Hacker News new | ask | show | jobs
by hnrodey 2019 days ago
The article says in San Mateo county there are an estimated 8000-15000 contagious people.

That’s several times higher than how many positive test results have been logged in the past 14 days.

https://thecovidcomplex.com/state/california/san-mateo/summa...

Disclosure: this is my website based on data from the New York Times.

3 comments

It’s important to distinguish between true case burden and diagnosed case burden.

Testing is prone to variance due to behavior, test availability, development of symptoms, contact tracing effectiveness, etc... I suspect their modeling efforts use a more principled measure than positive tests.

Without even seeing the data I would infer among all tests a high positivity rate (certainly double digit).

You'd be incorrect in that inference. According to state data, the positivity rate in San Mateo County is 2.6% currently.
Ah, thanks for checking. There are a few other hypotheses then:

Testing volumes are low, meaning we fail to capture a sizeable chunk of cases.

The testing strategy is bad. We’re testing people that are less likely to be positive.

And possibly, the model is just wrong.

I think it is a clear implications we still don't have enough testing even though CA as a whole is doing relatively well on testing. It seems like once your positive % goes above 5-10% it really means you're missing many cases.
Colorado is telling us that 2.5% of the state population is actively contagious. I don’t know the state number of positive cases off the top of my head but many counties I look at around right around 1% or less of positive cases / county population.

How can Colorado estimate at roughly 2.5x and San Mateo county is 4-7x? I’m shocked that the variance can be so great between counties.

Likely reflective of how little we really know. It’s also not clear from Morrow’s note where his estimate comes from.
Not sure if this is still the case, but in the past multiple people from one household / coming to the drive-through testing place together were not get all checked. It was "one of you has symptoms, you all isolate; one of you with symptoms tests positive, you're all positive". (but not recorded positive) Which means the true cases number is at least tests x avg-people-per-household + asymptomatic people not tested + ...
There’s something that should be said about people that are contagious yet should be in quarantine meaning they’re no longer actually infecting people.

Instead we get BS numbers that make it seem like every person that is contagious is still running around.