Thread by @DearPandemic

Q: What is the real death toll due to COVID-19? Are we OVER or UNDER counting deaths? Wouldn’t a lot of these people have died anyway?

A: The death toll has been HIGH, we are likely UNDER counting deaths, & some but NOT THIS MANY.

Let’s talk EXCESS MORTALITY. #epitwitter 1/

What is excess mortality? Very simply, it is the number of deaths above and beyond those that would have been expected based on previous years. 2/

People often have multiple contributing factors to death. There is room for error here and potential for both under and over counting of COVID-19 deaths.

A good overview of the challenges in measuring COVID-19 specific deaths from @LiveScience:
bit.ly/3jdbXQT
3/

Due to these challenges, researchers have awaited ALL-CAUSE mortality data, considered more reliable. In developed countries, overall death registration is quite good...governments don’t want to pay social security and pensions any longer than they have to :) 4/

From all registered deaths, we can calculate “EXCESS MORTALITY.” The simplest way to calculate this is to count how many more deaths there have been compared to an average of the previous few years over the same time frame. #poptwitter #demography 5/

The benefit of “excess mortality” is that it likely picks up “direct” COVID-19 related deaths that were not registered as such, especially early on when testing rates were low.... 6/

...as well as “indirect” deaths due to overwhelmed hospitals or people being deterred from seeking out needed health care. Deaths due to some causes such as car accidents may have declined during lockdown, counteracting some of the positive “excess” deaths. @AndrewNoymer 7/

Several websites have been providing great data visualizations of excess mortality during the pandemic, including @jburnmurdoch @FT for many countries and localities:

on.ft.com/2WurZvX

and @nytimes for U.S. States:

nyti.ms/395tYfm

8/

By the FT’s calculations:

The US has seen almost *150,000* excess deaths during the pandemic period, *23%* higher than typical mortality.

The UK has seen around *65,000* excess deaths, *45%* higher than typical mortality.

@FT @jburnmurdoch

9/

The country-level numbers mask even higher excess mortality in hotspots for example:

New York City saw over *27,000* excess deaths, *208%* higher than typical mortality

Madrid saw over *16,000* excess deaths, or *157%* higher than typical mortality.

10/

For those worried about over counting of COVID-19 deaths or counting people who got COVID-19 but were going to die soon anyway:

EXCESS mortality only counts deaths ABOVE AND BEYOND WHAT IS EXPECTED, giving us the most accurate picture of the mortality burden due to COVID-19.

Because there are changes over time in the age composition of populations and other underlying trends in mortality, demographers prefer to model the baseline of “expected” mortality based on these trends rather than just take the average of previous years. #poptwitter /12

This has led to slightly fine-tuned but qualitatively similar estimates for excess mortality in the U.S. and UK. Some links to these model-based estimates:

US: bit.ly/396l4hi
UK: bit.ly/30koOrx

#poptwitter #epitwitter

/13

In the US and the UK, excess mortality estimates are roughly 30% higher than counts due to official COVID-19 deaths, suggesting we are UNDER rather than over counting COVID-19 related deaths.

/14

So, it’s important to remember that while SOME people who died of COVID-19 would have died anyway over this period of time, this is NOT THE NORM and does not explain the huge increases in overall mortality we have seen in the hardest hit COVID-19 cities and countries.

/15

A caveat--while excess mortality is a very useful metric for countries where death registration overall is good, there are many countries where it is not & the true toll of COVID-19 mortality will be hard to measure:

https://twitter.com/helleringer143/status/1261867969992425474

/16

Finally for more details check out this fantastic excess mortality explainer from @MaxCRoser & @OurWorldInData

ourworldindata.org/excess-mortali…

Online dashboards for calculating excess mortality in different locales:
cdc.gov/nchs/nvss/vsrr…

mpidr.shinyapps.io/stmortality/ @HMDatabase

/17