Thread by @dhmontgomery

1/ This can be somewhat counterintuitive, so it's not surprising many people are confused.

First, numbers: For the period where we've got data (May-Nov.) MN has 1,868 COVID deaths, of which 718 were fully vaccinated & 1,150 not; most of these deaths were in the past few months.

https://twitter.com/nu_Antz/status/1470453514182791172

2/ Over this period, the share of MN's population that's fully vaccinated has grown from about 40% up to about 60% (with, again, more of the deaths happening in these later period where vax rates were higher).

3/ But remember 96.5% of MN's #COVID19 deaths have been over 50, and nearly 85% were 65+. So when estimating deaths, we REALLY need to adjust for age.

4/ When this period for which we have data began, more than 80% of MN seniors 65+ were fully vaccinated; that’s now over 90%.

Among 50-64-year-olds, fully vaxxed rates rose from about 50% to 75% in this same period.

This chart is by 1+ dose, not fully vaxed, but same picture:

5/ So in this time period among the age groups that account for almost all of Minnesota’s #COVID19 deaths, the vaccinated population VASTLY outnumbers the unvaccinated population. And yet despite that the raw count of deaths among the unvaxxed significantly exceeds the vaxxed.

6/ An example*: for the week of Nov. 7, there were approx. 170 COVID deaths in MN among those 65+. ~105 of those were unvaxxed, ~65 vaxxed. In other words, 61% unvaxxed.

But those ~105 unvaxxed deaths were out of ~60K unvaxxed seniors. The ~65 were out of 800K vaxxed seniors.

7/ In other words, in this week we had about 177 deaths per 100,000 unvaccinated seniors, and about 8 deaths per 100,000 vaccinated seniors.

THAT’S the difference, more than 20-fold.

https://twitter.com/dhmontgomery/status/1470469133691408385

8/ If you just try to divide cumulative totals by vaxxed/unvaxxed, you’ll miss the huge impact that age has on susceptibility to #COVID19 (as well as big differences in vax rates among different age groups).

9/ If you break it out by age group, the results are somewhat messy, but generally consistent, with the only really big gap by case prevalence (where breakthrough cases are comparatively rarer among children & the elderly than among working-age adults). Deaths & hosp v. similar.

10/ The * footnote to #6 here: These numbers are approximate. @mnhealth releases nominal overall figures for breakthrough cases, but only ratios for ages, so I had to reverse-engineer it. My numbers don’t quite match to our totals from another dataset, but are close-ish.

@mnhealth 11/ Another thing this analysis overlooks, of course, is partially vaccinate people (or, more recently, the role of boosters). We just have data by “fully vaccinated” and “not fully vaccinated,” when of course the real situation is more nuanced.