Read on Twitter
Nine months later I still cannot entirely wrap my head around the fact that getting a covid test is useful from a national statistical point of view but is much less useful to any one individual getting tested
We are at the point where if you think you have covid, you probably have covid. What’s more, you probably will not test positive for covid until after you show symptoms. And if you don’t have a severe case, the instruction is the same as 9 months ago: quarantine.
Since I got some pushback on this, I want to briefly lay out my reasoning and the people who know more than me can feel free to go to town on it. https://twitter.com/JakeAnbinder/status/1343286442181750786?s=20
There seems to be agreement that 5-6 days is the median incubation period, i.e. the time from infection to symptoms. But some studies seem to say it can be much longer, esp in older people. This one puts the 90th percentile for 55 year olds at 17 days. https://onlinelibrary.wiley.com/doi/full/10.1002/agm2.12137
At the same time, the consensus seems to be that people are most infectious around the day before or day of developing symptoms. https://www.nature.com/articles/s41591-020-1016-z
Yet this lit review says that during the course of infection, you will likely return a false negative every day *until the day you develop symptoms*. i.e. the idea of using a test to “catch” infection before you become symptomatic won’t often work. https://www.acpjournals.org/doi/10.7326/M20-1495
So it seems like if you had contact with someone infected, or even if you just wanted to see family or friends, you could very well test negative *while being contagious.*
And it seems like the only way to solve this is to take like ~half a dozen tests every couple days over at least a ~2 week period, and hope that some of them cover the span at the end of the incubation period and into the symptomatic period when you'd actually test positive.
But of course if you’re symptomatic, you shouldn’t need a test to tell you not to see people. And overall, there seems to be very little advantage offered by testing in this scenario that could not be better accomplished simply by quarantining.
So it seems like tests are useful but really only in a scenario of constant testing. And they don’t replace quarantining at all—in fact if you don’t substantially isolate yourself in the days in between your tests, the immediate subsequent tests become sort of worthless.
Yes but again only in the scenario where each person is being given constant testing. So, IRL, only people on college campuses and who work at hospitals. https://twitter.com/justjoshinyou13/status/1343428449499815943
I have never been tested so maybe they already do this, but it seems to me that the people administering the tests should try to warn people that the test is only effective if they take several more of them over the next couple weeks. https://twitter.com/Lydia_Beeyoobee/status/1343429694302126080
My sense is the public health people have been good at conveying that getting a test is not a license to act like there's no pandemic, but they've been less good at conveying that you might actually already be contagious on the same day you test negative https://twitter.com/xristopherlasch/status/1343432380468625408
This gets back to my original point: testing data is good for letting us make aggregate conclusions, but for most people who get tested, the manner in which they are using the testing infrastructure is, to them, virtually useless
And as far as I can tell, there has been very little effort to stop people from misusing the testing infrastructure in this way
You don't need to look very far to see proof of this. The colleges that are testing people constantly have had no big outbreaks. Outside of those places, the US is one of the top testing countries per capita, and yet also one of the most infected.
