Who are the unvaccinated?
By 
This is a guest post from commenter Leeward Mountain. He’s crunched a bunch of granular stats breaking down the US population in terms of the current age distribution of completely unvaccinated people. I’ll add a couple of comments at the end
LEEWARD MOUNTAIN
The question is, what is the age breakdown of the unvaccinated? Are they all old? All young? In between?
TLDR: Roughly speaking, the majority of unvaccinated Americans are minors. Around 2/3 are under 25, and over 80% are under 40.
The 1+-dose population is more relevant than the full-series population because the group of people who have gone to the trouble of getting minimally inoculated against Covid – since the summer, this group has always been at least 10% of the overall population – are more susceptible to public health messaging than those who have refused Covid vaccination up to now. Also, while I warn I am not up to date on the literature, there is probably at least some protection offered by a partial series of vaccination against Omicron serious illness/hospitalization compared to nothing at all.
This comparative benefit must still tell at the scale of tens of millions; the partial-series population is similar in size to that of the historical medical-uninsured population in the US for reference. The partial-series population age structure is now actually fairly similar to the overall age structure, in part because of the onset of 5-11 eligibility 3 months ago, but I won’t be commenting on it further.
The official 2020 census provides reliable population estimates with detailed age breakdown. Now, you will realize that the 2020 census is almost two years out of date and may not be the dataset the CDC is working against. This is alright, because we only need to work with proportions. It is probably not the case that the age distribution of the country in April 2020, when the census was largely complete, matches exactly that of February 2022. Everyone has aged up 2 years, some people were born, some died, and a few immigrated or emigrated. In April 2020 the death toll from Covid was still negligible compared to how it stands today. Nevertheless, I’m taking the proportions between age groups between 2020 and 2022 as close enough for our purpose of rough calculation. I couldn’t find age distribution data on the CDC site with the right cutoffs anyway, so one way or another there’s a compromise.
As it happens, the Census Bureau projected the US population at the beginning of 2022 to be 332,403,650, which is somehow less than the middle estimate of 332, 601,000 in the official 2020 census!
https://www.census.gov/library/stories/2021/12/happy-new-year-2022.html [Hyperlink]
So I don’t think these 2020 figures will throw us off the mark more than trivially. If anything, the (more than usual) disproportionate deaths among older Americans over the past two years should lead us to accept an upward buffer to the ultimate findings among younger groups.
So, the operation is pretty straightforward. Get the census numbers for each age group, their fraction of the total population, then multiply by current (Feb. 11) CDC figures for vaccination rate per age group to derive an age group’s component of the unvaccinated population. 0-4 age group is treated as effectively 100% unvaccinated. I use the mid-range estimates from the census (all population figures are by thousand).
Ages: 0-4 5-11. 12-17 18-24 25-39 40-49 50-64 65-74 75+ TOTAL
# Population 19,458 28,994 26,208 30,220 68,158 41,153 64,105 32,455 21,848 332,601 % Population 5.85% 8.72% 7.88% 9.09% 20.49% 12.37% 19.27% 9.76% 6.57% 100% Unvaxed as % Pop 5.85% 5.95% 2.61% 2.2% 4.47% 1.83% 1.58% 0.49% 0.33% 25.31%
Now, we begin to encounter the peril of fudging between data sets – the cumulative unvaxed proportion of the total population is 25.3% in my calculation. Our CDC reference is 76% one-plus, or 24% unvaxed, as of mid-February 2022 (80.8% one-plus, 19.2% unvaxed, among the 5+ eligible population). World in Data shows 76.5% for first-dose-plus rates, but I’ll take the CDC figure for greater consistency between sources as well as a pessimism allowance.
While a discrepancy of more than a percentage point on a cumulative figure is annoying – a small part of the problem may be that the CDC overalls are current from Feb. 15 while the age group rates are from Feb. 11 – what helps sustain my confidence in the extrapolability of proportions is that taking the unvax rate of just the eligible 5+ population from my set gives only a 0.26-point discrepancy from the current CDC figures (19.46 – 19.2 = 0.26).
In the within-set comparison, the age groups contributing the most unvaccinated are 5-11 (23.5% of the unvaccinated) and 0-4 (23.1%), followed by 25-39 (17.6%), 12-17 (10.3%) and 18-24 (8.7%), in that order, for 83.2% of the unvaxed population. Those darned Millennials! No, it’s OK, it’s just that there are so many of them alive (the second-largest generation), as the vax rates themselves remind us.
So to summarize, assuming a modest level of stability in the proportions of vaccination in age groups and of age groups in the general population between the statistics I manipulate and whatever the ‘true’ current statistics are for both population age structure and vaccination rates, my figures of the majority of unvaccinated Americans being minors, 2/3 under 25, and over 4/5 under 40 stand reasonably well. Or even more so if there are fewer elderly people, or 40+ altogether, in proportion to the population now than there were pre-pandemic. If anyone wants to take a crack at developing a more reliable calculation, they’re very welcome.
Setting specific estimates aside, what we can be confident about in all this is that in the United States today the unvaccinated cohort is very predominantly kids and young people. How does it change our perceptions if no more than 5% of the American population are unvaxed AND middle-aged or older – the demographics who have been causing us so much grief since the last summer (admittedly it was much more than 5% at that time)?
END LEEWARD MOUNTAIN POST
A couple of comments:
Per the CDC, here are the percentages of total COVID deaths since the beginning of the pandemic by age:
85+ 25.8
75-84 25.7
65-74 22.9
50-64 18.8
40-49 4.5
30-39 1.8
18-29 0.6
0-17 0.09
A surprising stat (to me) is that this distribution remains practically unchanged if we consider only COVID deaths last month:
85+: 25.6
75-84: 26.5
65-74: 22.9
50-64: 18.8
40-49: 4.0
30-39: 1.5
18-29: 0.6
0-17: 0.1
I would have thought that going from a completely unvaccinated population in which the overwhelming majority of COVID deaths were among old people, to one in which the overwhelming majority of old people are vaccinated, would have skewed the the age distribution of COVID deaths strongly downwards, but this isn’t happening currently. (The age distribution of COVID deaths did skew downwards somewhat in 2021 relative to 2020, but apparently this effect has disappeared during the Omicron wave).
The bottom line stat in Leeward Mountain’s breakdown that jumps out to me is that, currently, only 2.4% of the total US population consists of people who are 50 and older and completely unvaccinated. Yet this cohort is now generating around 93% of the roughly 100,000 COVID deaths we’ve seen in the last six weeks alone. (94% of all current COVID deaths are among people 50 or older, and vaccinated people in this cohort reduce their mortality risk from COVID by well over 90% relative to the unvaccinated).