Mortality with Meep: Total U.S. Excess Mortality in 2020, by Race and Ethnicity
Some stark patterns in racial/ethnic disparity
The age slices of U.S. mortality are important to look at, given age is a huge driving factor.
Let’s move on to racial/ethnic groups, which provide a new perspective. In this post, I’m showing results at a high level, and what I’ll do in the next post will be telegraphed in the video where I look at the CDC dashboard by race/ethnicity.
Total Mortality by Race/Ethnicity in 2020
Since the last time I graphed total mortality in 2020, there’s been a data update… and now I’ve got the data split out by race/ethnicity, not age.
Seeing the breakout, unlike with age, doesn’t really tell you a lot.
First, non-Hispanic White folks are not only the majority of the U.S. population (~60%), they’re also the oldest. Most of the people dying are over 80 years old, and most of those people are non-Hispanic White. That’s true even in “normal” years.
Second, while there are more Hispanic people (~19%) in the U.S. than Black (~13%), Hispanic mortality is actually lower than both Black and non-Hispanic White mortality rates. That’s even without digging into age distribution differences.
But it still doesn’t give us an idea how much in the way of excess deaths we’re seeing in 2020.
Excess Mortality by Race/Ethnicity in 2020
So let’s subtract off the average number of deaths for 2015-2019 for each group and see what results.
Hmmm, interesting. There are some artifacts of data issues I can see at the end of the year — yes, we have the deaths drop off in December, due to the lag in reporting.
But one item I see is that Black excess mortality drops off somewhere in October… you know, when North Carolina has stopped reporting its deaths to the CDC.
[death glare at North Carolina]
Let’s see if we can drill down better.
Percentage Excess Mortality by Race/Ethnicity
Let’s focus on the three largest racial/ethnic groups: Non-Hispanic White (~60% of population), Hispanic (~19%), and Non-Hispanic Black (~13%).
The other groups together make up less than 10% of the U.S. population, and really don’t have a lot of deaths in prior years. When you’ve got a small population, you get far more volatility in the mortality stats just to begin with.
Focusing on these three groups, here are the relative results — excess mortality as a percentage of the average number of deaths by week for 2015-2019.
Wow, huh?
This definitely shows the Black and Hispanic populations in the U.S. definitely have been harder-hit, percentage-wise, than the overall population.
Except.
Given what I know about Simpson’s paradox (that’s some paradox, the Simpson’s paradox (it’s my favorite)), I know that the overall excess percentages are almost definitely skewed because of the geographic hit COVID made — especially that first spike where both Black and Hispanic excess mortality went over 100%.
The first spike had the biggest hit in NYC, which has a unique population distribution compared to the rest of the U.S., for instance.
I will get to that eventually.
Video: CDC dashboard on racial/ethnic death patterns in 2020
I also look at the top 10 states (including NYC with NY) in population.
So you can see already the differences by state, though the CDC display isn’t showing the graphs the way I do.
Resources
Data source: Weekly counts of deaths by jurisdiction and race and Hispanic origin, accessed 28 January 2021.