U.S. Mortality Trends Through the Pandemic
COVID and non-COVID trends by age group
I gave a talk at the Actuaries Club of Hartford and Springfield on May 10, using data I had grabbed from CDC Wonder on May 6.
Here is a non-actuarial version of the talk… and yes, it runs almost an hour:
This is a look at mortality trends, primarily by age groups and major causes of death, both COVID and non-COVID causes, for 2020 and 2021.
I do explain what actuaries are generally thinking about as we look at the general population mortality trends, and why we care by cause of death, even when we are going to pay out the life insurance death benefits even when it doesn’t matter what you die from.
In short: we are trying to figure out which way mortality trends are going to go. Different causes of death have trended differently, and we are interested in teasing out the impact of COVID specifically from other causes of death that may continue having bad trends post-pandemic (even if there is a post-pandemic period — after all, COVID becoming endemic like the flu is a possibility.)
Time stamps
I put time stamps in the video, which makes YouTube break it up into chapters.
The big topics are as follows:
01:45 Insurance and mortality trends
05:59 All-cause mortality
11:07 Deaths by age group
23:53 Cause of death – major categories
26:42 COVID trends
32:35 Non-COVID causes of death
37:16 Physiological causes of death
42:50 External causes of death
There are a bunch more topics than that, but those are the major highlights.
Big result: worst excess mortality for ages 30-44
I discuss this in the portion of the talk about death rates by age group.
For 2021, the worst relative increase in mortality, compared to 2019, was for ages 30-44.
[I have called it the Millennial Massacre, but it obviously overlaps with Gen X…. and Middle Age Massacre doesn’t exactly work, either. Dang the allure of alliteration].
We will see in a moment that most of that mortality increase didn’t come from COVID.
If you look at overall mortality, obviously total mortality for this age group is much lower than for those much older.
A 5% increase in mortality for those aged 85+ will translate to a much larger number of deaths, but a 50% increase in mortality for those aged 40-44 is extremely worrisome to actuaries and insurers even if the absolute number of deaths is lower in impact. We’re setting reserves and expectations based on certain assumptions, and we’re generally not assuming fluctuations of 50% — that’s just nuts compared to our historical experience…..
…..until now.
The big question: how much of the excess mortality was COVID
A second key graph:
I discuss this graph at time stamp 30:55 Percentage of excess deaths explained by COVID.
For those age 85+, more than 100% of the excess mortality is explained by COVID for 2021 – there really is nothing to explain beyond COVID. Small kids really didn’t have excess mortality, so there’s nothing to explain there.
For older middle-aged adults, age 45-84, the majority of the excess is COVID. There still is a non-COVID component to explain, but it’s not the majority.
However, the part we really are worried about, age 10-44, and that age 30-44 in particular, less than half of that excess mortality was COVID. We need to dig into that if we’re to get a handle on mortality projections as actuaries.
I have talked about some of these major contributors to excess mortality for these age groups, and a lot of it was external causes. I am just talking at a high level in this video.
Links and more
A look at mortality trends, primarily by age groups and major causes of death, both COVID and non-COVID causes, for 2020 and 2021.
Dropbox link to slides and spreadsheets:
https://www.dropbox.com/sh/394386q974qe7hv/AADXaLYH5uekikjmpA-Hv8LMa?dl=0
Mary Pat Campbell links
Substack
Mortality with Meep
https://marypatcampbell.substack.com/s/mortality-with-meep
LinkedIn
https://www.linkedin.com/in/marypatcampbell/
Data sources:
Finalized mortality data: https://wonder.cdc.gov/ucd-icd10.html
Provisional mortality data: https://wonder.cdc.gov/mcd-icd10-provisional.html
Full time stamp list:
00:00 Intro
01:30 Topics
01:45 Insurance and mortality trends
05:59 All-cause mortality
08:17 Deaths by month
11:07 Deaths by age group
16:11 Death rate by age
18:57 Logarithmic scale
20:24 Percentage change from 2019
23:53 Cause of death – major categories
25:54 External deaths increased in 2020
26:42 COVID trends
28:15 Quarterly deaths
29:04 COVID deaths by age group
30:55 Percentage of excess deaths explained by COVID
32:35 Non-COVID causes of death
37:16 Physiological causes of death
37:21 Heart disease
38:10 Cancer
39:45 Flu/pneumonia
40:47 Diabetes
41:42 Liver disease
42:50 External causes of death
42:53 Homicide
44:58 Suicide
45:51 Accidents
46:13 Falls
47:53 Motor vehicle accidents
49:17 Drug overdoses
53:20 Links and outro
Haven't watched the YouTube video yet, but quick questions/feedback/notes:
1) In the XLSX file "Underlying Cause of Death 2021 Provisional Stats and Ranking 2022 May 7" I am trying to figure out what category/age group is missing from the various pivot tables which causes the total deaths to be artificially low.
Examples: Sheet "COD Pivot Table", filtering to all 10 year age groups, sum of deaths is 2,891,658 when it should be 3,461,321 according to sheet "Age Groups", row 53, which sums total deaths for 2021 (3.46 million also ties out to other various CDC reports). What explains the 569K gap? I scanned ICD list and not seeing anything missing, and appears all age groups included.
I see similar gap in the sheet "Graph Prep COD" - each year appears to be missing roughly 500-600K deaths. What were you filtering out?
2) Have you run similar analysis for other countries? Granted my skills are at "Grade school Tee Ball" to your "New York Yankees", but lots of interesting things when digging into all-cause morality and excess deaths using mortality.org data and other countries vital statistics. Sweden not having any excess deaths ages 0-65 throughout 2020 and 2021 is interesting, South Korea reporting unusually low death rates even before the pandemic strikes me as odd considering their aging population (death rates roughly equal to Colorado when I would have expected them to mirror closer to Arizona), etc.
3) I'm embarrassed I never knew of the "Sparkline" feature in Excel (As seen in "Total Mortality Level by State 1999-2021 provisional 6 May 2022", Sheet "Crude Death Rates", Column Y. ) I wonder how new that is... I used to tediously make tiny pictures of my charts and paste into cells a decade+ ago to achieve the same thing. If I find out that was there all along in the late aughts I'm gonna pull out what's left of my hair :)