here: how many of the asymptomatic will go on to be symptomatic? i.e. we’ve just caught them early. And how many will never show symptoms? This is why we really need a longitudinal study of cases.

The overall infection rate is low. Urgh. With street testing, could they have missed sick people who are isolated? With a long illness and recovery window you have to allow for that.

Otherwise…good. Nice to see a lot of the “commonsense” hypotheses being tentatively upheld.

2. Jack says:

What is hard about the concept of “representative sample”? Completely off topic but are presidential polls conducted this way, it would sure explain a lot.

3. Rishabh says:

It seems these 2% currently have the virus? I am guessing these were not antibody tests and not an indication of how many people have had the virus in the past. Is my understanding correct?

4. Andy says:

Why is there no consensus on what treatment to give if one is tested positive with such testing (as and when it will be done nationwide as being talked about )

5. Sam J. says:

I don’t believe these test are of much value at all. They have masses of false positives and sometimes they don’t identify accurately when someone does have corona.

The President of Tanzania sent in test of goats and the fruit paw paw and got positive result back.