Things that went whizzing by that I want to remember: So Friday morning I woke up with the sniffles in Carbon County, Montana—east northeast of the Yellowstone-Grand Teton National Park complex. No fever. Full senses of smell and taste. No coughing or perceived shortness of breath (after accounting for the altitude: I was at 5,500 feet instead of the sea level I am used to).... Hope is not a plan. The Delta plague wave will brush relatively lightly past Teton County, Wyoming. I fear what it will do to Carbon County, Montana.
The Canadian study done in long-term care homes found full vaccine efficacy to last five months. The CDC is getting six to nine months on a more general population.
It really is a coronavirus; the Uncommon Cold. You get a cold every year. There's no reason to suppose this is different in that respect, and the emerging data says it's not. (Various optimism about T-cells versus antibody counts notwithstanding.)
The way forward here is to make and apply enough effective vaccines to everyone on earth in a single three month period. Actually logistically doable, but the political will could be considered lacking.
Also, the morbidity doesn't get _enough_ better with vaccination. We're still looking at a debate about whether case severity correlates with degree of cognitive impairment (UK data says yes, US data says no) from the pre-vaccinated cases; we haven't got anybody (there hasn't been time!) publishing on whether vaccination reduces cognitive impairment or any other morbidity.
Bergstrom appears to argue the reason that we see purported efficacy declines in Israel and lack of staying power for 2x mRNA is due to population differences and relative age / efficacy.
The recent NEJM paper seems to indicate continued efficacy. Yes, the examples Eric Topol cites (Qatar / others) do not, but aren’t there significant problems extrapolating out from those? (Can’t really say the same with REACT though).
"The 2SLS estimates are often much bigger than the OLS estimates. Their discrepancy is negatively correlated with the strength of the IVs…"
Well that *could* be a violation of the exclusion restriction, as Xu suggests. But might it not alternatively be interpreted as "method hacking"? That is, a 2SLS analysis was selected for publication because it gave the most publishable results? That *could* be due to correlation between the instrument and the error term in the underlying probability generating process, but it could also just be spurious. Either way, I would put the blame on the selection effect.
One thing I'm confused about is the recent insistence that dropoff of Covid antibodies after a few months is a reason to get a 3rd, booster, shot. I seem to recall reading a few months ago that 'not to worry, antibodies always drop off, the important thing is that your killer-T cells have been trained and will trash Covid anyhow if it shows up in your body', or something to that effect.
Also, here in Southern Oregon, well-vaxxed in some spots and not in others, I'm waiting for the ICU's, now full, to overflow, and maybe get some antivaxxers to belatedly mask up.
The Canadian study done in long-term care homes found full vaccine efficacy to last five months. The CDC is getting six to nine months on a more general population.
It really is a coronavirus; the Uncommon Cold. You get a cold every year. There's no reason to suppose this is different in that respect, and the emerging data says it's not. (Various optimism about T-cells versus antibody counts notwithstanding.)
The way forward here is to make and apply enough effective vaccines to everyone on earth in a single three month period. Actually logistically doable, but the political will could be considered lacking.
Also, the morbidity doesn't get _enough_ better with vaccination. We're still looking at a debate about whether case severity correlates with degree of cognitive impairment (UK data says yes, US data says no) from the pre-vaccinated cases; we haven't got anybody (there hasn't been time!) publishing on whether vaccination reduces cognitive impairment or any other morbidity.
Bergstrom appears to argue the reason that we see purported efficacy declines in Israel and lack of staying power for 2x mRNA is due to population differences and relative age / efficacy.
https://twitter.com/ct_bergstrom/status/1427767356600688646?s=21
The recent NEJM paper seems to indicate continued efficacy. Yes, the examples Eric Topol cites (Qatar / others) do not, but aren’t there significant problems extrapolating out from those? (Can’t really say the same with REACT though).
"The 2SLS estimates are often much bigger than the OLS estimates. Their discrepancy is negatively correlated with the strength of the IVs…"
Well that *could* be a violation of the exclusion restriction, as Xu suggests. But might it not alternatively be interpreted as "method hacking"? That is, a 2SLS analysis was selected for publication because it gave the most publishable results? That *could* be due to correlation between the instrument and the error term in the underlying probability generating process, but it could also just be spurious. Either way, I would put the blame on the selection effect.
One thing I'm confused about is the recent insistence that dropoff of Covid antibodies after a few months is a reason to get a 3rd, booster, shot. I seem to recall reading a few months ago that 'not to worry, antibodies always drop off, the important thing is that your killer-T cells have been trained and will trash Covid anyhow if it shows up in your body', or something to that effect.
Also, here in Southern Oregon, well-vaxxed in some spots and not in others, I'm waiting for the ICU's, now full, to overflow, and maybe get some antivaxxers to belatedly mask up.
So sorry you had to go through that in Montana!