Must always consider mechanism and data quality in COVID19 curve. Factors Topol misses: (1) UK cases started rising early June, and deaths lag by 1mo (2) cases in the other countries cited may be underreported, increasing the death/case ratio (Russia curves esp make no sense)
Another thing to consider is if Topol included precisely those countries that have the highest death/case ratios to make a contrast with UK, then he would have filtered in exactly those countries that are underreporting their cases.
Kind of frustrating that Topol would continue to make the same mistakes in data interpretation (no mechanism, no awareness of data quality, no awareness of selection bias)
Of course deaths will be mostly reduced as at-risk people are well vaccinated in the UK (and vaccines work), but there will be some deaths from hypoimmunity and the occasional unexpected fatality. As it's too early for deaths, misleading to imply there won't be any.
Meaning Eric may be able to say in the end he was right on the conclusion of lower deaths due to vaccination (I'd certainly agree with that too) but misinterpreting data to support that conclusion now is not proper scientific methodology.
If it's motivated by a desire to show vaccines working in the real world, there are plenty of other ways to do it. In the meantime, the message backfires as the young unvaccinated think they are invincible and their spreading the disease to hypoimmune and children doesn't matter
If the goal is to show vaccines prevent deaths despite Delta surging, the best way is comparing an earlier wave to the Delta wave in the *same* country. At least there's a chance of reporting standards being similar, and you'll remember to correct for the case-to-death time lag
• • •
Missing some Tweet in this thread? You can try to
force a refresh
CDC advice that vaccinated should feel comfortable spending all the time they want indoors and unmasked with the 50% unvaccinated population while Delta surges through their communities doesn't sound so reassuring now, does it?
This thread, and probably the one it refers to by Topol, should be pulled. The study did not look at transmission from vaxxed people at all. Topol wrote "91% efficacy of blocking transmission of infection" when he should have just written "91% efficacy of blocking infection"
Actual results: 91% refers to the reduced infection of vaxxed people, similar to the 95% in clinical trials. All of this is pre-Delta. So kind of non-news.
Sorry gotta go against Fauci on this one, just as I did in March 2020 on masks. If you don't want to be a breakthrough infection and pass on the virus (vaxxed still have ~20% the chance of unvaxxed with Delta) and you're highly exposed, then wear a mask. Simple, cheap, convenient
👏LA's Ferrer "We have enough risk and enough unvaccinated people for Delta to pose a threat to our recovery. And masking up now could help prevent a resurgence in transmission. This is a precautionary recommendation, given that we don’t have all the information we wish we had.”
Ferrer said there’s “increasing evidence that a very small number of fully vaccinated individuals can become infected with the Delta variant and may be able to infect others.”
OK! Finally found data on breakthrough cases by each vaccine type. Includes hospitalizations and deaths too. Numbers currently too small to make any conclusions, but worth keeping an eye on this space, especially those curious about J&J efficacy: oklahoma.gov/covid19/newsro…
Yes these Oklahoma Health Dept reports are excellent. Here's the # of breakthrough cases up to last week by manufacturer. I ran some Fisher's exact tests. The difference between J&J and Moderna case numbers comes up stat sig. Nothing else does (small #'s means low stat power)
If we get more #'s, we can run more stat tests to see if any of these differences become statistically significant (not more than 5% probability of being due to chance). But even better would be if we can find numbers from more populous states. Please help look if you have time!
Some observations from the UK:
• Vaccines provide ~95% protection against death from Delta
• Deaths in vaccinees are in older people with preexisting conditions
Numbers as we expect for AZ and RNA vax, as is breakthrough severity risk factors
J&J is still expected to be a bit less efficacious. So rather than conclude everything's ok and that's the best we can do, I take it as certain people can use a booster if they are high-risk
We can do a back-of-envelope calculation for the expected impact. We’re predicted to get 30M cases, 1M hospitalizations, and 100k deaths in the next wave among unvaxxed. If our vaccines are 90/95/95% effective on these measures, and half of people are vaccinated, then...
The US at large is now in positive growth for COVID-19 cases, as are exactly 25 states. This reverses a trend of decreasing cases in most states until last month. I'd love for the "no surge" prediction to be true, but it's looking like math is winning over wishful thinking again.