Yeah, you'd think they could come up with something because, as you note, it would be massively beneficial.Arthur Dent wrote: ↑August 10 20, 12:32 pmI’d have to defer to people who know how to do good study design, but the problems you point out are basically the same ones faced in phase III vaccine trials. In those trials, some are vaccinated while other get a placebo, and they just go about their lives while you wait to see who happens to get infected through normal happenstance. I think the answer largely involves having to use a fairly large sample size to get enough statistical power. This is one of the reasons human challenge trials, where you deliberately infect vaccinated individuals, have been proposed.
The ethical case for a careful human challenge vaccine trial has some merit because the vaccine will already be proven to have a reasonable chance of working, and you can test (people you hope to be) low risk individuals while ensuring no further spread. Obviously, none of that applies starting an outbreak in the unprotected public.
Anyway, the costs of this disease running wild are enormous, so it’s hard for me to imagine the investment here would be too high even if the study requires tens of thousands of participants to work. The payoff of identifying a significant fraction of cases early would be enormous.
I'm not optimistic though which is why I kind of outlined the path of least resistance where these tests could at least be sold to the public just to get them out there. Let the public have access and if it gains momentum, which I have to think it would very quickly, then there would at least be pressure to recognize it and either governments would adopt it or at the very least studies could get going to prove the efficacy.
Echoing what I said about studies earlier though, I'm still not convinced you have access to a control group. Of course, I dont' know how to set up a study so maybe I'm making it too difficult. With your vaccine trials, presumably the mass amounts of people are similar with representative age, race, gender, location, etc included in both groups. So, for instance, you have 100 people from Austin in the control group and 100 people from Austin in the vaccine group and they're randomized such that all the demographics/races/genders/classes/etc are encapsulated. And, the data is collected and added to the overall data for all the other cities where the trial is underway.
But, you can't do that with a trial trying to compare the efficacy of certain testing strategies because, well, you inherently contaminate any control group the second the trial starts. If you unleash rapid testing in Austin giving everyone a test per day for the month of August, Austin can't be the control group because Austin in August even without rapid testing isn't the same as Austin in July, there will be some immunity due to the increased exposure that came about in July. Obviously, using a different area of the country or even state wouldn't make sense either as every community is on its own path with some areas that have been hit hardest already seeing massive reductions in cases.
In the vaccine trial taking the made up numbers I used above, it doesn't matter what's happening in Austin because you have 200 people living in the same environment, exposed to the same risks.
That's not to argue we shouldn't do it or anything like that, just food for thought on why its a challenge or what the challenges would be from what little I know about it. And why I'd push for no studies, cut the red tape, and let the companies sell these things already given they can show they detect the virus at somewhere around the 30 CT levels.
