1/

@NickHudsonCT inspired me to do something similar for the US.

We start w/ 10+ yrs of All-Cause.

Regression captures baseline drift, then fit a seasonal model.

We note that ~1/3 of the excess is outside the SD bars. 👍

From this, we can estimate weekly excess mortality.

https://twitter.com/NickHudsonCT/status/1361947880538468353

2/
From the seasonal model, we can estimate yearly mortality and yearly excess over the past decade.

We see the largest All-Cause excess in CY 2017 ~40K which is less just Alz excess this year.

All-Cause Excess includes excess mortality from Alz, MI, overdoses, suicides etc.

Understanding the Danish Mask Study:

1/
Whenever possible, you should always run your trial in software to see if the study design is sound, and capable of delivering the results. A power analysis is a start, but a simulation is better. 2/ @Lester_Domes
The study uses the classic Frequentist analysis with p-values. We will leave that conversation for another time.

Here we start with the actual study data and run the study a million times and observe the range of raw counts as well as the range of percentages.