Vaccinations

Regular readers have no doubt noticed that something like 80% of the new posts here at ACA Signups over the past 2-3 weeks have been obsessively tracking COVID-19 vaccination rates via various metrics (partisanship, income, geographic region, education level, etc). While COVID and the vaccination program are obviously heavily healthcare policy-related, they're also obviously not right at the core of what this site is normally about.

This week I'm finally easing off on the vaccination tracking stuff (I'll only be posting about them here weekly going forward--on Wednesdays--with rare exceptions), I wanted to explain why I've been so obsessive about this. A Twitter thread by nurse Julia Pulver (an old friend of mine here in Michigan) explains it better than I could. I've converted her thread into a more blog-friendly format with her permission:

One of the most traumatic experiences I ever had in an ICU was performing end of life care for a 34yr mother of 3. She had advanced breast cancer & there was nothing more that could be done-she was in multi-system organ failure.

NOTE: The original version of this post included a serious, bone-headed data error on my part, requiring me to pull the post and revamp it just minutes after it went live. After the main post I'll explain how I screwed up and how I've resolved the issue. The bottom line is that my premise may still be correct, but if so it won't be nearly as dramatically as I had originally thought.

As I noted a week ago, something very interesting has changed on the GOP side of the political and media punditry aisle:

Generally speaking, however, it sure sounds to me like someone in the GOP now believes that their 6-month anti-vaxx propaganda campaign is starting to kill off their own voter base, because everyone from Sean Hannity to Florida Governor Ron "Don't Fauci my Florida" DeSantis seem to have finally gotten the memo.

Today I'm looking at county-level U.S. vaccination rates via two other factors: Income and Education.

For Income, I'm using the USDA Economic Research Service's estimate of Median Household Income from 2019 (which was, of course, the last year before COVID-19 hit the U.S. in the first place).

The lowest countywide median household income in 2019 was around $25,000/year (Clay County, Georgia); the highest was roughly $152,000/year (Loudoun County, Virginia):

One of the biggest criticisms I've received with my county-level vaccination level project is that I haven't taken into account a rather obvious truth about the partisan divide in America: Democrats tend to cluster in much more densely-populated urban areas while Republicans tend to live in more sparsely-populated rural areas.

In addition, regardless of your political lean, you might expect it to be a lot more difficult to get vaccinated if you live out in the middle of the boonies where the nearest hospital, clinic or pharmacy is 50 miles away or whatever...not to mention that if you're the only one for miles around, you might be less likely to see getting vaccinated as a high-priority task regardless of your ideology.

Therefore, the reasoning goes, instead of looking at the partisan lean of each county, it would make much more sense to see how much correlation there is based on population density or whether it's a more urban or rural region, right?

Fair enough. I decided to do just that:

I was planning on only updating the county-level vaccination graph monthly, but given the attention this has received via high-profile folks like David Frum and Paul Krugman, I've decided to post updates weekly.

With the more aggressive Delta variant now spreading quickly among the unvaccinated in the U.S., 85% of the total population seems to indeed be the more likely threshold which will be needed to achieve herd immunity.

As a reminder:

Let's be clear: There's a good ~48 million or so children under age 12 who haven't been approved to take the COVID-19 vaccine yet, roughly 14% of the population. There's also perhaps 8 - 9 million Americans age 12+ who are immunocompromised or have serious allergic reactions to vaccinations, meaning the vaccine would have no effect on them or they can't take it at all. Combined, that's around 17% of the total U.S. population who can't get vaccinated, leaving it up to the other 83% of us who can to do so.

A few weeks ago I posted a look at the county-level COVID-19 vaccination rate outliers...the counties which don't fit the main pattern of heavily-Trump counties having low vaccination rates and heavily-Biden counties having high vaccination rates.

At the time, I sliced all 3,100+ counties across all 50 states + DC into 9 regions based on a 3x3 grid:

• Vaccination rates: Under 40%, 40-60%, over 60%
• Trump 2020 vote: Under 40%, 40-60%, over 60%

I then dug deeper into two of these: Low-Trump/Low-Vaxx counties and High-Trump/High-Vaxx counties.

I found 77 counties in the first category...the population in 62 of which is more than 40% Black (and another 7 are majority Native American).

At the opposite corner, there were only 4 counties...one of which includes The Villages, the massive, MAGA retirement community. I didn't have much insight into the other 3.

I was planning on only updating the county-level vaccination graph monthly, but given the attention this has received via high-profile folks like David Frum and Paul Krugman, I've decided to post updates weekly.

With the more aggressive Delta variant now spreading quickly among the unvaccinated in the U.S., 85% of the total population seems to indeed be the more likely threshold which will be needed to achieve herd immunity.

As a reminder:

It's been a week since I last ran a scatter plot displaying current COVID-19 vaccination rates across every COUNTY nationwide according to the 2020 election results...all 3,100+ of them.

As I noted last week, with the more aggressive Delta variant now spreading quickly among the unvaccinated in the U.S., 85% of the total population seems to indeed be the more likely threshold which will be needed.

For the updated graph below, I've made a few more tweaks to both my data sources and the graph itself, neither of which makes that much of a difference in terms of impact on the final results:

Yesterday I posted a completely updated county-level look at where COVID-19 vaccinations stand based on the 2020 Presidential election results.

As I've been noting for months (and as the mainstream press has finally started picking up on), while there are several factors explaining why so many people haven't gotten vaccinated yet, the single biggest determining factor is their political lean and, even more specifically, who they voted for President last fall.

However, the impact of partisan lean varies widely from state to state. In some states the regression line is a steep partisan slope, in others the slope is barely there at all. In some states the R-squared (R^2) figure, which helps determine correlation of the partisan lean variable, is high (above 0.7) while in others it's below 0.1.