Interesting study about Sturgis and its effect on covid transmission. Red = with Sturgis (ie, an estimate of what really occurred); blue = “synthetic” line without Sturgis.
Note the cost they estimate, and this is if no one dies. The costs are “external”, meaning costs to society (eg, healthcare, hospitals, etc.), but many will be “internal” (eg, you stay at home for 2+ weeks because you are sick).
The whole study is here:
And scope is stunning, almost 20% of ALL new cases of covid in the US:
Certainly plausible - lots of people, mostly unexposed to the virus, disregarding social distancing, etc., and then returning home to populations mostly unexposed…
This kind of thing is going on all over, only on a smaller scale. Nobody cares any more. I live in a small tourist hub. If I ask people to respect the rules or at least stay away from me, they act like I’m a bitch, or they start ranting about masks and conspiracies. It’s a shame that while so many people have lost their businesses or jobs or even their lives because of this, others think it’s okay to travel around, congregate with strangers, and breathe on everyone.
The estimated costs come to over $26,500 per person that attended. That seems like quite the hyperbole to me.
I would be interested in actual numbers- how many attendees were seriously ill? Or went to the hospital? Or died (I did read one person died. He was over 60 with multiple underlying conditions)? They used cell phone data to figure out where people were from and where they went back to. Why not check on hospital admissions in those places. Otherwise, knowing what we know about the failures of our current testing system, it really doesn’t mean much.
What if we just picked a random 450,000 people that would not have otherwise been tested and test them? Case numbers would jump then too I bet.
What else was going on at that time of year? School started for one. I can guarantee you that somewhere someone that doesn’t want school to open is using those same numbers to justify their position. Since all we have are fuzzy numbers from using ‘a synthetic control approach’ we have no idea if those new infections are 45 year old bikers or 4th graders.
It’s the butterfly effect. The costs cascade forever. Technically, the amount of money increases forever as time goes on.
I love math models. The way I read it, they used phone data to develop a model to track people around Sturgis, and then developed a model to correlate the phone model to the CDC test data.
The developed a “Synthetic control”? Does that mean we don’t need Randomized Controlled Tests any more? Is this the new gold standard? No pesky, expensive, control groups, just create one in the computer. Riii…ght. IOW, we have no reason to know that the blue curves wouldn’t have become the red curves without Sturgis happening at all because it’s not a controlled experiment. It’s association.
I think you’re exactly right @gme and would bow in your direction if I knew what direction that was.
I would make that last line “45 year old bikers, 4th graders or false positives.”
A statistician whom I read points out that the word case probably doesn’t mean what most people think. When we read “case” we think of someone who got the disease; what it almost always means in this context is someone who has tested positive with one of the available tests. I went through the paper looking to see what they meant and it doesn’t mention which tests were used. They say they used CDC data which could be several tests (as I understand it). Some tests have false positive rates around 90% (some of the Polymerase Chain Reaction, PCR, tests). It’s pretty much impossible to know what really went on.
I’m suspicious, but I worked in a field where the best and most expensive models got you close enough to make things work with enough blood, sweat and tears.
