More younger people at the ICU compared to past? Could be an outbreak affecting them more because of network structure or behavior, vaccine/past immunity among elderly, ICU capacity change compared to past or even temp reporting blip in some countries, among many other things.
-
Show this thread
-
A variant rising in an area could be founder or stochastic effects (one version happened to catch in a superspreading event—this pathogen is very overdispersed) or actually outcompeting others through more transmissibility. That’s why it took time to establish B.1.1.7 was.
2 replies 6 retweets 129 likesShow this thread -
Similarly, previous (not that well-established) claims of high infection rate through models based on (maybe, maybe-not accurate) nonrandom seroprevalance are NOT, by themselves proof of widespread reinfection, or even that reinfected cases are severe or equally contagious etc.
1 reply 5 retweets 107 likesShow this thread -
I used to teach methods and stats for sociology students, and epidemiology examples were always very useful because they really are among the thorniest to try to untangle. And I believe harmful to overclaim, and not necessary to do so to make basic public health recommendations.
2 replies 11 retweets 196 likesShow this thread -
zeynep tufekci Retweeted Eric Feigl-Ding
For example, the Brazil situation is horrific and deserves immediate attention, but this thread below is an example of how not to draw conclusions about P1 and it’s differential effect on young people. (The situation is bad enough without making unsupported claims like this).https://twitter.com/DrEricDing/status/1379995247812444162 …
zeynep tufekci added,
Eric Feigl-DingVerified account @DrEricDingLet this sink in—over **half of all intubated young adults** on ventilators now dies from#COVID19 in Brazil
. What’s odds last spring 2020? 1 in 3.
That is likely an effect of #P1 variant—not just ICU capacity because March 2021 wasn’t included yet.
(HT @terrence_mccoy) pic.twitter.com/mHY5Vfqu0iShow this thread12 replies 53 retweets 272 likesShow this thread -
Similarly, it’s likely that P1 is more transmissible than the wildtype, but what’s happening in Brazil is completely compatible with an unchecked epidemic causing hospital system collapse even without higher transmissibility let alone any—unproven, speculative—effect on severity.
6 replies 12 retweets 220 likesShow this thread -
zeynep tufekci Retweeted Alexis C. Madrigal
Yes, of course. There are *SO MANY* confounders like that. We saw admission criteria change even in the United States during surges, affecting a lot of hospital-level statistics. Making claims about characteristics of variants/surges is hard, and it’s important to be careful.https://twitter.com/alexismadrigal/status/1380152737204072448 …
zeynep tufekci added,
3 replies 5 retweets 173 likesShow this thread -
Replying to @zeynep
You are right in flagging those confounders. But as you said in a recent newsletter, "sometimes the plural of anecdote is, indeed, data". And here in Brazil (and now Canada) there are plenty of anecdotes about how younger people are arriving in hospitals in worse conditions.pic.twitter.com/6lGRw78IhC
1 reply 0 retweets 0 likes -
Replying to @Burgos
It raises the question but doesn’t answer. Good analysis is the ability to tell the difference rather than rigid rules like “correlation doesn’t imply causation” (except when it does) or “absence of evidence isn’t evidence of absence” (except when it is because you are looking).
1 reply 0 retweets 0 likes -
Replying to @zeynep
Agreed! It doesn't answer, it just helps to formulate better hypothesis. And I'm with you that we should refrain from statements like those from Eric Ding. But following closely the situation in Brazil it does seem that there is one additional, different variable this time around
1 reply 0 retweets 0 likes
Could be just more transmissibility plus hospital collapse. Don’t need much more than that, necessarily, to see what we’re tragically seeing there. May also be more severe—possible. But definitely too early, and “disporportionately severe on the young” seems like a giant stretch.
-
-
Replying to @zeynep
Yeah, I think with the data that we have it's very probable that the P1 variant is *not* "disproportionately severe" on the young. What many doctors are saying is that since younger people wait more for treatment, they tend to arrive in worse shape and stay longer in hospitals.
1 reply 0 retweets 1 like -
Replying to @Burgos
Exactly. A lot of things may well explain rise in younger patients dying, especially compared to a whole year ago.
1 reply 0 retweets 1 like - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.