Flawed as they are, AB tests can still be used effectively. Oliver’s piece was good for the casual pandemic viewer, but we need to think more deeply.
For starters, 50% false positive rate isn’t what they said; it was 50% of the positives in that particular location were false positives (the guy Oliver quoted actually said false positive rate of 50% but Oliver clarified, since an actual 50% false positive rate is lunacy). The rate of false positives as a proportion of positives is highly dependent upon the rate of true positives, and a rate of 50% isn’t alone enough to make the test worthless (for instance, if the population has a 0.5% true positive rate and the AB test gives 0.5% false positives, then false positives will be 50% of the positives, even though the test isn’t actually that inaccurate).
If you’re in a place where the true positive rate is 20%, then a 0.5% false positive rate is potentially palatable. Additionally, we can compare geographies with AB tests and glean extremely valuable insights about the spread of the disease even if the tests are flawed.
As long as we understand the limitations of the data we’re getting, all new data is useful. If we obtain mass AB testing and track the positives carefully, we should be able to pretty quickly draw conclusions about whether ABs confer immunity, which would be hugely valuable to know. And we can do this even with relatively shitty AB tests.
However, this requires competent administration of the testing, consistent populations, aggressive tracking, and synthesizing standardized data. I have no faith in the administration to do any of this. Maybe Canada or France can give it a shot?
Veterans Affairs has some nursing homes, usually (or maybe always) on the same campuses as their hospitals. I suspect that seeing as they’re federally run, deaths in those homes are likely reported to the VA, and not to the state that the home is in.
May also be the same for VA hospitals. Not sure on that, but I know I live near JBLM in Washington and there’s a hospital on base. I’m not sure if somebody dies there with CV if it’s reported to the state DoH or only federally…
Lol. Ok, I see it now. I tried to word that line specifically so it wouldnt sound that way. Fwiw nothing happened but the contact was closer than just casual person to person talking a few feet away.
I mean, if false positives are as rampant as they say - I dont want to give myself false confidence and then take a small risk like going golfing and getting sick because I’m dumb. It’s not that I want to know so I can go out and be buckwild, I just want the peace of mind because I have a very strong suspicion I may have actually contracted this.
The labs that are doing these tests seem reputable - but if the consensus here is that theyre crap i’ll just save my money.
The good news is that there’s been a lot of twitter pushback on the broad conclusions suggested in the article. But mostly of a science-y “stop drawing broad conclusions from studies esp ones that aren’t peer reviewed” variety.
My daughter in the Bronx, who’s pretty sure she had COVID-19, mentioned to me over the weekend that the local redi-care clinic now has AB tests. She was hoping to get over there this week.
Also, that number is laughable as well. We passed 70K yesterday. If we assume 2-3K deaths per day going forward, we will hit 134K by the end of MAY, not the end of July.
Did you want answers to the rest or just congrats on your sexy time?
The test is pretty much useless, my understanding is that if it’s positive it’s like 50-50 whether you actually had it. If it’s negative then it’s like 95% accurate.
I’m in at this point. Fuck it, let’s cancel federal taxes. Sorry, Kentucky, no red state bailout for you in the form of subsidies. Oh, look, New York’s state tax just went up by 15-20% and they’re still paying less in total than before AND they have more money being spent in the state? You hate to see it.