That was a pretty awesome blog!
thanks! More to come!
Once again, the NY Times is reporting on the red/blue impact of covid, but they are dancing around the headline and burying the lead:
Covid is Selectively Killing More Republicans than Democrats
The story itself is behind a paywall, so I don’t have access. Can anyone else look this up?
I am unable to find the supposed NYT article attached to this.
Here is something related:
It’s from July.
And then there’s this from October 1:
https://www.nytimes.com/2021/09/27/briefing/covid-red-states-vaccinations.html
But I don’t see that graph.
(and ha, I see they start it at Jan 1, 2021)
https://www.nytimes.com/2021/11/08/briefing/covid-death-toll-red-america.html
They start the line in January 2021, to show how the diversion of death rates changed once vaccines were available.
Seeing the same lines for 2020 might help show just how even the death rate was during non-vaccine COVID era, but I think it doesn’t show that, due to more deaths occurring in Blue urban areas (but, it’s per 100,000 people, so that SHOULD even out?).
Some conservative writers have tried to claim that the gap may stem from regional differences in weather or age, but those arguments fall apart under scrutiny. (If weather or age were a major reason, the pattern would have begun to appear last year.) The true explanation is straightforward: The vaccines are remarkably effective at preventing severe Covid, and almost 40 percent of Republican adults remain unvaccinated, compared with about 10 percent of Democratic adults.
Also, note that it’s COVID deaths, not excess deaths.
Lastly, “COVID” is still an acronym, so should be all caps, IMO. Probably a Times format thing, to save ink.
It’s a style thing. For my job, they keep changing COVID over to Covid-19 for anything I write.
In my blog, which I barely edit, I use COVID and Covid interchangeably.
Okay, I just realized one huge problem if they want to talk vaccine vs. no vaccine.
Vaccines became available in the U.S. at the beginning of December 2020. A lot of people had trouble getting vaccinated for quite a while (and I thank @ao_fan for getting me an appointment back in April) … and the peak of COVID deaths in all the four “waves” so far was in January 2021. Giving the timing of everything, very few people were vaccinated for that winter wave with the highest level of deaths.
So, if the core of their argument is that different vaccination rates are making the difference (because masks/etc were going on in 2020, so you have to stick to vaccines), you kind of have to exclude the “hangover” from January 2021.
So look at deaths from March 1, 2021 and on. They’ll obviously still get a gap as before – because that big summer jump up comes from the south (Texas/Florida/Tennessee).
It might be nice to contrast, say, Austin versus Houston or something like that. Then you have geographic proximity, but different politics, and you can check different vaccination rates & different case rates/death rates.
1 Like
So… that KFF survey:
Let’s grab this:
And let’s see how well it matches up with the CDC’s vaccine tracker:
So for men (65% at least one dose) and women (69% at least one dose)… but with KFF, we’ve got 68% vs 75%. I could see the men’s rate being within sampling rates, but women… a bit on the high side.
Then we’ve got age:
um, near 100% for age 65-74 for at least one dose? I am extremely skeptical of the CDC number.
I could graph the KFF survey responses against the CDC numbers, even I dialed it back to mid-October… (okay, let me do that)
Okay, the age 50-64 at 84% w/ one dose vs survey saying 72% with one dose. That’s a big difference.
Why do I get the feeling that if you ask people if they’re vaccinated, some will lie and say yes they are, even if they didn’t, because they feel like they were supposed to… and other will lie and say no, because it’s none of these people’s business or they’re pissed off about vaccine mandates.
[I do see in the footnotes that Texas data isn’t included in the age data, as they don’t record age of those vaccinated, evidently.]
I’m not doing the racial data, because it’s iffy – only about 70% of the vaccinated have race recorded.
In any case, there are some oddities in both the KFF survey results, and also in the CDC data. I would break out ASOP 23, but I think this is enough digging through these details to be skeptical of the latest convenient correlation.
My latest:
I was looking at 12 top locations in U.S. (and yes, NY is the state minus NYC).
here are pre-pandemic death distributions:
and here are COVID death distributions
and, for convenience, here’s the percentage point differences between the two
Another great blog. Thank you!
I was going to ask you about that Ohio Asian number, but then you explained it right below. How many Asians are there in Ohio?
274600, according to Google. So, is that jump in deaths within the confidence interval? (Not a life actuary, so asking.) Could it be more of a “which particular Asians?” (Middle East vs Far East vs Southern, versus other states’ mixes?)
Quick note: when discussing the Acela corridor, you repeat a sentence:
I used to take Amtrak from NYC to Raleigh, NC as well for many years – it’s the same route, essentially. I am extremely familiar with what it looks like along the way. In particular, I know what a really densely-populated area looks like versus… not so much. I used to take Amtrak from NYC to Raleigh, NC as well for many years – it’s the same route, essentially.
Oh, and those people you discuss at the very end aren’t looking for truth.
Yes, I know.
And I meant to say I drove I-95 & I took the train. Oh well.
Simpson’s Paradox keeps rearing its ugly head
4 Likes
I have concluded that covid is killing white people!
Well, yes, it does. It kills lots of kinds of people.
1 Like
Are any of the light blue portions Rx overdoses? Or are those baked into the expected mortality?
And if so, how much?
Yes, there are definitely drug overdoses in the light blue.
The excess mortality is just deaths over 2015-2019 average, because I didn’t want to deal with mortality rate trends, population projections (as people do move around, in addition to getting older and dying), etc. The SOA does those sorts of research projects w/ researchers who actually get paid. I’m just doing something quick. For free.
3 Likes
Deaths in the US have been trending up in recent years - you could be overestimating excess deaths by around 75k-100k based on a 2015-2019 average. That seems like it would be a decent chunk of non-COVID excess mortality if you made an adjustment.
Wave 3 non-COVID excess mortality for white seems like an odd outlier. I have seen data elsewhere that indicates high levels of non-COVID excess mortality since the pandemic started, but I still struggle with what is driving that. I get the increase in things like overdoses or suicides, but it feels like there are other causes of death related to COVID that are a result of the virus itself. Like someone had a mild undiagnosed case of COVID in spring 2020 that caused unknown damage to some organ that led to death 6 months later.
Yes, the SOA can make those adjustments in the research they do. You know, the people who get paid (and the people who can get access to the detailed CDC data, because they qualify… and I do not.)
As for sequelae from COVID, yeah, that one is going to be more difficult to get at. I did see an increase in deaths from diabetes and some other causes in 2020 that might be COVID-related. Could be due to missing treatment, but could be a causal chain, given the co-morbidities we’ve seen (and the ones we’ve not seen, such as cancer.)
Someone I know from HS lost her husband about a year ago and initially was quite bitter about it complaining that because of COVID he hadn’t sought help/been seen in as timely a fashion as she though he would have been had it not been for COVID. I expect this has lead to some degree of excess mortality.