There’s a great, long Twitter thread which looks at the effect of vaccines. (It summarizes a paywalled article, if you want to see the whole thing.) One of the elements which I found striking was how the risk of COVID-19 relative to the flu has dropped over time. In mid 2020, the infection fatality rate was about thirteen times higher as for the flu for people over 60 in England (NB: this is a semilog chart!):
By early 2022, however, the infection fatality rate had dropped to “only” twice as lethal as the flu, due to vaccinating hard the most vulnerable:
(Don’t forget, however, that COVID-19 is also much more transmissible than the flu. It’s “only” twice as lethal, but there are five times as many cases.)
Okay, data from BA.2 (a subtype of Omicron) is coming in. This preprint says that the odds of getting BA.2 are about twice that of getting BA.1 (<- which is what people casually call ‘Omicron’). Interestingly, the odds of getting it are 2-3x regardless of your vax status (compared to BA.1 with the same vax status), but for 2 or 3 dose vaccinated people, the chance of transmitting it is actually lower.
This tweet (in French) has a nice visualization of data (from this report from the French stats agency) shows the different hospitalization risk, by age and vaccination status for people with Omicron vs. with Delta:
I was surprised at how much bigger the risk was for people over 80 from Delta than from Omicron, especially in the unvaccinated. Yikes!
In Friday’s General post, I mentioned a study which said that Long COVID in kids is rare and resolves quickly. This thread and its followup thread take exception to Friday’s paper. The biggest issue he points out is that the survey asks about symptoms after COVID-19 diagnosis for one arm, but for symptoms any time in the past year for the control arm. The kids in the control arm could have been sick with some other illness at some point in the past year (and given that RSV made a roaring comeback in many locations, it might be likely that the control kids were sick at some point).
There also was no attempt to find out if any of the kids in the control group had been infected with COVID-19 asymptomatically. If a lot of kids were infected but didn’t know it (and then got Long COVID, even a mild case), that would make Long COVID appear to be less common.
The author’s Part 2 is a criticism of the type of the study. He says that looking at the prevalence of individual symptoms is stupid, the thing to do is to look for clusters of symptoms.
He also says that the survey doesn’t do a good job of quantifying how severe the symptoms are, and I’m with him on that. If you just ask 100 people if they have had fatigue at any time in the last year, I bet close to 100% of the people will say “yes”. You’d get a more meaningful result if you merely asked if there was any unusual fatigue, and even more if you gave a Likert scale of how severe or longlasting the fatigue was.
Something he did not mention that bothered me is that the survey was sent to the parents. While I understand that the parents are the one with the mailing address, and that little kids might not have the sophistication to fill out the survey themselves, I worry that that might introduce another degree of separation. Now you’re asking the parents to remember and report on how the child felt sometime in the past year. :-/
In this study, the authors did a Fourier analysis to try to figure out what the periodicity was. They found no evidence for any year-based seasonality, nor any consistent periodicity.
This article reports that Prime Minister Justin Trudeau tested positive this morning. He says he is boosted and asymptomatic.
This article reports that the Canadian Blood Services has put out a call for blood donation.
This is a great article talking about the state of free (as in speech, not as in beer) vaccines, i.e. with patent restrictions eliminated. One of the things it talks about is how difficult it is to roll a vaccine out, even if you don’t have patent restrictions. Please go read this.