Conversation

@MLevitt_NP2013

This is a very important issue. It means that excess death is not well-defined. I am puzzled that USA seems to have half the flu deaths per million as Europe. US has about twice the COVID-19 deaths as EU. Is this due to baselines including a lot of normal annual flu in US?

Quote Tweet

Ray J Wallin

@Ray_J__

Sep 18, 2021

Replying to @PienaarJm and @MLevitt_NP2013

I've posted this before, but yearly averages are thrown off by winter flu etc outbreaks. Year-to-year variations become more evident when summer-only data is used for forecasting. And then, there is the chaotic behavior of fish in a pond... youtube.com/watch?v=m7TXb7

4:06 AM · Sep 19, 2021

Replying to

The difference in deaths between EU and USA may be due to Hope Simpson, not flu. USA Covid deaths split above and below the 37th parallel show that in 2021 summer, we peaked at ~3x deaths in S states vs N. in 2020 summer, deaths in S states peaked at ~5x the N states. 1/3

Replying to

and

I live in North Carolina (southern-ish), and during many summers, a lot of people get colds, but not flu. This leads me to wonder if seasonality causes a coronavirus surge in summer.

Show replies

@IamAllBannedSEXfeminists Honk

Replying to

We have this excellent graph for Sweden, we need same for other countries.

Replying to

Not double the COVID deaths per capita. US is at 2002/million while EU is 1710/million. It was close to even a few months ago until the Delta surge, but EU lags behind the US so they'll probably catch up later.

ourworldindata.org

COVID-19 Data Explorer

Research and data to make progress against the world’s largest problems

Replying to

Jig is up

Replying to

I might have misunderstood the last para but don’t understand why including normal flu deaths in baseline would increase Covid deaths? Excess deaths are the deaths in excess of the expected baseline for that week which in SA includes winter peaks due to flu.

Replying to

If you only take yearly (not monthly) excess death statistics, you could use a rolling window of width=2 years to even out bumps as well. Or with monthly data one can just not artificially cut at year boundaries and just take the data as is, as a timeseries.

Replying to

'Excess death' is not well defined. I have suggested determining excess death results from 2 or more methods, from opposed viewpoints, then audit each. Methods and time periods can be chosen to benefit one's opinion/stance, but I believe Levitt's method is fair.

Replying to

Don’t hospitals get paid more if they are covid deaths?

Show more replies