Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We’re Testing the Wrong People

theatlantic.com

Post reply on HN
Print view View original

41–50 of 155 posts

Re: We’re Testing the Wrong People

#41

Post by yason »

alex_young wrote: post #3

The only data that is somewhat reliable in this pandemic is COVID deaths. Testing the population that is very sick / dying is crucial to understanding the baseline for the worst part of the problem. Once you have that under control, and we are not nearly there yet, you have to get every first responder tested regularly so you can prevent them from spreading the disease and give them some assurance that coming to work…

Deaths can lie too. What does it take for a death to be labelled as covid-19 death? That the person who died was merely tested positive, or do they do autopsies for all those who tested positive? How about covid-19 deaths that aren't counted as covid-19 deaths, for a reason or another? How about people dying less of regular influenza this year because of covid-19 and lockdowns?

Even looking back at monthly 2020 deaths in comparison to earlier years might not truly reveal the final impact of covid-19.

Re: We’re Testing the Wrong People

#42

Post by ctoth »

Maybe someone here can explain to me how this is supposed to work. Let's assume we have enough tests in the US to test every single person in the country every 3 days, and are contact tracing everyone with a smartphone, and providing smartphones to everyone without.

Now, with all this in place, isn't the best we can do to alert people after they have already been exposed to the virus?

It feels a bit like the safe sex crusaders switching to say everybody can have unprotected sex now, we'll let you know if you catch something, hopefully before you infect your next partner. Which I suppose is more reasonable for the group, but ... as the individual who is expected to actually go out in the world to consume or whatever, this doesn't feel like a viable solution. What am I missing?

Re: We’re Testing the Wrong People

#43

Post by beamatronic »

munificent wrote: post #30
bo1024 wrote: post #20

But the tests aren't perfect, so how to get actionable data from them isn't clear. We saw this with the flawed west coast studies. If a test has false positive or false negative rates even as high as 1%, and we run hundreds of thousands to millions of tests per day...that's thousands or tens of thousands wrongly quarantined for 14 days, and the same number who are contagious but told it's okay to go out.

We are currently quarantining about 300 million here in the US, so tens of thousands would be a massive improvement. The way to think about this is: who do we need to test so we can start un -quarantining people. The goal is not to find out who has it because right now, we are essentially behaving as if the entire US population is infected. The goal is to efficiently determine who is very unlikely to have it so that…

It's all about putting people into risk categories. Then each group's members can act in the optimal way for their group. Then retest/rebalance periodically.

Re: We’re Testing the Wrong People

#44

Post by godtoldmetodoit »

RandallBrown wrote: post #25

I've been saying this since the beginning of this whole thing. I don't understand why we haven't focused all of our testing on people that aren't sick so we can catch the asymptomatic cases that are out in the world spreading it to everyone.

I've been wondering why there hasn't been more programs to test representative samples of individuals, just like a political poll.

If I was in a Governor's or Mayor's shoes right now and I know there aren't enough resources to test everyone, I'd want some data like that to work with to make policy decisions.

While you are in the process of opening things back up, you keep a close eye on your daily/weekly representative sample numbers to see if anything needs to be dialed back. It gives you much closer to immediate feedback.

I suppose the immediate problem is getting a hold of a representative sample of people willing to be tested, but that feels solvable. I'm sure I'm not the first one to think of it, just curious why it isn't being done, as it feels immensely useful (maybe this is just the 538 political nerd in me wanting virus polls?).

Re: We’re Testing the Wrong People

#45

Post by dragonwriter »

No, we need to test everybody, so we can't be testing the wrong people (the reason we need to test sick people is different than the reasons we need to test other people, but they all need tested.)

But we don't have nearly enough testing capacity (either tests or the organized infrastructure to administer them, though both problems are being worked on) to meaningfully test the general population, whether or not we focussed our entire effort on the most important subset for reopening, so it would be extraordinarily stupidly premature to shift our current testing priority in that direction. This also means that general reopening is, a fortiori, extraordinarily stupidly premature.

Re: We’re Testing the Wrong People

#46

Post by eckesicle »

munificent wrote: post #32

> Another promising pathway is to pool many tests and run them together. If a pooled sample tests negative, everyone in the pool is negative. If it is positive, the members of the pool can be tested individually. Well, it took a global pandemic, but we finally found a use for bloom filters!

We have been working on productizing this for the past couple of weeks.

A bottle neck seems to be that it's not really possible to pool more than at most 32 patients at a time.

It's an early work in progress but have a look: https://www.pcr-pooling.com

Re: We’re Testing the Wrong People

#47

Post by dragonwriter »

ctoth wrote: post #42

Maybe someone here can explain to me how this is supposed to work. Let's assume we have enough tests in the US to test every single person in the country every 3 days, and are contact tracing everyone with a smartphone, and providing smartphones to everyone without. Now, with all this in place, isn't the best we can do to alert people after they have already been exposed to the virus? It feels a bit like the safe sex…

> Now, with all this in place, isn't the best we can do to alert people after they have already been exposed to the virus?

No, the best we can do in that case is treat COVID-19 like a dangerous communicable disease that isn't too widespread for surveillance and do specific quarantines of the infected and exposed, which then vastly limits the possibility that anyone else gets exposed, while not requiring shelter-in-place restrictions for those not already infected/exposed.

That's the whole point of disease surveillance with widespread testing and extensive contact tracing, and why it is key for a general reopening of the economy in the absence of an effective cure/vaccine.

Re: We’re Testing the Wrong People

#48

Post by jeremyjh »

dragonwriter wrote: post #45

No, we need to test everybody, so we can't be testing the wrong people (the reason we need to test sick people is different than the reasons we need to test other people, but they all need tested.) But we don't have nearly enough testing capacity (either tests or the organized infrastructure to administer them, though both problems are being worked on) to meaningfully test the general population, whether or not we fo…

The article explained its reasoning. Maybe you could explain yours. Specifically what is the value of testing very sick people when it cannot affect the treatment plan ?

Re: We’re Testing the Wrong People

#49

Post by heraclius »

ctoth wrote: post #42

Maybe someone here can explain to me how this is supposed to work. Let's assume we have enough tests in the US to test every single person in the country every 3 days, and are contact tracing everyone with a smartphone, and providing smartphones to everyone without. Now, with all this in place, isn't the best we can do to alert people after they have already been exposed to the virus? It feels a bit like the safe sex…

The target is to ensure that each case infects no more than one additional person (‘REach infected person will come into contact with some other people. Testing allows three things—

- first, we can tell this person to stay at home and so stop them from infecting more people;

- second, we can tell the people they have seen to stay at home with good contact-tracing, so that even if they get it they don’t spread it further; and

- third, with enough tests and contact tracing, those who were exposed but not infected don’t have to stay at home.

We are winning so long as RThe other way of reducing R requiring much less testing and contact tracing is a lockdown. This has obvious downsides. The point of a programme of testing and conyact tracing is that we can turn the cost of avoiding disaster from a lockdown to sticking some things up people’s noses.

edit: typo in first para

Re: We’re Testing the Wrong People

#50

Post by rckoepke »

markrages wrote: post #26
gumby wrote: post #12

Earlier quoted context omitted.

> The only data that is somewhat reliable in this pandemic is COVID deaths. I agree,, especially as you included the word “somewhat”. We have pretty good numbers on people who died in hospitals, but people who died at home appear to be seriously under counted. Additionally, consequential deaths are currently invisible. These are people who died, say, ecause the coronavirus crowded out their access to hospitals. There…

You can get part of that with excess death rates. From the weeks overall deaths subtract the average death rate for this week in past years. That should give both people who died at home and consequential deaths.

I'd reckon it's "doable" but not "easy", putting it in the realm of professional epidemiology. People are also dying less in car crashes, pollution-induced asthma, etc. There's quite a bit that may need to be accounted for. I suspect analyses will have decent error bars a year or so from now, but it's unlikely we'll get precise realtime estimates from this methodology. Would love to be proved wrong.
Post reply on HN