Death rates diverged at the point in time at which the vax was introduced. Isn’t that itself strong evidence? Unlikely that the confounders changed substantially at that exact time.
It’s only not strong enough evidence if you are desperate to believe vaccines don’t work. Saying that the “weak” already died early in the pandemic in liberal leaning states is ridiculous. Massive waves of COVID ran through all the states by the time vaccines were available, and the Delta wave in particular was pretty bad in places like Florida. It also doesn’t take into account any amount of natural aging that the population everywhere would be going through... people are getting older every day, more people are becoming diabetic, developing cancer, etc.-- the people pushing anti vaccine arguments are just looking for confirmation that their assumption “the vaccine doesn’t work” is true; desperately flailing for one piece of anecdotal evidence is not going to prove any of them right unless nearly the entire medical establishment is in conspiracy together-- a virtual
No. Because blue states were hit early in the epidemic. That killed off their most vulnerable and gave some immunity to most of the rest. So as COVID later spread to red states, you’d naturally expect higher mortality in red states than blue during that time period, which overlapped vaccine introduction. Attributing that to the differential vaccine rates is pretty stupid. And I say this as someone who was in the phase 3 Pfizer trial, so not an anti-vaxxer, just an opponent of dopey analysis. Ecological fallacy is another issue, which Nate seems to dismiss for no good reason. Hey Nate, take a look at the mortality rate for blue Navajoland vs red Pinal County and tell me what you see. And as for age, just using 65+ is dopey, because risk to an 85+ person was vastly greater than risk to a 65+ person, so finer granularity, say deciles or better, ventiles.
I have done a gazillion regressions in my day and written FORTRAN programs to do them, not just using R or Stata like twerps today, though I’ve used those too. Point is that there is a very major logical flaw in Nate’s analysis. It’s up to him to think these things through before presenting them as something worth reading.
Here you are bragging about how much you've done, yet you cannot produce an analysis? I think Nate has probably thought this through quite a bit, considering all the work he shows. Whereas I don't see that from you.
Just as the AIDS epidemic needed a sociologist working with a public health officials and statisticians to solve the problem because the experts needed an expert on homosexual behavior to understand spread. To properly analyze this data one needs an intellectually honest political scientist working with other experts. I’m not a political scientist but I happen to know a lot about southeastern demographics and politics and so I can easily conclude that NPIs had a significant impact on lowering the Covid death rate.
Btw, did NPIs work with the AIDS epidemic?? I know in Brazil and India and Africa it was solved by the medications Big Pharma developed.
If I had the data and time of course I could do it. Any idiot can run a regression with R or Python or even Excel, and indeed idiots regularly do. I’ve pretty much lost interest in this sort of analysis because demonstrating a causal relationship is nearly impossible outside an experimental setting. You don’t need to tarry over causality to come up with effective forecasting models, of course, which is what Nate usually does, but here he is trying to demonstrate a causal relationship with an obviously deficient model.
There were experimental settings (you were in one!) that found the vaccines reduced mortality substantially. There was then differential uptake by state, and there was also differential mortality by state, in exactly the pattern you'd expect if differences in uptake caused differences in mortality.
I can't randomize states to take or not take the vaccine, so I don't really know what else it would take to convince you
Of course not, I’m suggesting that red state infections lagged many months behind blue states, a fact easily observed in the data, which would cause the effect I described. If you look at data from April or May 2020, using Nate’s crude reasoning, you would conclude Trump voters were immune to COVID.
But what does this have to do with vaccine's becoming widespread at the beginning of 2021 and the difference in death rates that emerges then? COVID was clearly already everywhere in the country by then.
The difference didn't emerge then. The difference emerged in summer 2020 and was fairly consistent from that point forward. The difference is mainly caused by obesity, healthcare, and age differences, with obesity being a primary driver. A state obesity chart looks similar to a state covid death rate chart. When vaccines were introduced, the slope of this difference appears to have increased modestly - that's probably the effect of different vaccination rates. But that is NOT the primary driver of differences between red and blue states. The primary drivers are obesity and age, only one of which Nate corrected for here.
This I agree with. Red states are on average more obese than blue states and weight should have been included. But whether the virus arrived in red/blue states in march/april/may isn't a relevant factor assuming all states were saturated by the time vaccines arrived.
Agree. Nate is doing the kind of doofus analysis that data journalists do. Saying that a primitive model is better than a more elaborate one because overfitting can sometimes be a problem is really lame. (And let's not hear any babbling about instrumentation variables as a cure, that's nonsense 99% of the time.)
It sounds like you are making a testable statistical claim. Something like "Cases before the introduction of vaccine are negatively correlated with deaths after the vaccine was available".
And also the claim "Once you control for this vaccination rates no longer predict deaths after the vaccines".
Have you run the numbers on either claim? At least the second seems dubious to me, but I would be happy to change my mind if you ran the analysis.
I suggest you review the ample research on the effects of natural immunity after COVID infection. For example, "Past SARS-CoV-2 infection protection against re-infection: a systematic review and meta-analysis," the Lancet, February 16, 2023: "protection [from prior infection] against severe disease remained high for all variants, with 90·2% (69·7–97·5) for ancestral, alpha, and delta variants, and 88·9% (84·7–90·9) for omicron BA.1 at 40 weeks." 40 WEEKS. So, if persons previously infected (more common in blue states because of the way the epidemic spread) were more likely to be vaccinated (more common in blue states) then you have an obvious confounding problem and will (in Nate's simple model) attribute the differential in mortality to vaccination rates, when prior infection accounts for much of it. That was my first point.
And I should not need to cite research to show that once the most vulnerable have been killed off, the epidemic will not kill the same people again. I think we can agree on that, which was my second point. Most of the vulnerable people in blue states had been killed off before vaccines arrived, while most of the vulnerable people in red state had not been killed off. Another confounder.
So those are two omitted variables obviously confounding any attempt to attribute differential mortality rates to vaccination rates at the state level. Another, as another commentator has observed, is obesity rates. I could add diabetes and other comorbidities as relevant.
If you want an analogy, consider comparing death rates from the Black Death in Italy vs England in 1348 vs 1349. If a useless treatment had been introduced in Italy early in 1349, would we conclude that it protected the Italian population simply because England had a much higher mortality rate in 1349? The real reason, of course, is that Italy was hit earlier in the epidemic.
I am not disputing the existence of natural immunity, or of other factors that influence covid deaths. But the claim that these variables can explain the demonstrated effects of vaccination is a much bigger claim.
It is also analysis that can be done. The data is available. If you think you have some factor or set of factors that can explain Nate's results then you should do the statistical analysis to demonstrate as such.
To close this out, I ran my own regression by state. COVID deaths per million (as of 10/23/23) as the dependent variable. X variables: % of population African American (2020); % of population Urban (2020); % of population 65+ (2020); and % of vote for Trump in 2016 election.
R2 was .60 (adjusted R2 .57). F significance 1.4x10^-8. Intercept and all coefficients significant at less than the .001 level.
Now, you might say that % vote for Trump might be just a proxy for vaccination level. Nope. Add % of population with one or more vaccine doses into the model, and the % vaccine variable has a P value of only .70. If you get rid of the Trump variable and keep the vaccine % variable, the P value for the vaccine variable rises to .0005, which is interesting but still a lower P value than any of the other coefficients. And the adjusted R2 falls to .36, which is far worse than the .57 of the original model.
As for what the % vote for Trump means, it likely means (in my opinion) that Trump voters have less fear of death (consistent with greater religiosity) and therefore are not obsessive in safety measures such as vaccination, social distancing, and so on. I think the Navajo here in Arizona are similar in not being terrified of death, and perhaps the African-American population as well. Is this good or bad? Not for me to judge.
Did you look at the data? Doesn't support your statements. (Also some obvious garbage data, such as a 1.2% seroprevalence for Vermont in early February 2021, when the case count for Vermont exceeded 13,000 (>2.0%) and the number of infections would be, of course, many times greater than the case count, making the estimate off by perhaps an order of magnitude.)
As long as we are talking, if you plan on running a regression, I suggest including percentage of state population that is African American and also Native American, since those subpopulations had much higher COVID death rates than White or Asian subpopulations. Percentage African American is highest in southern states red states), which of course elevates the COVID death rates in those states, despite the subgroup being (as with Native Americans) blue voters. Obvious ecological fallacy going on in analyses that fail to look at that. It's all very interesting but you'll need a very sophisticated model to untangle this stuff. Lots of variables in play.
That is an awfully convoluted story where a much more concise story is adequate. There are also controlled studies that demonstrate lower rates of severe COVID. The burden on your position is well beyond a single counter example and recommending a different age cut. Do better.
Oh for heaven's sake, duh, of course the vaccines reduce mortality. That's not the question. The question is how much of the mortality differential observed between blue and red states after the introduction of vaccines can be attributed to vaccines versus the timeline of the epidemic and other factors. Entirely different question. I have no doubt that some of the differential is attributable to vaccines, but a lot less than Nate's simple analysis suggests. See my response to Not a Lib above for details.
Since we agree on basically everything, why are you arguing secondary effects? This isn't intended for a journal, he's just demonstrating vaccine uptake has macro impact on death rates and uptake has a political component. Look at the SEs on his regression and realize this is not intended as a exercise in precision but instead a blog post.
My point is that it does not demonstrate that, at all. I agree that vaccines reduce individual probability of death from COVID, ceteris paribus, because randomized controlled trials have shown that, but Nate’s analysis does not show that.
Sigh. He's demonstrated different populations have different death rates in a manner that is highly correlated with vaccine uptake while also considering age effects. We have a strong prior based on individual testing that death rates decline with vaccination. Basic first order effect should be that populations with more vaccinations have lower death rates. This correlation is present in the data, even when considering age as an alternative.
He didn't do exactly what your wanted him to do, but he ran a pretty simple set of correlations that reinforce priors based on individual studies. You would need fairly strong evidence to demonstrate political alignment -> vaccination rate -> death rate is unreasonable or insignificant. You point out some reasonable second order channels, but I'd need some math to show they are game changing. I fundamentally don't by the burn out story as many states saw multiple severe waves : LA, AZ
It is not clear at all that the vaccines reduce mortality. There are unknown amounts of people that were dumped into the unvaccinated category upon death if they were not 14 days past the second dose and that is a data crime that skews all subsequent analysis.
The JAMA study reported in the Summary Table the raw number of deaths revealing that the age-specific Rep death rate in Ohio was 13% less than the Dem rate but they didn't share how that varied over time.
What they did report was a bizarrely calculated *excess* death rate at the county level that was adjusted by the partisan difference in death rates at the state level.
So the "higher death rates" are only relative to a very rough model of what the rates should have been with the *state level difference erased*. And they didn't share the data or even the unadjusted rates over time. That's unethical and dishonest.
But Nate apparently didn't even read the study because he happily skips directly to "death rates" which he has no information on, unless he secretly obtained the data the authors did not disclose.
Eh? Excess mortality data is readily available, and there are not "corrections" at the state level. Just counting number of dead people, county by county week by week by coroners who who might be red/blue or anything in-between.
Excuse me, I did read the study. But I don’t care about the study since the unmassaged data is readily available. I did my own downloads and some of my own analysis. I was also suggesting that Nate use that data so that people like yourself wouldn’t carp about all this. So… where, in your own proof, did you see that older democrats died at a higher rate? Do you have a worked example? I’d love to see that.
No, we don't know what happened to death rates -- they didn't share that data temporally. We do know from eTable 1 that overall Democrats had higher age-specific death rates.
The PROJECTED death rates were calculated separately for each party so the EXCESS rates started out converged BY DEFINITION. They initially shifted higher in Democrats but then had an inflection point in June 2020 and started shifting higher in Republicans relative to the separate projections. Nothing special happened in April 2021 except that the trend line in each age bucket continued.
thats not evidence the vaccines are not effective. the vast majority of people, especially old and immune-compromised people most susceptible to covid caused death were vaccinated. a much larger percentage of a much smaller group can be a smaller raw value than the smaller percentage of the larger group. sure the vaccines were not 100% effective at preventing death, but literally all available evidence says they made a pretty dramatic difference
More importantly it's not evidence that the vaccines were effective against death (and hospitalization). In fact, the reanalysis combined MRNA and PFE showed that when looking at severe adverse events of special interest (determined prior to the studies), there were more in the treatment arm and it was statistically significant.
I think a little more work needs to be to establish that as a fact. Seems like the inflection point for differential mortality rates was September 2020 (graph c from the previous post). Look at it this way, if the graph wasn't labeled with the vax date where would you say the change in mortality started taking place? In fact, it is a strange coincidence that the excess death rate for Republicans is so much less than for Democrats at the exact time that the vaccine is widely available... I wouldn't make much of that but I also wouldn't take it as pointing to vax rates as the main culprit for the difference.
Hawaii did the best and Puerto Rick did great because of NPIs. So if you believe the GBD halfwits once the NPIs were eliminated Covid would have ravaged those low natural immunity populations…but it didn’t because slowing spread until the availability of the vaccines and then getting everyone vaccinated worked to mitigate severity and prevent hospitalizations and save lives.
What worked for Hawaii & PR was their island status making it easy to prevent importation of the virus. Much as it did for Australia & New Zealand. This was not an option for non-island states. You cannot close the land borders between states in the US. People flowed freely across them and made stopping spread of the virus impossible.
I know that…but it shows NPIs work. Travel restrictions are an NPI and Vermont and Maine and Florida Keys and even Nebraska benefited from effective travel restrictions in that nobody was driving through those states. The Dakotas simply threw caution to the wind and encouraged people to travel to states nobody wanted to travel through and had big spikes because of it.
Travel restrictions only work in very limited scenarios - mainly islands or places that can easily control all travel in and out. Vermont, Maine, & Florida Keys are also not typical places. They are all fairly isolated, wealthy, & healthy.
Vermont and Maine are super old. And so is Puerto Rico which is also poor although I understand PR doesn’t really have a peer state to compare to. Arizona encouraged people to travel to it in winter 2020/21 and it is the largest state with fewest NPIs in 2020 and sure enough it has the highest Covid death rate.
I find this analysis all pretty compelling. Thanks for the follow up, Nate. One item that does catch my eye, though, is this:
>>Are all states counting COVID deaths in the same way? (Probably not.)<<
It probably would be interesting to look at excess mortality instead of COVID deaths. I do recall seeing nation-to-nation comparisons that appears to differ quite a bit when using the former rather than the latter. But I haven't looked into whether the US (or individual states) compile state-level excess mortality numbers.
At least in Iowa, the covid numbers began to get LESS reliable (in the direction of undercounting) as vaccination and vaccine uptake became political footballs. So I would guess that is true in most red states: if you think “covid isn’t real LOLZ” then your numbers are going to underestimate how much worse things got in their states when they (duh) could have gotten better with vaccines.
This sounds reasonable. On the other hand even in more progressive jurisdictions, it's pretty likely, I think, that SARS-CoV-2's contribution to increased mortality has been underestimated: surely substantial numbers of infections go undetected.
I've long suspected it's necessary to look at raw excess mortality numbers (at least for the period 2020-2022) to get an accurate picture as to the pandemic's death toll.
Yep; basically came here to say exactly this. The labeling for deaths is an inexact process that makes it hard to parse data along that dimension. But cats in boxes aside, it’s pretty simple to say if someone is dead or alive. Look at all cause deaths, back out suicides and drug overdoses (I think this an extremely important factor especially in 2020-21) and then see how the numbers look.
You're not wrong (it would be interesting). However, you're also not wrong in that it likely wouldn't change the results all too much. Minor changes aren't going to result in major differences across all states (maybe it results in a measurable change in a few).
I also suspect it wouldn't result in major changes. Though it would be nice to be sure either way. I recall looking at an excess mortality vs. covid death comparison between the UK and Germany that made the latter's advantage over Britain in terms of covid deaths a lot less impressive (perhaps the NHS, for all its faults, is good at record-keeping and accurate reporting of data?).
EDIT: Indeed: if you held a gun to my head and forced me to guess, that guess would be that, if we looked at state-level total excess mortality, the covid toll in the red states has actually been worse vis-a-vis the blue states than we've been led to believe based on covid death numbers alone.
I think your final conclusion is reasonable considering what we know from certain states. If anything (the most glaring example being FL) red states were less likely to call something a covid death and the blue states were more likely. In the end, it was probably relatively minor in terms of this analysis as the total effect would be low. Most (say 90%) were probably reported correctly.
Right, and also, it seems probable that, in jurisdictions where anti-vax and anti-NPI activism were fiercest, less coronavirus testing was occuring, and consequently less identification of SARS-CoV-2 infections.
This is what I came to post, but I am tired, and burnt out discussing the topic, but yes - we can easily compute excess deaths and I have gone into great detail tracking excess deaths/all-cause mortality through several reports the CDC publishes weekly.
Here is a high-level summary in google sheets for anyone interested, all cause deaths, by state, by year, going back to 2014. I also included population (from Excel population data), Obesity (BMI from a CDC report), and then columns M-AB I did a few quick-and-dirty calculations (excess deaths using the "3 year average compare"*, deaths as % of population, deaths per million residents).
I don't see the effect with excess deaths Nate sees with covid mortality. Some examples:
Maine averaged 14,589 deaths per million 2017-2019, and that average holds true 2014-2019 as well with little deviation. 2020 they rise to 15,859 deaths, but the "vaccinated years" of 2021 and 2022 they hit 17,095, and 17,076 deaths despite being one of the most highly vaccinated states. Why didn't they return back to ~15,000 deaths projection?
Similar issues for Vermont, Washington, Oregon, New Hampshire, Delaware - all had elevated mortality not recede in 2021 and 2022. Why?
"Red" Ohio has a 14% absolute vaccination difference from "Blue" Pennsylvania (that's quite a bit) [2], but their excess deaths are fairly close at 13% and 16% respectively - and if this 3% absolute difference was the result of a large difference in vaccination, then why did other neighboring states with similar gap in vaccination have identical outcome? Example, South Dakota much higher vaccinated than North Dakota, yet ND has slightly less excess deaths (16% vs 17%) and both countries saw mortality return close to baseline by 2022 - also wasn't South Dakota "ravaged by (Kristi Noem's) 'Trumpian response' to the pandemic" [3] - how did they do so well with all those motorcycle rallies, kids going to school, not wearing masks?
Why did low vaccinated Iowa, the state that didn't "care whether you live or die" (WaPo) [4] have one of the lowest excess death rates in the country? (14%)
Why did California and Florida have near identical excess deaths (21.1% and 21.9%) despite California having a much healthier population (in terms of BMI and other metrics), a younger population, and importantly - these calculations use a fixed population dataset - when we know that Florida increased it's population the last 3 years while California decreased.\
I have talked about this in depth with many "true believers" of (for lack of a better term) the "official narrative" and, I largely get hit with goalpost shifting ("what about population density", "what about early spread in liberal states", etc) and accusations of "cherry picking" - but not only are there a shit ton of cherries, but a good hypothesis should welcome counterfactuals.
It seems to me the base "health" of a state population - obesity, alcohol consumption, age, smoking, location (being in a rural area has long been known as a contributing factor to mortality - too lazy to post myriad of studies covering this - it's common sense) - all seem far more indicative of covid outcome than vaccination rate.
ESPECIALLY because the country level detail offers damning counterfactuals - specifically that South Korea had the 2nd largest year-over-year mortality spike in 100 years, of any country in the world in 2022 - jumping from 310K deaths to 372K deaths.
* The "3 year average compare" method is similar to how OWID was originally calculating excess deaths before switching to the Karlinsky-Kobach formula which is similar to a linear regression on weekly data.
(careful with 2023 data, it lags quite a bit.... 6 weeks back is roughly 90% complete, 12 weeks back 98% complete, etc... they update this every Wednesday
**note that this is weekly data some years (2020, 2015, etc) will have 53 weeks instead of 52. I have a similar analysis through CDC wonder which goes by year, but I like the drill down of weeks to see trends (not included in this google sheet, but the above links provide it to the curious)
[2] I keep meaning to add this data into the spreadsheet, though not sure if still accurate
>2020 they rise to 15,859 deaths, but the "vaccinated years" of 2021 and 2022 they hit 17,095, and 17,076 deaths despite being one of the most highly vaccinated states. Why didn't they return back to ~15,000 deaths projection?<
I can think of a few possibilities: 1) SARS-CoV-2 had become more widespread (edged closer to endemicity) by 2021-2022 than in 2020, when it first arrived. Indeed, nationally, 2021 was the deadliest year for Covid mortality. I'd expect most places in the US saw more excess mortality in 2021 (and perhaps in 2022?) than in 2020. There was simply a much greater chance of becoming infected as the pandemic wore on, and this partly swamped the protective effects of vaccines. 2) Delayed effects of SARS-CoV-2 infections: we know the virus affects cardiovascular health. It seems likely in certain cases the accumulating effects of heart problems took some time to result in premature death. 3) End of social distancing benefits: we've seen pretty clear evidence from Australia and New Zealand that social distancing and NPI practices—while undoubtedly unpleasant and in some cases psychologically damaging—held down numbers of certain kinds of health problems in addition to Covid cases (ie, reduced flu infections, reduced intestinal viruses, reduced car accidents, etc). So, the end of lockdowns would likely result in an uptick in some causes of death, that, combined with the lingering mortality increase associated with SARS-CoV-2 endemicity (we're basically stuck for time being with the mortality equivalent of an extra flu outbreak—and a fairly deadly one at that—per year), resulted in a net increase in overall mortality relative to both the pre-pandemic period and the early period of the pandemic itself (say, first half of 2020).
My guess is that, in the next couple of years, we'll see all cause mortality retreat from the 2021-2022 numbers, but probably not quite get back to the pre 2020 years: we've now got a new disease likely to kill (or contribute to the deaths of) several tens of thousands of Americans annually for the foreseeable future.
Thanks for the feedback and hypothesis testing Charles.
My main point is that when reviewing excess deaths at the state level, - I don't find the original claim Nate proposes ("all states had the same initial outcome, but after vaccinations blue states improved") holds when viewing excess deaths, hence the "counterfactual" examples I pointed to.
A few thoughts on the possibilities you propose:
CR: >>1) SARS-CoV-2 had become more widespread (edged closer to endemicity) by 2021-2022 than in 2020, when it first arrived.<<
We found that Covid spread earlier than we realized. In the US [1], France, [2], China [3][4], and Italy [5], so this hypothesis may or may not be true. Seems to me it was spreading long before we realized it, which is interesting because there was zero signal in excess deaths until we took action.
CR: >>2) Delayed effects of SARS-CoV-2 infections: we know the virus affects cardiovascular health.<<
We "know" the virus affects cardiovascular health because we have scrutinized this virus, and this virus alone, to such a high degree, to search for anything and everything it may or may not do. What we don't have, is the "compared to what?" question, the foundation of scientific skepticism. What if the same methodology applied to Covid was applied to the flu, or one of the other dozens of Coronaviruses, or RSV, etc? Might we also find correlation with cardiovascular issues?
I think of Feynman's take on this "may be true, may not be true" [6]
It seems possible that if searched for this signal with the flu, we may also have found it. We just never looked before.
CR: >>3) End of social distancing benefits: we've seen pretty clear evidence from Australia and New Zealand that social distancing and NPI practices...—held down numbers of certain kinds of health problems in addition to Covid cases (ie, reduced flu infections, reduced intestinal viruses, reduced car accidents, etc<<
A competing hypothesis explains why "flu went away", and it has nothing to do with social distancing and masks - viral interference. I wrote a detailed counter to this hypothesis to Jeremy Faust on his substack here:
The "TL;DR" is that places which didn't participate in social distancing and masks also saw Flu disappear. The Nordic countries all returned kids to school, largely unmasked, short lockdowns, and flu disappeared there too. Same in South Dakota, Florida, etc. Additionally we had seen a new virus knock out other viruses before, like Swine Flu did to RSV.
Also you mentioned "reduced car accidents", but that isn't true (at least not for 2020). Surprisingly car accidents increased while we were social distancing.
Additionally, all accidental deaths (sum of car, homicide, suicide, drowning, poisoning, fire, etc) rose significantly in 2020 (+50K year of year) and even increased in 2021 and 2022 (+60K compared to 2019 each year).
CR: >>My guess is that, in the next couple of years, we'll see all cause mortality retreat from the 2021-2022 numbers, but probably not quite get back to the pre 2020 years:<<
It's possible that 2023 sees return to "baseline" mortality, at least in the US. For the first 26 weeks of 2023 we have 1,559,204 deaths. Using linear regression of 2014-2019 first 26 weeks to predict deaths for 2020-2023, we would expect ~1,552,602 deaths at this point. Really just depends on what happens the back half of 2023. First half of 2022 by comparison we had 1,687,056 deaths. Might be possible, I hope.
"My main point is that when reviewing excess deaths at the state level, - I don't find the original claim Nate proposes ("all states had the same initial outcome, but after vaccinations blue states improved") holds when viewing excess deaths, hence the "counterfactual" examples I pointed to."
This is not what Nate said above.
He said:
"Until vaccines became available, there was little difference in COVID death rates between blue states and red states.
After vaccines became available, there were clear differences, with red states having higher death rates, almost certainly as a result of lower vaccine uptake among Republicans."
So no, not all states had the same outcome in 2020. Just red and blue were not significant variables in 2020.
And more important he did not say that blue states improved in 2021/2022. They just did better than red states.
In general Covid killed a lot more people died in 2021 and 2022 than in 2020. That's because 1) Delta was more virulent and more deadly and 2) It took quite a while to spread to everybody.
Also, obviously, blue and red states only differ modestly. Even the bluest state is 30% Trump, and will still have plenty of anti-vaxxers.
Anyway, if you want to do your analysis like Nate above, using excess death statistics, you can. It won't be true for every single state, but the effects should be similar.
I redid my own regression looking at the predictors of the change in overall death rate as you suggested. I did find a less significant effect than simply running the official COVID deaths but it is significant and in the same direction as Nate found.
Percent getting vaccinated is the only significant variable when run together with %over 65 and %of Democratic Vote. Vaccine uptake is more important than partisanship but those two variables are confounded.
I just looked at the correlation for % vaccinated and Democratic Percentage vote and it's a modest r=0.39 however staff vaccinated is substantially higher at r=0.86.
This difference surprised me to the extent that I reran the regression using %of Staff... and there were no significant effects on the change in 20-21 mortality or COVID Deaths.
All together there is an effect on the change in overall death rates but there does seem to be a little relative increase of the official COVID numbers in "bluer" states. Bias against counting COVID in "reder" States? Or inflation of COVID associated deaths in "blue" states could explain this difference. Either way thanks for making me check my biases.
The mortality actuary "Meep" (Mary Pat Campbell) just posted an interesting rebuttal to Nate's analysis as well. Just going through it now, you may find it interesting.
I'd say it's fun exploration of the data, though not really a rebuttal. Making scatterplots by age category is clever, but it points to the same effect (even the same R^2), for both sets of ages.
I would almost say that this article is a good rebuttal to MPC's approach. That is, if you're arguing on the internet, it makes more sense to -start- by presenting a simple analysis. But you have to go into it -knowing- that your simple model will yield the same results as someone else's more complicated model.
Interesting point. I ran that correlation obesity and death rate and there is a r= 0.25 pre vaccine and 0.53 post vaccine correlation. Way more than age related factors.
Sure (although not that many - weight and preexisting conditions had an effect, but a much, much smaller one than age). But I can't think of any that would have changed right when the vaccines were introduced.
Re: the specific point that Shaked made, would that health care quality have gotten suddenly worse when vaccines were introduced? If not, then it's very unlikely to be a relevant confounding variable.
Why not? The whole "flatten the curve" argument was that health care resources could be overwhelmed by a surge of patients. In rural counties with a lower number of health care workers per capita a large increase in patient numbers could have detrimental effects on the general standard of care.
I redid a regression and controlled for Obesity rates and % over 65 and the predictive power of dem_percent for the change in excess deaths in 2020 and 2021 goes away. However, this doesn't necessarily mean that these rates can be the cause of the change in death rate. The relationship between Obesity and party voting might be damping down a true benefit of vaccination behavior. But the idea that there might be other explanations other than vaccination behavior driven by party voting can't be ruled out. Obesity rates correlate negatively with percent of democratic votes in 2020 presidential election ... R^2=0.27!
So all we can say is that we can't demonstrate a partisan relationship to the change in death rate over and above the relationship of obesity/age/disability and party voting.
Yes I checked this assumption and the correlation of death rate and %demVotes is negative for 2020 before the vaccines. So the more a state voted Democratic the fewer deaths pre vaccine.
Since the COVID numbers are in question I did an overall death correlation and this effect is slightly stronger so the difference in death rate and official COVID count is definitely not the cause of that finding.
It's good for trusted sources to occasionally demonstrate their rigor on takes to show that they should continue to be trusted. You're right that doing that every time is overkill, but never doing that also allows for take artists to conform their takes to their priors if they never have to go up against strict scrutiny.
Yes, I loved that part. I think you might be off because he's not talking to you.
If you're a super-nerd who over-thinks everything, then you're always going to have a "more robust" analysis up your sleeve. No matter what. You look at everything 12 different ways because you can. And nobody wants to hear about it. Nobody on the internet. Nobody in your business. And nobody in your family or friends group.
It's better, in every situation-- to start with a simple claim that you know is bullet-proof, and put all the gross details in the addenda.
And then just wait for replies before explaining-- "Yes, I looked into that too" or "Oh you want to know about that? Great here is some more."
Glad you at least briefly addressed the first claim in Friday's post, and admitted it might not be as robust.
First, I'll note your brief summary today is technically stronger than your original assertion: here you say NPIs had no effect, whereas Friday your explicit claim was simply that there was little difference in death rates between red and blue states. But the summary, that NPIs had *no* effect, while not explicitly stated Friday, I think reasonably reflects the sense readers would take away from the post.
I think your hostility towards NPIs remains overstated and I would strongly urge you to back off any implication that all NPIs were entirely useless. It is reasonable to at least point out other costs of NPIs, and to question whether the benefits outweigh them, but categorically rejecting them is counterproductive.
First, we know now they vary a lot in effectiveness. Masking is a good example, as N95 masks are far more effective than cloth or lower grade paper masks at reducing COVID spread, but also a poorly worn mask is much less effective than one that fits well. Also even a good, well fitting mask does more to prevent you from spreading COVID to others than it does to reduce your risk of catching it. Masking also reduces the odds of spreading other illnesses, which is why for years mask wearing was non-controversially routine for health care workers in closer contact settings, like dentistry or surgery.
Reducing people gathering also varied in effectiveness. On the one hand, outdoor gatherings such as spring break beach trips or public protests after George Floyd's death did not result in surges of COVID cases many people feared. The better ventilation of being outdoors was sufficient to keep risk of transmission in those settings low. On the other hand, early contact tracing often found links to unmasked indoor gatherings, especially of people from very different areas, like a wedding or conference.
I appreciate the attempt to quantify and simplify analysis by focusing on state-level death rates, but I'd note that especially early in the pandemic, prevalence varied widely across states, while if anything government mandates about NPIs was more uniform. Most of the country put in place mask mandates and bans on large gatherings (even outdoors). We don't have a good control group to compare states with high prevalence that did *not* have lockdowns or mask mandates before vaccination was available, so it is harder to quantify how effective those NPIs were at reducing spread and death.
I am comfortable saying NPIs were less effective than vaccines, even way less effective, but I'd note they were pushed most strongly before vaccines were available, and when less was known about COVID spread. We can and should be smarter about when and where to use, and in particular, mandate them.
At a minimum, I'd say encouraging people to stay home when possible if sick, or at least mask when they have to go out when sick, would be a positive development, not simply with respect to COVID, but infectious disease in general. What helped COVID spread widely was that people were often infectious before becoming symptomatic, and also that it spread primarily by air and not necessarily direct contact.
The next pandemic, however, may spread differently, in which case some NPIs that were largely ineffective for COVID may be more valuable.
Finally, I suggest you not frame things as simply vaccines good, NPIs bad. While I'd agree vaccines have been more effective, NPIs covered a broad range of things, and well targeted NPIs may still have value, especially before vaccines are developed. Stigmatizing NPIs as being inherently ineffective and not worth their costs risks overfitting your model. It is far easier to find in hindsight which interventions were ineffective against this pandemic than it is in the face of a new threat to figure out what works.
It's also entirely possible certain NPIs are not as effective, or not worth their costs, in different environments and cultures!
Is closing indoor recreation like a church (where few people are singing in a chorus) as valuable as closing sport stadiums (with more people, but outdoors)? Who really knows?
Is closing either one, or both of them, worth it, if it means schools can stay open without turning children into vectors of transmission? Who really knows?
Different states did varying degrees of NPIs…NC and Louisiana probably found the best balance. Louisiana’s numbers look bad because it was in the initial wave and its economy has been negatively impacted by numerous unprecedented weather events…but it did a good job.
I'm fine with the second claim but am still not convinced of the first claim. I don't really buy that masks and distancing had zero benefits. I see two potential confounding factors:
1. Possibly blue states had equal deaths pre-vaccine because they got hit earlier. NY and NJ being high in the first list lends to this.
2. Even though the restrictions between states were different, many people in red states may have still been masking and distancing (not sure what the data says in this).
So while some of the blue state restrictions may have been unnecessary, I am nervous people are interpreting this data as saying in the next pandemic we shouldn't bother with masking or distancing.
I don't think you should read that as "masking was ineffective" as much as "making lots of restrictive rules did not meaningfully reduce COVID transmission below people's natural cautiousness."
As others note, Nate's analysis more measures the effects of policy than any specific change in behavior. Look, if during COVID you locked yourself in your apartment and didn't go out, you were pretty unlikely to get COVID, or a cold, or the flu or anything else. Which we DID see spike once the restrictions came off, so clearly some sort of transmissible disease reduction was taking place. So Nate isn't making statements about whether any individual can/should have masked/distanced/isolated for their own safety. Just whether or not mandating it (and a lot of people doing their best to comply, but not everyone equally) had any effect.
So I read the Cochrane review, and while one of the authors has certainly made his opinion clear to the media, THE STUDY ITSELF didn't actually make those conclusions. Like Nate, the actual study, as written, said that there was no discernable difference between more/less restrictive POLICIES, across some of the recent transmissible diseases like SARS, COVID, etc.
In science the burden of proof is on the person advancing the hypothesis. You cannot prove a negative.
If you believe that masks make a difference you must present the evidence that backs up that assertion. All these studies are doing is pointing out that the evidence fails to meet that threshold.
I think that's a distinction without a meaningful difference. Plus given that Cochrane reviewed RCT's on masking--where the focus is on the practice of masking rather than government policy--I'm not sure that actually makes sense.
So, considering and then accepting your statement that there isn’t a meaningful distinction between the efficacy of masking and the policies of masking. Obviously they would be correlated. How about that? Someone on the internet considering a statement and revising their own assertion! Will miracles never cease.
Yes. It was clear from the report that we are way, way overdue for a more rigorous testing of masking, which is shocking in a way. I suspect that while the medical literature is quite clear on the efficacy of say, Dr.'s wearing masks in surgery, we don't know how well, or with what parameters, that might be extended to the general public. E.g., what if the government had shipped everyone a basket on N95's at the beginning of the outbreak instead of people wandering around with bandannas over their mouths? Would that have made a difference? We just don't know.
Yeah I just don’t know how you parse any of that in a data-centered way; these confounders get to things like when states were hit and compliance rates, which are difficult to pin down. I was fine with his initial claim (not much variance=unclear effects) but here he actually says no effects which I think a lack of obvious difference does not support
Yeah I guess he's saying the costs are clear and the benefits are not, so it really deserves its own post examining both of those, because yeah I think there's a big difference between saying we don't know whether NPIs were effective and saying we know NPIs were not effective. But maybe if you make your argument about the costs then that distinction matters less, idk.
If masks and distancing had strong effects, then one should be able to look at curves of death rates and infections and see inflection points where mandatory masking was introduced, and be able to, at a glance, discern between states that masked and that didn’t. I have yet to see examples of that data.
That’s not to say that in highly controlled containment scenarios that masks don’t work, or that for the right individual in the right setting they don’t provide some marginal benefit. It’s just that that benefit is either so minor or so random that its effects fail to emerge at the population level.
Your posts are really well thought out and refreshing to read. I think you admirably try to produce fragments of truth despite the maelstrom of emotional bickering that surrounds politics and COVID.
I am sure you already know this, but I hope to remind you that the insane leftists who have hounded you on COVID are self-selected, very online leftists and surely don't represent the progressive movement. The more rational a person is, the less likely they are to waste their time tweeting etc at prominent people (I'm drunk-writing this right now).
I can imagine it feels like the whole world is ganging up on you when you post on twitter, even if you know that's not the case. So being told this may have little effect, but please let your thoughts and writing be affected by online morons as little as possible.
To be clear I am not criticising this post where you address points made by a person regarded as an expert. I just get the impression it weighs on you at times. I know it can't be helped because you're human, and I wouldn't be able to stand it as well as you can, but I want to remind you that the sane, busy (and sober) majority just reads and doesn't post.
And please keep posting on twitter, and ignore the fickle mob!
Good comment - I hope he reads this. One thing I would say is that it might be reasonable to be concerned that the bullies on the left will push everyone else on the Democratic side in a bad direction, because bullies tend to get their way (just look at the Republicans). So we always have to figure out when and how to stand up to bullies and when to ignore them (i.e. don't feed the trolls).
I appreciate this analysis, but I think one pretty major issue in it is that vaccine uptake is strongly correlated with taking additional precautions - particularly in the 2021-2022 time frame. Amongst people who were getting vaccinated early, they didn’t just decide to go live life like normal right away - there were still restrictions in place. The anti vax crowd living in these red states had very little of that.
So there’s a very strong latent variable here - which admittedly has waned over time - so I’d like to see how this analysis looks since some time in 2022, basically when other preventative measures were eliminated and masking stopped at a broad level in this country.
That's fair except that in the 2020 (pre-vaccine) time frame, even red states were still taking precautions. Obviously this needs to be a lot more robust of an analysis to say for sure but my feeling was around the same time of the vaccine rollout was around when red states effectively let it rip, while blue states were still imposing restrictions. I know the level of restriction was by no means uniform in 2020, but still.
And famously anyone living in Florida or familiar with Florida at the time knows the major population centers of Florida continued to have restrictions for months until DeSantis started to issue rules with preemption language. Broward, Dade, PBC, Hillsborough, Duval, etc all had restrictions in place long after that particular statement by DeSantis.
By that point, a lot of people who weren't vaccinated had already had COVID, which also increases the immune response. And they had it at a higher percentage than those who weren't vaccinated/didn't take precautions. So you'd need to account for that as well.
Republican counties have worse health outcomes than Democratic counties in general.
Disclaimer: I am posting this link as a jumping off point. Personally, my guess is that most of the disparity is probably due to poverty, lower number of health care workers per capita in rural counties, etc.
Why is the US the only country to recommend another round of boosters to everyone over 6 months old, regardless of previous COVID vaccination or health status?
Why has booster uptake declined every round?
How many boosters have you gotten?
The CDC and Pfizer claimed that the COVID jabs prevent transmission and infection, but don’t affect menstrual cycles and breast milk. Since real world data proves all of those claims false, were they spreading lies and misinformation?
Why did Dr. Aseem Malhotra change his mind about the COVID jabs?
How much money have Pfizer and Moderna made on COVID jabs?
Most African countries didn’t have jabs and were fine because covid had a 99%+ survival rate in healthy young people, enjoy bowing to big pharma boosters in perpetuity
Wow, the broadly low African life-expectancy is really a huge benefit, because that means they never live long enough to become truly vulnerable to COVID, get certain cancers, or have to pay for any related medical treatments!
Really an amazing insight! There's always a silver lining!
So, you agree that there's essentially no benefit to the covid vaccines in the young, only in the elderly? I think that most so-called “anti-vaxxers” would be quite happy to agree with you on that.
Young people can and did still get sick from and die due to COVID, just not as often so the benefit was smaller, but still existed. Also, young people can interact with and infect old people who will then die from COVID.
Public health isn't just about individuals, but the *public* at large.
Tell me without telling me that you didn't follow any covid info after ~2021.
Not even Pfizer still claims that their covid vaccine inhibits transmission in any way. They don't even claim it makes the recipient less likely to get covid. It has some benefit in reducing severity of outcomes in the elderly. If you're over-60 and in poor health, it's a good idea. Otherwise, you're wasting your time.
Also, even if the covid vaccine did inhibit transmission (which it doesn't), it's a fundamental principle in medicine that you can't force someone to receive an intervention for someone else's benefit.
You can inject yourself every day if you want. Just don't try to make me do the same.
Protip: almost any internet comment that starts with "So," is drawing some wild conclusion that you can't actually read in the original text. In this case it's still true.
IMO, while there is less benefit to young people for taking the vaxx, I still think the benefits heavily outweigh the costs.
Come on, dude, cost-benefit analysis existing in a variable spectrum isn't exactly an advanced concept.
I like how antivaxxers look solely at COVID death rates and assume that everyone who got COVID and didn’t die are “fine” yet go into full freak out mode over the potential non-death side effects of the vaccine.
1. Because the US health authorities are a larger, more thorough body than other countries.
2. Because non-elderly people are generally tired of hearing about COVID and more likely to simply risk COVID than bother themselves to get follow up vaccines.
3. I'm not Nate, so I don't know.
4. No -- there are literally tens of thousands of possible effects to study, and the fact that eventually 2 additional mild vaccine effects emerged is perfectly normal. Additionally, the NIH was the entity that studied these issues, further proving their commitment to giving the public an accurate picture of the rare/mild vaccine side effects.
5. Because discredited scientists and authors often make waves by making nonsensical claims as a means of extracting money from gullible audiences.
6. To copy another comment from below -- not as much as they deserve.
As far as I can tell, the US decline in boosters is a function of a the fragmented, privatized healthcare system. Already with this latest round of boosters, folks are unsure of whether they will have to pay, or where to find it, and having trouble getting appointments. The most motivated folks are already giving up because the amount of resistance in the system, while less motivated folks might not even try if it's likely to be a multi-day endeavor, or if it's even worth it. This is in combination with how most Americans are already familiar and skeptical of their ability to file a liability claim against an insurance provider if something does go wrong. That federated, profit-based nature makes even aggressive health recommendations from the state less trustworthy because there's less expectation the state will later help undo any harms.
By only asking follow-up questions, are we to assume you don't have any disagreement with Nate's claim, the one that this substack post is actually about? "After vaccines became available, there were clear differences, with red states having higher death rates, almost certainly as a result of lower vaccine uptake among Republicans"?
What he is positing is a two-stage model where partisanship explains vax rates and vax rates and age explain death rates, but he is actually only running a one-stage regression. So in his model including Biden vote share, age, and vax rate, Biden’s vote share coefficient is near zero and statistically insignificant. If you are taking his model seriously, then you should think partisanship is not a good explanation.
That there is a partisan divide in both vax rates (which Silver actually only takes for granted but there is evidence for… just google it) and mortality means there’s is probably something there, but in his previous post on the topic, his own analysis shows an inflection point for difference in mortality rates started in September 2020. That makes me think it’s vax uptake plus behavior.
Instead of Nate Silver nailing it, I’d say he whiffs it"
Also ". Let's abstract away a bit. Suppose you have an outcome variable y and 3 explanatory variables: x1, x2, and x3.
Then you run the model y = a*x1 + b*x2 + c*x3, and you find that the coefficient on x1 is near zero and statistically insignificant and the coefficients on x2 and x3 are large and statistically significant. Additionally, when you drop x1, you find r-squared stays the same."
Well, I'm not Nate Silver himself, but I've been a professional data scientist for about 10 years. I read JFA's comments on MR like 5 times and I still don't understand what JFA actually disagrees with.
JFA says: "If you are taking his [Nate's] model seriously, then you should think partisanship is not a good explanation.... makes me think it’s vax uptake plus behavior"
Nate says: "I don’t mean to imply that COVID is intrinsically more likely to target Republicans or anything like that. Rather, my claim is that COVID is considerably more deadly in people who haven’t been vaccinated, and ... state partisanship serves as a proxy for this"
Same. I have some basic understand of statistics and I didn't understand it either which is I why asked the help of someone with more knowledge. It seems that JFA guy comment is just incomprehensible.
It's not just age. It's also obesity, poverty, worse health care in rural counties, etc. We know (or strongly suspect) that those other factors have a significant effect on health outcomes so they need to be included in any analysis.
As a simple check, run the regression for cancer diagnoses and see what the result is.
Poverty level is the best proxy. Florida has the healthiest elderly population which is why it generally has a low age adjusted flu death rate with Vermont and a few very healthy states beating it out. Basically had DeSantis done what Cooper did in NC FL would have around 20,000 fewer Covid deaths…and NC had just as strong an economy as FL.
Can you do this analysis with excess deaths instead of COVID deaths to see if counting COVID death statistics makes any difference? Though I suspect the states with the higher rates are actually the ones more likely to undercount.
This is a useful analysis, but “vaccines really did work” probably won’t be particularly surprising for people without their heads in the sand.
I feel like there is a lot more room for argument on the “did anything else work?” side. Certainly nothing worked in the way vaccines did, but I would have hoped that was uncontroversial. The underlying patterns are messy before schools reopened. One could suggest that the data are consistent with the idea that dems were more susceptible (e.g. due to community density or living condition density) but sometimes took precautions that neutralized that difference.
Nate, this isn't a Presidential election with an electoral college. States are arbitrary units of observation for this outcome variable – such analyses can only lose information, or distort reality for dumb people.
States vary enormously in population, and in countless other ways, and there are more "red" states than "blue". You're giving a bunch of weight to Southern states – you're basically running up a score when you count very similar Southern states over and over and over.
It would be just as valid to group states together into Silver Regions 1 through 11, and run your silly model. It adds no info to compare state-level variables. If you want to say something about vax efficacy, just cut straight to the science at the human individual and population levels, ideally with a good cost-benefit grouped by age (by decade), sex, and health status.
As always, it helps to zoom out and be a good philosopher. COVID carries no mortality risk for most humans. It's been very difficult to get any details or cost-benefit analyses from US government agencies, but it looks like it doesn't kill healthy people under 50. (I'm being conservative on the age. Obviously what "healthy" comes down to will matter. Again, details are scarce.)
This means that your outcome variable is irrelevant to most people's decisions to take the shots. Now, if we're talking something like hospitalization or a severe symptoms outcome, then we can address those outcomes with solid data – maybe the same pattern will hold, or maybe not.
You're not even scratching the surface here in terms of getting at whether it's rational for someone to take the shots re: their own mortality risk (or others' – they don't prevent transmission; it's possible they reduce it – I've seen estimates of 20% reduction, but *how* it would do this is unexplained in the CDC's junk comms, and I'm not clear on the validity of the measures).
IFR for most is well below 0.1%. Say the shots knock that down to, what, 0.01%? At this point, it probably doesn't matter what the specific value is because humans can't necessarily do anything with differences in probabilities like 0.0001 vs. 0.00008 given all the other factors that we'd think about. Plug in other expected benefits. Then plug in that it's brand new, experimental pharma, maybe existing for six months on this planet at rollout. Humans don't normally take experimental pharma, any pharma that has only existed for six months, or even five years. We've never done this before, and it would be a disastrous habit to normalize the way you seem to want. So against the shaving down of a 0.1% or 0.03% IFR, a person needs to weigh that this is experimental pharma whose effects are largely unknown. (We don't have any new scientific methods or epistemics that can tell us the long-term effects of brand new pharma, nothing like the accelerated aging methods in materials science, structural engineering, etc. We're centuries away from that kind of scientific ability.)
Almost everyone who passes on the shots lives, and avoids any of the known and unknown side effects of them. Maybe 0.1% (probably less) end up dying. It looks like most of those would've been saved by the shots, but a good chuck would've died anyway (I saw a 25% figure recently). But for those who died, they chose to pass on experimental pharma in the face of a 0.1% IFR, weighing the unknown risks heavier. They made their choice, and were unlucky. It looks like a reasonable choice, arguably the most rational choice (and we'll always need some people to make that choice – we couldn't possibly allow 100% of humans or of a country to take experimental pharma). If they were informed, they knew the odds, made their choice. They might make the same choice again, or in any future similar scenario. Nothing about those odds and weights changes if you end up being in the 0.1%, right? People make their choices, and live with them, or not. Counting up deaths in a massive country and presenting relative stats without even discussing the fact that this doesn't apply to most people is just too neurotic – other people's choices are theirs. We don't know anything here about contexts or nature of the deaths. You might be talking about a year of life in many cases, or less, and I'm not sure what high-risk people's contexts might be, and how they're different from my own.
In any case, it's incredibly dangerous to normalize most humans, or even 5%, taking experimental pharma in the dark. This is a "Do NOT do this again" situation. There's no reason to settle for the lack of rigor, the stunning recommendations for young, healthy people, even infants, when there's no mortality risk there (possibly not much severity risk either). Humans need to be a lot more conscious and scientifically serious than this if we want to survive the century. These are the worst-performing vaccines most of us have ever taken, with the possible exception of the flu shots – they don't stop infection, or transmission, they don't last, and their long-term risks are unknown. It's crazy to promote them as the choice of the enlightened. Politics is just so toxic right now, especially leftist/Dem politics and prejudice.
It's also very early. This whole discussion is bogus without accurate COVID deaths data, which you don't have. Ideally we'd have the discipline and rigor to say "We don't have accurate deaths data" and walk away, tell you come back when you do. But I saw the other points as important enough to entertain your bad data. (Localities have warped financial incentives to code deaths as COVID deaths. It's the with/from problem. And the CDC seems to overcount COVID deaths every few months, sometimes children specifically, sometimes total. They've been embarrassingly corrected by stay-at-home moms – true story – which also led to the revision of a peer-reviewed journal article.) This is the kind of thing we don't sort out until years later, so it's more interesting to me to see what we've got in 2027 or so, assuming we don't have a White House that would censor scientific discourse and rigor (the way Biden did on things like side effects, the superior protection of natural immunity, even policy views, etc.)
p.s. Black population is the strongest demographic predictor of your bogus mortality data. I ran it and GOP registration. It's crazy that you didn't control for race or population, just nuts. Blacks are much less vaccinated than Republicans, something like 44% vs 63%. Your pattern of "red" states is strikingly similar to Newsom-style scams on gun violence, homicides, homicide-by-gun on weekends, etc. where he tries to paint "red" states as stained with blood. Much or most of that is driven by blacks in leftist-run Southern cities within "red" states, another example of the arbitrariness of the states comparison.
p.s. 2: You falsely made NH and VA blue in your table, and NC red. This subtly affects the appearance of your ridiculous eyeballing method. You also have Florida, Kentucky, and Texas as red, and Colorado as blue, when they're purple by registration. This wouldn't apply to your regression of fascist Biden voters, but it definitely weakens your color method.
p.s. 3: Most people couldn't get the vax by Feb 1, 2021, so I have no idea why you started there. This is an extremely stupid analysis – starting at some date, sacrificing three bats, etc – but if you're going to do this, you should choose a much later date after it was widely available. I wasn't eligible until the summer. Ideally, you'd run it for several different start dates, say each month of 2021, and see what's what.
We have data by county…in the southeast the counties with the highest vaccination rates tend to be the few ultra wealthy GOP counties and thus they have the lowest Covid death rates. The next counties with below state average death rates are the major urban Democratic counties with large African American populations and young educated people.
And NC has a significantly lower Covid death rate than the adjacent SC and GA because the Democratic governor was a little more aggressive with NPIs in 2020 and then he allowed public health officials to continue them for longer. York county in SC which is a suburb of Charlotte even benefited from Charlotte’s NPIs.
I don’t know why it’s so hard for people to believe that low educated Republicans fell for dangerous disinformation about vaccinations and NPIs ?!?
Hey, incidentally, have you ever examined the demographics of the “low educated” states and counties that didn’t have lots of immediate vaccine uptake, or might have had disproportionately high deaths? I have; I recommend others do so before they start throwing insults.
Yeah, West Virginia is 95% white…duuuuuuuuh. I was involved in GOP politics in the southeast…I know demographics and politics at the county level which is why all of this data is very easy for me to analyze. Basically the higher % of hillbillies in a population the higher the Covid death rate…I might do a medical journal article about the health implications of being a hillbilly. ;)
WV is 95% white with a very high Covid death rate because POVERTY is the biggest risk factor all things being equal. In February 2020 most people would have probably guessed that would be the case because it makes perfect sense. But guess what—with a super young population like Haiti the Covid death rate will never look that bad because as a % the population just doesn’t have a large group of vulnerable people.
The poorer states have lower life expectancy…excess deaths factor in life expectancy of various populations. Poverty essentially prematurely ages Americans which is why % below poverty level is the biggest factor in a population’s Covid death rate all things being equal.
Isn't education level itself one of the biggest predictors of health outcomes in an of itself, along with race and (of course the most important factor) SES?
In other words, doesn't higher education in itself predict better overall health outcomes, even when adjusted for race and SES?
Once again, you are somehow missing the point that education level by itself (just like race by itself) is a clearly recognized social determinant of health, even when adjusted for SES.
Maybe I missed it, but I would be curious to see the same regression analysis run for pre-COVID times (and after the first wave, although it might be considered cherry-picking data).
Also, I would be curious as to what geographic level - county, city, state, etc. - yields more statistically significant results.
Death rates diverged at the point in time at which the vax was introduced. Isn’t that itself strong evidence? Unlikely that the confounders changed substantially at that exact time.
It’s only not strong enough evidence if you are desperate to believe vaccines don’t work. Saying that the “weak” already died early in the pandemic in liberal leaning states is ridiculous. Massive waves of COVID ran through all the states by the time vaccines were available, and the Delta wave in particular was pretty bad in places like Florida. It also doesn’t take into account any amount of natural aging that the population everywhere would be going through... people are getting older every day, more people are becoming diabetic, developing cancer, etc.-- the people pushing anti vaccine arguments are just looking for confirmation that their assumption “the vaccine doesn’t work” is true; desperately flailing for one piece of anecdotal evidence is not going to prove any of them right unless nearly the entire medical establishment is in conspiracy together-- a virtual
Impossibility.
No. Because blue states were hit early in the epidemic. That killed off their most vulnerable and gave some immunity to most of the rest. So as COVID later spread to red states, you’d naturally expect higher mortality in red states than blue during that time period, which overlapped vaccine introduction. Attributing that to the differential vaccine rates is pretty stupid. And I say this as someone who was in the phase 3 Pfizer trial, so not an anti-vaxxer, just an opponent of dopey analysis. Ecological fallacy is another issue, which Nate seems to dismiss for no good reason. Hey Nate, take a look at the mortality rate for blue Navajoland vs red Pinal County and tell me what you see. And as for age, just using 65+ is dopey, because risk to an 85+ person was vastly greater than risk to a 65+ person, so finer granularity, say deciles or better, ventiles.
"There are always other things you can bring up." --Nate
Charles: just bringing up things is useless. Show us your regression analysis.
I have done a gazillion regressions in my day and written FORTRAN programs to do them, not just using R or Stata like twerps today, though I’ve used those too. Point is that there is a very major logical flaw in Nate’s analysis. It’s up to him to think these things through before presenting them as something worth reading.
Here you are bragging about how much you've done, yet you cannot produce an analysis? I think Nate has probably thought this through quite a bit, considering all the work he shows. Whereas I don't see that from you.
Just as the AIDS epidemic needed a sociologist working with a public health officials and statisticians to solve the problem because the experts needed an expert on homosexual behavior to understand spread. To properly analyze this data one needs an intellectually honest political scientist working with other experts. I’m not a political scientist but I happen to know a lot about southeastern demographics and politics and so I can easily conclude that NPIs had a significant impact on lowering the Covid death rate.
Btw, did NPIs work with the AIDS epidemic?? I know in Brazil and India and Africa it was solved by the medications Big Pharma developed.
If I had the data and time of course I could do it. Any idiot can run a regression with R or Python or even Excel, and indeed idiots regularly do. I’ve pretty much lost interest in this sort of analysis because demonstrating a causal relationship is nearly impossible outside an experimental setting. You don’t need to tarry over causality to come up with effective forecasting models, of course, which is what Nate usually does, but here he is trying to demonstrate a causal relationship with an obviously deficient model.
We’re asking this idiot to.
If you’ve lost interest maybe don’t enter the discussion just to muck things up.
There were experimental settings (you were in one!) that found the vaccines reduced mortality substantially. There was then differential uptake by state, and there was also differential mortality by state, in exactly the pattern you'd expect if differences in uptake caused differences in mortality.
I can't randomize states to take or not take the vaccine, so I don't really know what else it would take to convince you
Are you suggesting it took COVID almost an entire year to spread from blue to red states?
Of course not, I’m suggesting that red state infections lagged many months behind blue states, a fact easily observed in the data, which would cause the effect I described. If you look at data from April or May 2020, using Nate’s crude reasoning, you would conclude Trump voters were immune to COVID.
But what does this have to do with vaccine's becoming widespread at the beginning of 2021 and the difference in death rates that emerges then? COVID was clearly already everywhere in the country by then.
The difference didn't emerge then. The difference emerged in summer 2020 and was fairly consistent from that point forward. The difference is mainly caused by obesity, healthcare, and age differences, with obesity being a primary driver. A state obesity chart looks similar to a state covid death rate chart. When vaccines were introduced, the slope of this difference appears to have increased modestly - that's probably the effect of different vaccination rates. But that is NOT the primary driver of differences between red and blue states. The primary drivers are obesity and age, only one of which Nate corrected for here.
This I agree with. Red states are on average more obese than blue states and weight should have been included. But whether the virus arrived in red/blue states in march/april/may isn't a relevant factor assuming all states were saturated by the time vaccines arrived.
Agree. Nate is doing the kind of doofus analysis that data journalists do. Saying that a primitive model is better than a more elaborate one because overfitting can sometimes be a problem is really lame. (And let's not hear any babbling about instrumentation variables as a cure, that's nonsense 99% of the time.)
It sounds like you are making a testable statistical claim. Something like "Cases before the introduction of vaccine are negatively correlated with deaths after the vaccine was available".
And also the claim "Once you control for this vaccination rates no longer predict deaths after the vaccines".
Have you run the numbers on either claim? At least the second seems dubious to me, but I would be happy to change my mind if you ran the analysis.
I suggest you review the ample research on the effects of natural immunity after COVID infection. For example, "Past SARS-CoV-2 infection protection against re-infection: a systematic review and meta-analysis," the Lancet, February 16, 2023: "protection [from prior infection] against severe disease remained high for all variants, with 90·2% (69·7–97·5) for ancestral, alpha, and delta variants, and 88·9% (84·7–90·9) for omicron BA.1 at 40 weeks." 40 WEEKS. So, if persons previously infected (more common in blue states because of the way the epidemic spread) were more likely to be vaccinated (more common in blue states) then you have an obvious confounding problem and will (in Nate's simple model) attribute the differential in mortality to vaccination rates, when prior infection accounts for much of it. That was my first point.
And I should not need to cite research to show that once the most vulnerable have been killed off, the epidemic will not kill the same people again. I think we can agree on that, which was my second point. Most of the vulnerable people in blue states had been killed off before vaccines arrived, while most of the vulnerable people in red state had not been killed off. Another confounder.
So those are two omitted variables obviously confounding any attempt to attribute differential mortality rates to vaccination rates at the state level. Another, as another commentator has observed, is obesity rates. I could add diabetes and other comorbidities as relevant.
If you want an analogy, consider comparing death rates from the Black Death in Italy vs England in 1348 vs 1349. If a useless treatment had been introduced in Italy early in 1349, would we conclude that it protected the Italian population simply because England had a much higher mortality rate in 1349? The real reason, of course, is that Italy was hit earlier in the epidemic.
I am not disputing the existence of natural immunity, or of other factors that influence covid deaths. But the claim that these variables can explain the demonstrated effects of vaccination is a much bigger claim.
It is also analysis that can be done. The data is available. If you think you have some factor or set of factors that can explain Nate's results then you should do the statistical analysis to demonstrate as such.
Nope. Numerous seroprevalence studies have been done. Red states had more antibodies (and thus more previous infections) by the end of 2020.
Which means red states *should* have had an advantage going into 2021. And yet they died at far higher rates.
Cite me one study. This one from JAMA does not support you https://cdn.jamanetwork.com/ama/content_public/journal/intemed/938677/ioi200107f3_1617393574.02664.png?Expires=1698753936&Signature=SzgraLLuW-TtwxrFhBxaCgX4UuIgphLkpxXpHdTuoJXfC-VzQQ52uc8~NDKT75x-~crhtUKTNZC40LRtYwlauhYpkNu7tyz~qmd7zVlkn-hZHIQ7hz5NtpVq7lI6r21axASnaDYq75NEi1NGYSwUjJOElLOBQlurtpGMi01stSCMN3iw6dyZi8eePSl-AayOkKwWppB4-p6hSeQF~s00iqR22ZaI0G7HdSdRkiNdSeW8w2ZYrDiwkPeXkjZQeT-oblbwxcbT~Fd6N5xQoPRml19lQ9HxosxQi5nOqCbf49weWdoP-PQIoZTcIirYqTLTPdUV5cIo165aKS0Fe0vHxg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA
Here's a multitude compiled:
https://covid19serohub.nih.gov/
The one you cited only runs though September 2020, lol (and also shows red states trending higher outside of the top 4).
To close this out, I ran my own regression by state. COVID deaths per million (as of 10/23/23) as the dependent variable. X variables: % of population African American (2020); % of population Urban (2020); % of population 65+ (2020); and % of vote for Trump in 2016 election.
R2 was .60 (adjusted R2 .57). F significance 1.4x10^-8. Intercept and all coefficients significant at less than the .001 level.
Now, you might say that % vote for Trump might be just a proxy for vaccination level. Nope. Add % of population with one or more vaccine doses into the model, and the % vaccine variable has a P value of only .70. If you get rid of the Trump variable and keep the vaccine % variable, the P value for the vaccine variable rises to .0005, which is interesting but still a lower P value than any of the other coefficients. And the adjusted R2 falls to .36, which is far worse than the .57 of the original model.
As for what the % vote for Trump means, it likely means (in my opinion) that Trump voters have less fear of death (consistent with greater religiosity) and therefore are not obsessive in safety measures such as vaccination, social distancing, and so on. I think the Navajo here in Arizona are similar in not being terrified of death, and perhaps the African-American population as well. Is this good or bad? Not for me to judge.
Did you look at the data? Doesn't support your statements. (Also some obvious garbage data, such as a 1.2% seroprevalence for Vermont in early February 2021, when the case count for Vermont exceeded 13,000 (>2.0%) and the number of infections would be, of course, many times greater than the case count, making the estimate off by perhaps an order of magnitude.)
As long as we are talking, if you plan on running a regression, I suggest including percentage of state population that is African American and also Native American, since those subpopulations had much higher COVID death rates than White or Asian subpopulations. Percentage African American is highest in southern states red states), which of course elevates the COVID death rates in those states, despite the subgroup being (as with Native Americans) blue voters. Obvious ecological fallacy going on in analyses that fail to look at that. It's all very interesting but you'll need a very sophisticated model to untangle this stuff. Lots of variables in play.
That is an awfully convoluted story where a much more concise story is adequate. There are also controlled studies that demonstrate lower rates of severe COVID. The burden on your position is well beyond a single counter example and recommending a different age cut. Do better.
Oh for heaven's sake, duh, of course the vaccines reduce mortality. That's not the question. The question is how much of the mortality differential observed between blue and red states after the introduction of vaccines can be attributed to vaccines versus the timeline of the epidemic and other factors. Entirely different question. I have no doubt that some of the differential is attributable to vaccines, but a lot less than Nate's simple analysis suggests. See my response to Not a Lib above for details.
Since we agree on basically everything, why are you arguing secondary effects? This isn't intended for a journal, he's just demonstrating vaccine uptake has macro impact on death rates and uptake has a political component. Look at the SEs on his regression and realize this is not intended as a exercise in precision but instead a blog post.
My point is that it does not demonstrate that, at all. I agree that vaccines reduce individual probability of death from COVID, ceteris paribus, because randomized controlled trials have shown that, but Nate’s analysis does not show that.
Sigh. He's demonstrated different populations have different death rates in a manner that is highly correlated with vaccine uptake while also considering age effects. We have a strong prior based on individual testing that death rates decline with vaccination. Basic first order effect should be that populations with more vaccinations have lower death rates. This correlation is present in the data, even when considering age as an alternative.
He didn't do exactly what your wanted him to do, but he ran a pretty simple set of correlations that reinforce priors based on individual studies. You would need fairly strong evidence to demonstrate political alignment -> vaccination rate -> death rate is unreasonable or insignificant. You point out some reasonable second order channels, but I'd need some math to show they are game changing. I fundamentally don't by the burn out story as many states saw multiple severe waves : LA, AZ
It is not clear at all that the vaccines reduce mortality. There are unknown amounts of people that were dumped into the unvaccinated category upon death if they were not 14 days past the second dose and that is a data crime that skews all subsequent analysis.
Not true.
The JAMA study reported in the Summary Table the raw number of deaths revealing that the age-specific Rep death rate in Ohio was 13% less than the Dem rate but they didn't share how that varied over time.
What they did report was a bizarrely calculated *excess* death rate at the county level that was adjusted by the partisan difference in death rates at the state level.
So the "higher death rates" are only relative to a very rough model of what the rates should have been with the *state level difference erased*. And they didn't share the data or even the unadjusted rates over time. That's unethical and dishonest.
But Nate apparently didn't even read the study because he happily skips directly to "death rates" which he has no information on, unless he secretly obtained the data the authors did not disclose.
Eh? Excess mortality data is readily available, and there are not "corrections" at the state level. Just counting number of dead people, county by county week by week by coroners who who might be red/blue or anything in-between.
Yes, and just counting the bodies showed that Democrats died at higher rates for their age bucket.
And like 90% of the people expressing opinions, you didn't read the study.
They didn't use the "readily available" excess mortality data. They calculated separate baselines FOR EACH PARTY.
Excuse me, I did read the study. But I don’t care about the study since the unmassaged data is readily available. I did my own downloads and some of my own analysis. I was also suggesting that Nate use that data so that people like yourself wouldn’t carp about all this. So… where, in your own proof, did you see that older democrats died at a higher rate? Do you have a worked example? I’d love to see that.
No, we don't know what happened to death rates -- they didn't share that data temporally. We do know from eTable 1 that overall Democrats had higher age-specific death rates.
The PROJECTED death rates were calculated separately for each party so the EXCESS rates started out converged BY DEFINITION. They initially shifted higher in Democrats but then had an inflection point in June 2020 and started shifting higher in Republicans relative to the separate projections. Nothing special happened in April 2021 except that the trend line in each age bucket continued.
https://xkcd.com/2400/
Except more deaths in the vaccine arms in both mRNA studies
thats not evidence the vaccines are not effective. the vast majority of people, especially old and immune-compromised people most susceptible to covid caused death were vaccinated. a much larger percentage of a much smaller group can be a smaller raw value than the smaller percentage of the larger group. sure the vaccines were not 100% effective at preventing death, but literally all available evidence says they made a pretty dramatic difference
More importantly it's not evidence that the vaccines were effective against death (and hospitalization). In fact, the reanalysis combined MRNA and PFE showed that when looking at severe adverse events of special interest (determined prior to the studies), there were more in the treatment arm and it was statistically significant.
I think a little more work needs to be to establish that as a fact. Seems like the inflection point for differential mortality rates was September 2020 (graph c from the previous post). Look at it this way, if the graph wasn't labeled with the vax date where would you say the change in mortality started taking place? In fact, it is a strange coincidence that the excess death rate for Republicans is so much less than for Democrats at the exact time that the vaccine is widely available... I wouldn't make much of that but I also wouldn't take it as pointing to vax rates as the main culprit for the difference.
But NPIs worked great against the Flu in 2020. Surely they also worked for covid??
Hawaii did the best and Puerto Rick did great because of NPIs. So if you believe the GBD halfwits once the NPIs were eliminated Covid would have ravaged those low natural immunity populations…but it didn’t because slowing spread until the availability of the vaccines and then getting everyone vaccinated worked to mitigate severity and prevent hospitalizations and save lives.
What worked for Hawaii & PR was their island status making it easy to prevent importation of the virus. Much as it did for Australia & New Zealand. This was not an option for non-island states. You cannot close the land borders between states in the US. People flowed freely across them and made stopping spread of the virus impossible.
I know that…but it shows NPIs work. Travel restrictions are an NPI and Vermont and Maine and Florida Keys and even Nebraska benefited from effective travel restrictions in that nobody was driving through those states. The Dakotas simply threw caution to the wind and encouraged people to travel to states nobody wanted to travel through and had big spikes because of it.
Travel restrictions only work in very limited scenarios - mainly islands or places that can easily control all travel in and out. Vermont, Maine, & Florida Keys are also not typical places. They are all fairly isolated, wealthy, & healthy.
Vermont and Maine are super old. And so is Puerto Rico which is also poor although I understand PR doesn’t really have a peer state to compare to. Arizona encouraged people to travel to it in winter 2020/21 and it is the largest state with fewest NPIs in 2020 and sure enough it has the highest Covid death rate.
I find this analysis all pretty compelling. Thanks for the follow up, Nate. One item that does catch my eye, though, is this:
>>Are all states counting COVID deaths in the same way? (Probably not.)<<
It probably would be interesting to look at excess mortality instead of COVID deaths. I do recall seeing nation-to-nation comparisons that appears to differ quite a bit when using the former rather than the latter. But I haven't looked into whether the US (or individual states) compile state-level excess mortality numbers.
At least in Iowa, the covid numbers began to get LESS reliable (in the direction of undercounting) as vaccination and vaccine uptake became political footballs. So I would guess that is true in most red states: if you think “covid isn’t real LOLZ” then your numbers are going to underestimate how much worse things got in their states when they (duh) could have gotten better with vaccines.
This sounds reasonable. On the other hand even in more progressive jurisdictions, it's pretty likely, I think, that SARS-CoV-2's contribution to increased mortality has been underestimated: surely substantial numbers of infections go undetected.
I've long suspected it's necessary to look at raw excess mortality numbers (at least for the period 2020-2022) to get an accurate picture as to the pandemic's death toll.
Yep; basically came here to say exactly this. The labeling for deaths is an inexact process that makes it hard to parse data along that dimension. But cats in boxes aside, it’s pretty simple to say if someone is dead or alive. Look at all cause deaths, back out suicides and drug overdoses (I think this an extremely important factor especially in 2020-21) and then see how the numbers look.
You're not wrong (it would be interesting). However, you're also not wrong in that it likely wouldn't change the results all too much. Minor changes aren't going to result in major differences across all states (maybe it results in a measurable change in a few).
I also suspect it wouldn't result in major changes. Though it would be nice to be sure either way. I recall looking at an excess mortality vs. covid death comparison between the UK and Germany that made the latter's advantage over Britain in terms of covid deaths a lot less impressive (perhaps the NHS, for all its faults, is good at record-keeping and accurate reporting of data?).
EDIT: Indeed: if you held a gun to my head and forced me to guess, that guess would be that, if we looked at state-level total excess mortality, the covid toll in the red states has actually been worse vis-a-vis the blue states than we've been led to believe based on covid death numbers alone.
I think your final conclusion is reasonable considering what we know from certain states. If anything (the most glaring example being FL) red states were less likely to call something a covid death and the blue states were more likely. In the end, it was probably relatively minor in terms of this analysis as the total effect would be low. Most (say 90%) were probably reported correctly.
Right, and also, it seems probable that, in jurisdictions where anti-vax and anti-NPI activism were fiercest, less coronavirus testing was occuring, and consequently less identification of SARS-CoV-2 infections.
This is what I came to post, but I am tired, and burnt out discussing the topic, but yes - we can easily compute excess deaths and I have gone into great detail tracking excess deaths/all-cause mortality through several reports the CDC publishes weekly.
Here is a high-level summary in google sheets for anyone interested, all cause deaths, by state, by year, going back to 2014. I also included population (from Excel population data), Obesity (BMI from a CDC report), and then columns M-AB I did a few quick-and-dirty calculations (excess deaths using the "3 year average compare"*, deaths as % of population, deaths per million residents).
https://docs.google.com/spreadsheets/d/1zjKW2k8ODgkoPt61xGsLITLWvRYY2ymZvNzFS9yeRxk/edit?usp=sharing [1]
I don't see the effect with excess deaths Nate sees with covid mortality. Some examples:
Maine averaged 14,589 deaths per million 2017-2019, and that average holds true 2014-2019 as well with little deviation. 2020 they rise to 15,859 deaths, but the "vaccinated years" of 2021 and 2022 they hit 17,095, and 17,076 deaths despite being one of the most highly vaccinated states. Why didn't they return back to ~15,000 deaths projection?
Similar issues for Vermont, Washington, Oregon, New Hampshire, Delaware - all had elevated mortality not recede in 2021 and 2022. Why?
"Red" Ohio has a 14% absolute vaccination difference from "Blue" Pennsylvania (that's quite a bit) [2], but their excess deaths are fairly close at 13% and 16% respectively - and if this 3% absolute difference was the result of a large difference in vaccination, then why did other neighboring states with similar gap in vaccination have identical outcome? Example, South Dakota much higher vaccinated than North Dakota, yet ND has slightly less excess deaths (16% vs 17%) and both countries saw mortality return close to baseline by 2022 - also wasn't South Dakota "ravaged by (Kristi Noem's) 'Trumpian response' to the pandemic" [3] - how did they do so well with all those motorcycle rallies, kids going to school, not wearing masks?
Why did low vaccinated Iowa, the state that didn't "care whether you live or die" (WaPo) [4] have one of the lowest excess death rates in the country? (14%)
Why did California and Florida have near identical excess deaths (21.1% and 21.9%) despite California having a much healthier population (in terms of BMI and other metrics), a younger population, and importantly - these calculations use a fixed population dataset - when we know that Florida increased it's population the last 3 years while California decreased.\
I have talked about this in depth with many "true believers" of (for lack of a better term) the "official narrative" and, I largely get hit with goalpost shifting ("what about population density", "what about early spread in liberal states", etc) and accusations of "cherry picking" - but not only are there a shit ton of cherries, but a good hypothesis should welcome counterfactuals.
It seems to me the base "health" of a state population - obesity, alcohol consumption, age, smoking, location (being in a rural area has long been known as a contributing factor to mortality - too lazy to post myriad of studies covering this - it's common sense) - all seem far more indicative of covid outcome than vaccination rate.
ESPECIALLY because the country level detail offers damning counterfactuals - specifically that South Korea had the 2nd largest year-over-year mortality spike in 100 years, of any country in the world in 2022 - jumping from 310K deaths to 372K deaths.
https://docs.google.com/spreadsheets/d/1wXnEwk4jNuPaQA_dROqHQ_-koEKBzDBNsTKIIS3rwwc/edit?usp=sharing
________________________
* The "3 year average compare" method is similar to how OWID was originally calculating excess deaths before switching to the Karlinsky-Kobach formula which is similar to a linear regression on weekly data.
[1] sources are:
2014-2019 weekly data
https://data.cdc.gov/NCHS/Weekly-Counts-of-Deaths-by-State-and-Select-Causes/3yf8-kanr
2020-2023 weekly data
https://data.cdc.gov/NCHS/Weekly-Provisional-Counts-of-Deaths-by-State-and-S/muzy-jte6
(careful with 2023 data, it lags quite a bit.... 6 weeks back is roughly 90% complete, 12 weeks back 98% complete, etc... they update this every Wednesday
**note that this is weekly data some years (2020, 2015, etc) will have 53 weeks instead of 52. I have a similar analysis through CDC wonder which goes by year, but I like the drill down of weeks to see trends (not included in this google sheet, but the above links provide it to the curious)
[2] I keep meaning to add this data into the spreadsheet, though not sure if still accurate
https://usafacts.org/visualizations/covid-vaccine-tracker-states
[3] https://www.rollingstone.com/politics/politics-features/south-dakota-kristi-noem-covid-1142068/
[4] https://www.washingtonpost.com/outlook/2021/02/10/iowa-lift-all-restrictions/
[5] https://www.usnews.com/news/best-states/articles/2022-04-27/a-third-of-states-lost-population-in-2021-due-to-covid-19-pandemic-report-finds
>2020 they rise to 15,859 deaths, but the "vaccinated years" of 2021 and 2022 they hit 17,095, and 17,076 deaths despite being one of the most highly vaccinated states. Why didn't they return back to ~15,000 deaths projection?<
I can think of a few possibilities: 1) SARS-CoV-2 had become more widespread (edged closer to endemicity) by 2021-2022 than in 2020, when it first arrived. Indeed, nationally, 2021 was the deadliest year for Covid mortality. I'd expect most places in the US saw more excess mortality in 2021 (and perhaps in 2022?) than in 2020. There was simply a much greater chance of becoming infected as the pandemic wore on, and this partly swamped the protective effects of vaccines. 2) Delayed effects of SARS-CoV-2 infections: we know the virus affects cardiovascular health. It seems likely in certain cases the accumulating effects of heart problems took some time to result in premature death. 3) End of social distancing benefits: we've seen pretty clear evidence from Australia and New Zealand that social distancing and NPI practices—while undoubtedly unpleasant and in some cases psychologically damaging—held down numbers of certain kinds of health problems in addition to Covid cases (ie, reduced flu infections, reduced intestinal viruses, reduced car accidents, etc). So, the end of lockdowns would likely result in an uptick in some causes of death, that, combined with the lingering mortality increase associated with SARS-CoV-2 endemicity (we're basically stuck for time being with the mortality equivalent of an extra flu outbreak—and a fairly deadly one at that—per year), resulted in a net increase in overall mortality relative to both the pre-pandemic period and the early period of the pandemic itself (say, first half of 2020).
My guess is that, in the next couple of years, we'll see all cause mortality retreat from the 2021-2022 numbers, but probably not quite get back to the pre 2020 years: we've now got a new disease likely to kill (or contribute to the deaths of) several tens of thousands of Americans annually for the foreseeable future.
Thanks for the feedback and hypothesis testing Charles.
My main point is that when reviewing excess deaths at the state level, - I don't find the original claim Nate proposes ("all states had the same initial outcome, but after vaccinations blue states improved") holds when viewing excess deaths, hence the "counterfactual" examples I pointed to.
A few thoughts on the possibilities you propose:
CR: >>1) SARS-CoV-2 had become more widespread (edged closer to endemicity) by 2021-2022 than in 2020, when it first arrived.<<
We found that Covid spread earlier than we realized. In the US [1], France, [2], China [3][4], and Italy [5], so this hypothesis may or may not be true. Seems to me it was spreading long before we realized it, which is interesting because there was zero signal in excess deaths until we took action.
CR: >>2) Delayed effects of SARS-CoV-2 infections: we know the virus affects cardiovascular health.<<
We "know" the virus affects cardiovascular health because we have scrutinized this virus, and this virus alone, to such a high degree, to search for anything and everything it may or may not do. What we don't have, is the "compared to what?" question, the foundation of scientific skepticism. What if the same methodology applied to Covid was applied to the flu, or one of the other dozens of Coronaviruses, or RSV, etc? Might we also find correlation with cardiovascular issues?
I think of Feynman's take on this "may be true, may not be true" [6]
It seems possible that if searched for this signal with the flu, we may also have found it. We just never looked before.
CR: >>3) End of social distancing benefits: we've seen pretty clear evidence from Australia and New Zealand that social distancing and NPI practices...—held down numbers of certain kinds of health problems in addition to Covid cases (ie, reduced flu infections, reduced intestinal viruses, reduced car accidents, etc<<
A competing hypothesis explains why "flu went away", and it has nothing to do with social distancing and masks - viral interference. I wrote a detailed counter to this hypothesis to Jeremy Faust on his substack here:
https://insidemedicine.substack.com/p/covid-19-metrics-update-august-28/comment/39190372
The "TL;DR" is that places which didn't participate in social distancing and masks also saw Flu disappear. The Nordic countries all returned kids to school, largely unmasked, short lockdowns, and flu disappeared there too. Same in South Dakota, Florida, etc. Additionally we had seen a new virus knock out other viruses before, like Swine Flu did to RSV.
Also you mentioned "reduced car accidents", but that isn't true (at least not for 2020). Surprisingly car accidents increased while we were social distancing.
Additionally, all accidental deaths (sum of car, homicide, suicide, drowning, poisoning, fire, etc) rose significantly in 2020 (+50K year of year) and even increased in 2021 and 2022 (+60K compared to 2019 each year).
CR: >>My guess is that, in the next couple of years, we'll see all cause mortality retreat from the 2021-2022 numbers, but probably not quite get back to the pre 2020 years:<<
It's possible that 2023 sees return to "baseline" mortality, at least in the US. For the first 26 weeks of 2023 we have 1,559,204 deaths. Using linear regression of 2014-2019 first 26 weeks to predict deaths for 2020-2023, we would expect ~1,552,602 deaths at this point. Really just depends on what happens the back half of 2023. First half of 2022 by comparison we had 1,687,056 deaths. Might be possible, I hope.
______________
[1] https://academic.oup.com/cid/article/72/12/e1004/6012472
[2] https://www.sciencedirect.com/science/article/pii/S0924857920301643
[3] https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1009620
[4] https://www.washingtonpost.com/opinions/2021/08/02/new-report-says-covid-emerged-in-wuhan-months-earlier/
[5] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7428442/
[6] https://www.youtube.com/watch?v=tWr39Q9vBgo
"My main point is that when reviewing excess deaths at the state level, - I don't find the original claim Nate proposes ("all states had the same initial outcome, but after vaccinations blue states improved") holds when viewing excess deaths, hence the "counterfactual" examples I pointed to."
This is not what Nate said above.
He said:
"Until vaccines became available, there was little difference in COVID death rates between blue states and red states.
After vaccines became available, there were clear differences, with red states having higher death rates, almost certainly as a result of lower vaccine uptake among Republicans."
So no, not all states had the same outcome in 2020. Just red and blue were not significant variables in 2020.
And more important he did not say that blue states improved in 2021/2022. They just did better than red states.
In general Covid killed a lot more people died in 2021 and 2022 than in 2020. That's because 1) Delta was more virulent and more deadly and 2) It took quite a while to spread to everybody.
Also, obviously, blue and red states only differ modestly. Even the bluest state is 30% Trump, and will still have plenty of anti-vaxxers.
Anyway, if you want to do your analysis like Nate above, using excess death statistics, you can. It won't be true for every single state, but the effects should be similar.
Thanks for pointing out this resource. If nothing else I'm adding obesity to my dataset.
I redid my own regression looking at the predictors of the change in overall death rate as you suggested. I did find a less significant effect than simply running the official COVID deaths but it is significant and in the same direction as Nate found.
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1708.92 1075.24 -1.589 0.11883
Percent.of.residents.who.are.up.to.date.on.their.vaccines 36.87 12.55 2.939 0.00513 **
Age.Percent.65.and.Older -47.96 45.87 -1.045 0.30126
dem_percent 127.58 192.58 0.662 0.51098
Percent getting vaccinated is the only significant variable when run together with %over 65 and %of Democratic Vote. Vaccine uptake is more important than partisanship but those two variables are confounded.
I just looked at the correlation for % vaccinated and Democratic Percentage vote and it's a modest r=0.39 however staff vaccinated is substantially higher at r=0.86.
This difference surprised me to the extent that I reran the regression using %of Staff... and there were no significant effects on the change in 20-21 mortality or COVID Deaths.
Almost significant if you look at the official COVID death numbers change ;-/
All together there is an effect on the change in overall death rates but there does seem to be a little relative increase of the official COVID numbers in "bluer" states. Bias against counting COVID in "reder" States? Or inflation of COVID associated deaths in "blue" states could explain this difference. Either way thanks for making me check my biases.
Here's my spreadsheet. https://docs.google.com/spreadsheets/d/1anDokZ7m_EqoHTOgsJYSHl_xyXowO02azPz0m-HG7TQ/edit?usp=sharing
The mortality actuary "Meep" (Mary Pat Campbell) just posted an interesting rebuttal to Nate's analysis as well. Just going through it now, you may find it interesting.
https://marypatcampbell.substack.com/p/geeking-out-part-2-nate-silver-followup
I'd say it's fun exploration of the data, though not really a rebuttal. Making scatterplots by age category is clever, but it points to the same effect (even the same R^2), for both sets of ages.
I would almost say that this article is a good rebuttal to MPC's approach. That is, if you're arguing on the internet, it makes more sense to -start- by presenting a simple analysis. But you have to go into it -knowing- that your simple model will yield the same results as someone else's more complicated model.
Wouldn't kulldorff's argument also imply that red states should have had higher fatality rates pre-vaccine? So it fails that test too.
Surely there are more variables about disease transmission and co-morbidites than just age, right?
Interesting point. I ran that correlation obesity and death rate and there is a r= 0.25 pre vaccine and 0.53 post vaccine correlation. Way more than age related factors.
Sure (although not that many - weight and preexisting conditions had an effect, but a much, much smaller one than age). But I can't think of any that would have changed right when the vaccines were introduced.
Health care in poor, rural counties is much worse than in richer suburban/urban counties. Look at the disparity in outcomes for cancer diagnoses.
Re: the specific point that Shaked made, would that health care quality have gotten suddenly worse when vaccines were introduced? If not, then it's very unlikely to be a relevant confounding variable.
Why not? The whole "flatten the curve" argument was that health care resources could be overwhelmed by a surge of patients. In rural counties with a lower number of health care workers per capita a large increase in patient numbers could have detrimental effects on the general standard of care.
Behavior is highly related with transmission, and can change quite starkly.
And weight alone is clearly a confounding variable more akin to shoe-size and age.
I redid a regression and controlled for Obesity rates and % over 65 and the predictive power of dem_percent for the change in excess deaths in 2020 and 2021 goes away. However, this doesn't necessarily mean that these rates can be the cause of the change in death rate. The relationship between Obesity and party voting might be damping down a true benefit of vaccination behavior. But the idea that there might be other explanations other than vaccination behavior driven by party voting can't be ruled out. Obesity rates correlate negatively with percent of democratic votes in 2020 presidential election ... R^2=0.27!
So all we can say is that we can't demonstrate a partisan relationship to the change in death rate over and above the relationship of obesity/age/disability and party voting.
P= .058 for the official COVID death rate change after this correction.
Yes I checked this assumption and the correlation of death rate and %demVotes is negative for 2020 before the vaccines. So the more a state voted Democratic the fewer deaths pre vaccine.
Since the COVID numbers are in question I did an overall death correlation and this effect is slightly stronger so the difference in death rate and official COVID count is definitely not the cause of that finding.
It's good for trusted sources to occasionally demonstrate their rigor on takes to show that they should continue to be trusted. You're right that doing that every time is overkill, but never doing that also allows for take artists to conform their takes to their priors if they never have to go up against strict scrutiny.
Yes, or at least share a robust analysis that you find convincing.
Amazing admission here: Yes this analysis is schlock but don't worry I did a more robust analysis that would only confuse my target audience.
Yes, I loved that part. I think you might be off because he's not talking to you.
If you're a super-nerd who over-thinks everything, then you're always going to have a "more robust" analysis up your sleeve. No matter what. You look at everything 12 different ways because you can. And nobody wants to hear about it. Nobody on the internet. Nobody in your business. And nobody in your family or friends group.
It's better, in every situation-- to start with a simple claim that you know is bullet-proof, and put all the gross details in the addenda.
And then just wait for replies before explaining-- "Yes, I looked into that too" or "Oh you want to know about that? Great here is some more."
Good point.
Glad you at least briefly addressed the first claim in Friday's post, and admitted it might not be as robust.
First, I'll note your brief summary today is technically stronger than your original assertion: here you say NPIs had no effect, whereas Friday your explicit claim was simply that there was little difference in death rates between red and blue states. But the summary, that NPIs had *no* effect, while not explicitly stated Friday, I think reasonably reflects the sense readers would take away from the post.
I think your hostility towards NPIs remains overstated and I would strongly urge you to back off any implication that all NPIs were entirely useless. It is reasonable to at least point out other costs of NPIs, and to question whether the benefits outweigh them, but categorically rejecting them is counterproductive.
First, we know now they vary a lot in effectiveness. Masking is a good example, as N95 masks are far more effective than cloth or lower grade paper masks at reducing COVID spread, but also a poorly worn mask is much less effective than one that fits well. Also even a good, well fitting mask does more to prevent you from spreading COVID to others than it does to reduce your risk of catching it. Masking also reduces the odds of spreading other illnesses, which is why for years mask wearing was non-controversially routine for health care workers in closer contact settings, like dentistry or surgery.
Reducing people gathering also varied in effectiveness. On the one hand, outdoor gatherings such as spring break beach trips or public protests after George Floyd's death did not result in surges of COVID cases many people feared. The better ventilation of being outdoors was sufficient to keep risk of transmission in those settings low. On the other hand, early contact tracing often found links to unmasked indoor gatherings, especially of people from very different areas, like a wedding or conference.
I appreciate the attempt to quantify and simplify analysis by focusing on state-level death rates, but I'd note that especially early in the pandemic, prevalence varied widely across states, while if anything government mandates about NPIs was more uniform. Most of the country put in place mask mandates and bans on large gatherings (even outdoors). We don't have a good control group to compare states with high prevalence that did *not* have lockdowns or mask mandates before vaccination was available, so it is harder to quantify how effective those NPIs were at reducing spread and death.
I am comfortable saying NPIs were less effective than vaccines, even way less effective, but I'd note they were pushed most strongly before vaccines were available, and when less was known about COVID spread. We can and should be smarter about when and where to use, and in particular, mandate them.
At a minimum, I'd say encouraging people to stay home when possible if sick, or at least mask when they have to go out when sick, would be a positive development, not simply with respect to COVID, but infectious disease in general. What helped COVID spread widely was that people were often infectious before becoming symptomatic, and also that it spread primarily by air and not necessarily direct contact.
The next pandemic, however, may spread differently, in which case some NPIs that were largely ineffective for COVID may be more valuable.
Finally, I suggest you not frame things as simply vaccines good, NPIs bad. While I'd agree vaccines have been more effective, NPIs covered a broad range of things, and well targeted NPIs may still have value, especially before vaccines are developed. Stigmatizing NPIs as being inherently ineffective and not worth their costs risks overfitting your model. It is far easier to find in hindsight which interventions were ineffective against this pandemic than it is in the face of a new threat to figure out what works.
It's also entirely possible certain NPIs are not as effective, or not worth their costs, in different environments and cultures!
Is closing indoor recreation like a church (where few people are singing in a chorus) as valuable as closing sport stadiums (with more people, but outdoors)? Who really knows?
Is closing either one, or both of them, worth it, if it means schools can stay open without turning children into vectors of transmission? Who really knows?
Different states did varying degrees of NPIs…NC and Louisiana probably found the best balance. Louisiana’s numbers look bad because it was in the initial wave and its economy has been negatively impacted by numerous unprecedented weather events…but it did a good job.
I'm fine with the second claim but am still not convinced of the first claim. I don't really buy that masks and distancing had zero benefits. I see two potential confounding factors:
1. Possibly blue states had equal deaths pre-vaccine because they got hit earlier. NY and NJ being high in the first list lends to this.
2. Even though the restrictions between states were different, many people in red states may have still been masking and distancing (not sure what the data says in this).
So while some of the blue state restrictions may have been unnecessary, I am nervous people are interpreting this data as saying in the next pandemic we shouldn't bother with masking or distancing.
I don't think you should read that as "masking was ineffective" as much as "making lots of restrictive rules did not meaningfully reduce COVID transmission below people's natural cautiousness."
There's a known effect (https://ethicsunwrapped.utexas.edu/glossary/moral-equilibrium) where people will compensate for virtuous behaviors with viceful behaviors, and I have to wonder if that negated a lot of the blue state NPI rules.
As others note, Nate's analysis more measures the effects of policy than any specific change in behavior. Look, if during COVID you locked yourself in your apartment and didn't go out, you were pretty unlikely to get COVID, or a cold, or the flu or anything else. Which we DID see spike once the restrictions came off, so clearly some sort of transmissible disease reduction was taking place. So Nate isn't making statements about whether any individual can/should have masked/distanced/isolated for their own safety. Just whether or not mandating it (and a lot of people doing their best to comply, but not everyone equally) had any effect.
The real reason to question the efficacy of masking is studies like the Cochrane review.
So I read the Cochrane review, and while one of the authors has certainly made his opinion clear to the media, THE STUDY ITSELF didn't actually make those conclusions. Like Nate, the actual study, as written, said that there was no discernable difference between more/less restrictive POLICIES, across some of the recent transmissible diseases like SARS, COVID, etc.
In science the burden of proof is on the person advancing the hypothesis. You cannot prove a negative.
If you believe that masks make a difference you must present the evidence that backs up that assertion. All these studies are doing is pointing out that the evidence fails to meet that threshold.
Yes, agreed. Just that the specific conclusions of the Cochrane paper was that the evidence in support of these POLICIES failed to hold up.
I think that's a distinction without a meaningful difference. Plus given that Cochrane reviewed RCT's on masking--where the focus is on the practice of masking rather than government policy--I'm not sure that actually makes sense.
So, considering and then accepting your statement that there isn’t a meaningful distinction between the efficacy of masking and the policies of masking. Obviously they would be correlated. How about that? Someone on the internet considering a statement and revising their own assertion! Will miracles never cease.
Yes. It was clear from the report that we are way, way overdue for a more rigorous testing of masking, which is shocking in a way. I suspect that while the medical literature is quite clear on the efficacy of say, Dr.'s wearing masks in surgery, we don't know how well, or with what parameters, that might be extended to the general public. E.g., what if the government had shipped everyone a basket on N95's at the beginning of the outbreak instead of people wandering around with bandannas over their mouths? Would that have made a difference? We just don't know.
Yes, trust the experts!! Lol, you keep making the same mistake. Let’s invade Iraq because the “experts” said we should do it!! ;)
Yeah I just don’t know how you parse any of that in a data-centered way; these confounders get to things like when states were hit and compliance rates, which are difficult to pin down. I was fine with his initial claim (not much variance=unclear effects) but here he actually says no effects which I think a lack of obvious difference does not support
Yeah I guess he's saying the costs are clear and the benefits are not, so it really deserves its own post examining both of those, because yeah I think there's a big difference between saying we don't know whether NPIs were effective and saying we know NPIs were not effective. But maybe if you make your argument about the costs then that distinction matters less, idk.
If masks and distancing had strong effects, then one should be able to look at curves of death rates and infections and see inflection points where mandatory masking was introduced, and be able to, at a glance, discern between states that masked and that didn’t. I have yet to see examples of that data.
That’s not to say that in highly controlled containment scenarios that masks don’t work, or that for the right individual in the right setting they don’t provide some marginal benefit. It’s just that that benefit is either so minor or so random that its effects fail to emerge at the population level.
Your posts are really well thought out and refreshing to read. I think you admirably try to produce fragments of truth despite the maelstrom of emotional bickering that surrounds politics and COVID.
I am sure you already know this, but I hope to remind you that the insane leftists who have hounded you on COVID are self-selected, very online leftists and surely don't represent the progressive movement. The more rational a person is, the less likely they are to waste their time tweeting etc at prominent people (I'm drunk-writing this right now).
I can imagine it feels like the whole world is ganging up on you when you post on twitter, even if you know that's not the case. So being told this may have little effect, but please let your thoughts and writing be affected by online morons as little as possible.
To be clear I am not criticising this post where you address points made by a person regarded as an expert. I just get the impression it weighs on you at times. I know it can't be helped because you're human, and I wouldn't be able to stand it as well as you can, but I want to remind you that the sane, busy (and sober) majority just reads and doesn't post.
And please keep posting on twitter, and ignore the fickle mob!
Good comment - I hope he reads this. One thing I would say is that it might be reasonable to be concerned that the bullies on the left will push everyone else on the Democratic side in a bad direction, because bullies tend to get their way (just look at the Republicans). So we always have to figure out when and how to stand up to bullies and when to ignore them (i.e. don't feed the trolls).
I appreciate this analysis, but I think one pretty major issue in it is that vaccine uptake is strongly correlated with taking additional precautions - particularly in the 2021-2022 time frame. Amongst people who were getting vaccinated early, they didn’t just decide to go live life like normal right away - there were still restrictions in place. The anti vax crowd living in these red states had very little of that.
So there’s a very strong latent variable here - which admittedly has waned over time - so I’d like to see how this analysis looks since some time in 2022, basically when other preventative measures were eliminated and masking stopped at a broad level in this country.
The lack of variability pre-vaccine seems like (moderate) evidence that correlated behaviors didn't have much effect though.
That's fair except that in the 2020 (pre-vaccine) time frame, even red states were still taking precautions. Obviously this needs to be a lot more robust of an analysis to say for sure but my feeling was around the same time of the vaccine rollout was around when red states effectively let it rip, while blue states were still imposing restrictions. I know the level of restriction was by no means uniform in 2020, but still.
DeSantis famously said Florida was 'open for business' in September 2020.
And famously anyone living in Florida or familiar with Florida at the time knows the major population centers of Florida continued to have restrictions for months until DeSantis started to issue rules with preemption language. Broward, Dade, PBC, Hillsborough, Duval, etc all had restrictions in place long after that particular statement by DeSantis.
By that point, a lot of people who weren't vaccinated had already had COVID, which also increases the immune response. And they had it at a higher percentage than those who weren't vaccinated/didn't take precautions. So you'd need to account for that as well.
https://www.scientificamerican.com/article/people-in-republican-counties-have-higher-death-rates-than-those-in-democratic-counties/#:~:text=Sehgal%20and%20his%20colleagues%20found,relative%20to%20majority%2DDemocratic%20counties.
Republican counties have worse health outcomes than Democratic counties in general.
Disclaimer: I am posting this link as a jumping off point. Personally, my guess is that most of the disparity is probably due to poverty, lower number of health care workers per capita in rural counties, etc.
My data gathering and subsequent regressions could take a look at this general health/poverty question.
https://docs.google.com/spreadsheets/d/e/2PACX-1vTEImSvXyduS_-l7QdNtZHYDpYCTLOQFRladPAlIXNaA3TslGfcuPFva6mMO0BpQkkRU7KgRvLnbjtE/pub?gid=0&single=true&output=csv
A few follow up questions:
Why is the US the only country to recommend another round of boosters to everyone over 6 months old, regardless of previous COVID vaccination or health status?
Why has booster uptake declined every round?
How many boosters have you gotten?
The CDC and Pfizer claimed that the COVID jabs prevent transmission and infection, but don’t affect menstrual cycles and breast milk. Since real world data proves all of those claims false, were they spreading lies and misinformation?
Why did Dr. Aseem Malhotra change his mind about the COVID jabs?
How much money have Pfizer and Moderna made on COVID jabs?
I can answer the last one - not enough, they should've gotten easily ten times as much for saving the world from a global pandemic.
Most African countries didn’t have jabs and were fine because covid had a 99%+ survival rate in healthy young people, enjoy bowing to big pharma boosters in perpetuity
Wow, the broadly low African life-expectancy is really a huge benefit, because that means they never live long enough to become truly vulnerable to COVID, get certain cancers, or have to pay for any related medical treatments!
Really an amazing insight! There's always a silver lining!
Just think how much money the society in Logan’s Run saves in entitlements. And no medical care induced bankruptcies!
Oh shit my palm diamond just changed.
So, you agree that there's essentially no benefit to the covid vaccines in the young, only in the elderly? I think that most so-called “anti-vaxxers” would be quite happy to agree with you on that.
Nope, try again.
Young people can and did still get sick from and die due to COVID, just not as often so the benefit was smaller, but still existed. Also, young people can interact with and infect old people who will then die from COVID.
Public health isn't just about individuals, but the *public* at large.
Tell me without telling me that you didn't follow any covid info after ~2021.
Not even Pfizer still claims that their covid vaccine inhibits transmission in any way. They don't even claim it makes the recipient less likely to get covid. It has some benefit in reducing severity of outcomes in the elderly. If you're over-60 and in poor health, it's a good idea. Otherwise, you're wasting your time.
Also, even if the covid vaccine did inhibit transmission (which it doesn't), it's a fundamental principle in medicine that you can't force someone to receive an intervention for someone else's benefit.
You can inject yourself every day if you want. Just don't try to make me do the same.
Protip: almost any internet comment that starts with "So," is drawing some wild conclusion that you can't actually read in the original text. In this case it's still true.
IMO, while there is less benefit to young people for taking the vaxx, I still think the benefits heavily outweigh the costs.
Come on, dude, cost-benefit analysis existing in a variable spectrum isn't exactly an advanced concept.
I like how antivaxxers look solely at COVID death rates and assume that everyone who got COVID and didn’t die are “fine” yet go into full freak out mode over the potential non-death side effects of the vaccine.
Yup!
1. Because the US health authorities are a larger, more thorough body than other countries.
2. Because non-elderly people are generally tired of hearing about COVID and more likely to simply risk COVID than bother themselves to get follow up vaccines.
3. I'm not Nate, so I don't know.
4. No -- there are literally tens of thousands of possible effects to study, and the fact that eventually 2 additional mild vaccine effects emerged is perfectly normal. Additionally, the NIH was the entity that studied these issues, further proving their commitment to giving the public an accurate picture of the rare/mild vaccine side effects.
5. Because discredited scientists and authors often make waves by making nonsensical claims as a means of extracting money from gullible audiences.
6. To copy another comment from below -- not as much as they deserve.
As far as I can tell, the US decline in boosters is a function of a the fragmented, privatized healthcare system. Already with this latest round of boosters, folks are unsure of whether they will have to pay, or where to find it, and having trouble getting appointments. The most motivated folks are already giving up because the amount of resistance in the system, while less motivated folks might not even try if it's likely to be a multi-day endeavor, or if it's even worth it. This is in combination with how most Americans are already familiar and skeptical of their ability to file a liability claim against an insurance provider if something does go wrong. That federated, profit-based nature makes even aggressive health recommendations from the state less trustworthy because there's less expectation the state will later help undo any harms.
If the vaccine works, do you think it's wrong for the companies to make a lot of money on it? I mean, some might, but do you?
By only asking follow-up questions, are we to assume you don't have any disagreement with Nate's claim, the one that this substack post is actually about? "After vaccines became available, there were clear differences, with red states having higher death rates, almost certainly as a result of lower vaccine uptake among Republicans"?
Your piece was posted on MarginalRevolution.com and the top comment right now is this. I am not statistically literate enough to know if he is right so I'd like if someone (preferrably Nate Silver himself) who is, to respond to this (https://marginalrevolution.com/marginalrevolution/2023/10/sunday-assorted-links-438.html?commentID=160661089)
"Silver isn’t quite as convincing as he thinks.
What he is positing is a two-stage model where partisanship explains vax rates and vax rates and age explain death rates, but he is actually only running a one-stage regression. So in his model including Biden vote share, age, and vax rate, Biden’s vote share coefficient is near zero and statistically insignificant. If you are taking his model seriously, then you should think partisanship is not a good explanation.
That there is a partisan divide in both vax rates (which Silver actually only takes for granted but there is evidence for… just google it) and mortality means there’s is probably something there, but in his previous post on the topic, his own analysis shows an inflection point for difference in mortality rates started in September 2020. That makes me think it’s vax uptake plus behavior.
Instead of Nate Silver nailing it, I’d say he whiffs it"
Also ". Let's abstract away a bit. Suppose you have an outcome variable y and 3 explanatory variables: x1, x2, and x3.
Then you run the model y = a*x1 + b*x2 + c*x3, and you find that the coefficient on x1 is near zero and statistically insignificant and the coefficients on x2 and x3 are large and statistically significant. Additionally, when you drop x1, you find r-squared stays the same."
Well, I'm not Nate Silver himself, but I've been a professional data scientist for about 10 years. I read JFA's comments on MR like 5 times and I still don't understand what JFA actually disagrees with.
JFA says: "If you are taking his [Nate's] model seriously, then you should think partisanship is not a good explanation.... makes me think it’s vax uptake plus behavior"
Nate says: "I don’t mean to imply that COVID is intrinsically more likely to target Republicans or anything like that. Rather, my claim is that COVID is considerably more deadly in people who haven’t been vaccinated, and ... state partisanship serves as a proxy for this"
Same. I have some basic understand of statistics and I didn't understand it either which is I why asked the help of someone with more knowledge. It seems that JFA guy comment is just incomprehensible.
It's not just age. It's also obesity, poverty, worse health care in rural counties, etc. We know (or strongly suspect) that those other factors have a significant effect on health outcomes so they need to be included in any analysis.
As a simple check, run the regression for cancer diagnoses and see what the result is.
Poverty level is the best proxy. Florida has the healthiest elderly population which is why it generally has a low age adjusted flu death rate with Vermont and a few very healthy states beating it out. Basically had DeSantis done what Cooper did in NC FL would have around 20,000 fewer Covid deaths…and NC had just as strong an economy as FL.
https://www.seattletimes.com/nation-world/why-major-study-argues-floridas-covid-death-rate-compares-favorably-to-californias/
Keep listening to the experts! ;)
Can you do this analysis with excess deaths instead of COVID deaths to see if counting COVID death statistics makes any difference? Though I suspect the states with the higher rates are actually the ones more likely to undercount.
This is a useful analysis, but “vaccines really did work” probably won’t be particularly surprising for people without their heads in the sand.
I feel like there is a lot more room for argument on the “did anything else work?” side. Certainly nothing worked in the way vaccines did, but I would have hoped that was uncontroversial. The underlying patterns are messy before schools reopened. One could suggest that the data are consistent with the idea that dems were more susceptible (e.g. due to community density or living condition density) but sometimes took precautions that neutralized that difference.
Nate, this isn't a Presidential election with an electoral college. States are arbitrary units of observation for this outcome variable – such analyses can only lose information, or distort reality for dumb people.
States vary enormously in population, and in countless other ways, and there are more "red" states than "blue". You're giving a bunch of weight to Southern states – you're basically running up a score when you count very similar Southern states over and over and over.
It would be just as valid to group states together into Silver Regions 1 through 11, and run your silly model. It adds no info to compare state-level variables. If you want to say something about vax efficacy, just cut straight to the science at the human individual and population levels, ideally with a good cost-benefit grouped by age (by decade), sex, and health status.
As always, it helps to zoom out and be a good philosopher. COVID carries no mortality risk for most humans. It's been very difficult to get any details or cost-benefit analyses from US government agencies, but it looks like it doesn't kill healthy people under 50. (I'm being conservative on the age. Obviously what "healthy" comes down to will matter. Again, details are scarce.)
This means that your outcome variable is irrelevant to most people's decisions to take the shots. Now, if we're talking something like hospitalization or a severe symptoms outcome, then we can address those outcomes with solid data – maybe the same pattern will hold, or maybe not.
You're not even scratching the surface here in terms of getting at whether it's rational for someone to take the shots re: their own mortality risk (or others' – they don't prevent transmission; it's possible they reduce it – I've seen estimates of 20% reduction, but *how* it would do this is unexplained in the CDC's junk comms, and I'm not clear on the validity of the measures).
IFR for most is well below 0.1%. Say the shots knock that down to, what, 0.01%? At this point, it probably doesn't matter what the specific value is because humans can't necessarily do anything with differences in probabilities like 0.0001 vs. 0.00008 given all the other factors that we'd think about. Plug in other expected benefits. Then plug in that it's brand new, experimental pharma, maybe existing for six months on this planet at rollout. Humans don't normally take experimental pharma, any pharma that has only existed for six months, or even five years. We've never done this before, and it would be a disastrous habit to normalize the way you seem to want. So against the shaving down of a 0.1% or 0.03% IFR, a person needs to weigh that this is experimental pharma whose effects are largely unknown. (We don't have any new scientific methods or epistemics that can tell us the long-term effects of brand new pharma, nothing like the accelerated aging methods in materials science, structural engineering, etc. We're centuries away from that kind of scientific ability.)
Almost everyone who passes on the shots lives, and avoids any of the known and unknown side effects of them. Maybe 0.1% (probably less) end up dying. It looks like most of those would've been saved by the shots, but a good chuck would've died anyway (I saw a 25% figure recently). But for those who died, they chose to pass on experimental pharma in the face of a 0.1% IFR, weighing the unknown risks heavier. They made their choice, and were unlucky. It looks like a reasonable choice, arguably the most rational choice (and we'll always need some people to make that choice – we couldn't possibly allow 100% of humans or of a country to take experimental pharma). If they were informed, they knew the odds, made their choice. They might make the same choice again, or in any future similar scenario. Nothing about those odds and weights changes if you end up being in the 0.1%, right? People make their choices, and live with them, or not. Counting up deaths in a massive country and presenting relative stats without even discussing the fact that this doesn't apply to most people is just too neurotic – other people's choices are theirs. We don't know anything here about contexts or nature of the deaths. You might be talking about a year of life in many cases, or less, and I'm not sure what high-risk people's contexts might be, and how they're different from my own.
In any case, it's incredibly dangerous to normalize most humans, or even 5%, taking experimental pharma in the dark. This is a "Do NOT do this again" situation. There's no reason to settle for the lack of rigor, the stunning recommendations for young, healthy people, even infants, when there's no mortality risk there (possibly not much severity risk either). Humans need to be a lot more conscious and scientifically serious than this if we want to survive the century. These are the worst-performing vaccines most of us have ever taken, with the possible exception of the flu shots – they don't stop infection, or transmission, they don't last, and their long-term risks are unknown. It's crazy to promote them as the choice of the enlightened. Politics is just so toxic right now, especially leftist/Dem politics and prejudice.
It's also very early. This whole discussion is bogus without accurate COVID deaths data, which you don't have. Ideally we'd have the discipline and rigor to say "We don't have accurate deaths data" and walk away, tell you come back when you do. But I saw the other points as important enough to entertain your bad data. (Localities have warped financial incentives to code deaths as COVID deaths. It's the with/from problem. And the CDC seems to overcount COVID deaths every few months, sometimes children specifically, sometimes total. They've been embarrassingly corrected by stay-at-home moms – true story – which also led to the revision of a peer-reviewed journal article.) This is the kind of thing we don't sort out until years later, so it's more interesting to me to see what we've got in 2027 or so, assuming we don't have a White House that would censor scientific discourse and rigor (the way Biden did on things like side effects, the superior protection of natural immunity, even policy views, etc.)
p.s. Black population is the strongest demographic predictor of your bogus mortality data. I ran it and GOP registration. It's crazy that you didn't control for race or population, just nuts. Blacks are much less vaccinated than Republicans, something like 44% vs 63%. Your pattern of "red" states is strikingly similar to Newsom-style scams on gun violence, homicides, homicide-by-gun on weekends, etc. where he tries to paint "red" states as stained with blood. Much or most of that is driven by blacks in leftist-run Southern cities within "red" states, another example of the arbitrariness of the states comparison.
p.s. 2: You falsely made NH and VA blue in your table, and NC red. This subtly affects the appearance of your ridiculous eyeballing method. You also have Florida, Kentucky, and Texas as red, and Colorado as blue, when they're purple by registration. This wouldn't apply to your regression of fascist Biden voters, but it definitely weakens your color method.
p.s. 3: Most people couldn't get the vax by Feb 1, 2021, so I have no idea why you started there. This is an extremely stupid analysis – starting at some date, sacrificing three bats, etc – but if you're going to do this, you should choose a much later date after it was widely available. I wasn't eligible until the summer. Ideally, you'd run it for several different start dates, say each month of 2021, and see what's what.
We have data by county…in the southeast the counties with the highest vaccination rates tend to be the few ultra wealthy GOP counties and thus they have the lowest Covid death rates. The next counties with below state average death rates are the major urban Democratic counties with large African American populations and young educated people.
And NC has a significantly lower Covid death rate than the adjacent SC and GA because the Democratic governor was a little more aggressive with NPIs in 2020 and then he allowed public health officials to continue them for longer. York county in SC which is a suburb of Charlotte even benefited from Charlotte’s NPIs.
I don’t know why it’s so hard for people to believe that low educated Republicans fell for dangerous disinformation about vaccinations and NPIs ?!?
Hey, incidentally, have you ever examined the demographics of the “low educated” states and counties that didn’t have lots of immediate vaccine uptake, or might have had disproportionately high deaths? I have; I recommend others do so before they start throwing insults.
Also can you travel between different states?
Yeah, West Virginia is 95% white…duuuuuuuuh. I was involved in GOP politics in the southeast…I know demographics and politics at the county level which is why all of this data is very easy for me to analyze. Basically the higher % of hillbillies in a population the higher the Covid death rate…I might do a medical journal article about the health implications of being a hillbilly. ;)
Oh, I get it, you’re a troll! Phew, that’s usually much more difficult to figure out. Saved me a bunch of time and mental energy, appreciate it.
WV is 95% white with a very high Covid death rate because POVERTY is the biggest risk factor all things being equal. In February 2020 most people would have probably guessed that would be the case because it makes perfect sense. But guess what—with a super young population like Haiti the Covid death rate will never look that bad because as a % the population just doesn’t have a large group of vulnerable people.
How to explain low covid deaths in Equatorial Africa?
Ironically, Haiti’s poverty might’ve helped them in this case: https://www.iie.org/programs/iie-centennial-fellowship/centennial-fellows-blogs/2020-blog/jean-max-charles-3.
If they are all that dumb, wouldn't we expect them to have a higher mortality rate on that basis alone?
Isn't education level itself one of the biggest predictors of health outcomes along with SES?
The poorer states have lower life expectancy…excess deaths factor in life expectancy of various populations. Poverty essentially prematurely ages Americans which is why % below poverty level is the biggest factor in a population’s Covid death rate all things being equal.
Isn't education level itself one of the biggest predictors of health outcomes in an of itself, along with race and (of course the most important factor) SES?
In other words, doesn't higher education in itself predict better overall health outcomes, even when adjusted for race and SES?
Once again, poverty includes education levels.
Once again, you are somehow missing the point that education level by itself (just like race by itself) is a clearly recognized social determinant of health, even when adjusted for SES.
Maybe I missed it, but I would be curious to see the same regression analysis run for pre-COVID times (and after the first wave, although it might be considered cherry-picking data).
Also, I would be curious as to what geographic level - county, city, state, etc. - yields more statistically significant results.