effective covid mitigations: all cause death analysis

"more burpee, less fauci:" a real public health policy for america

longstanding gatopal ben M @USMortality who has been tracking overall US mortality rates (all cause deaths etc) was kind enough to let me have a bit of a rummage around in some of his data and boy was there some interesting stuff.

first a few words of definition:

excess mortality is defined as all cause deaths in excess of predicted all cause deaths. actually making that prediction is always a bit of an exercise in modeling, but it’s also one with which humans have a great deal of experience. insurance actuaries have been doing it for centuries and they have models with strong enough predictive ability that you can price life insurance to the nickel with them. so this is real actuarial science, not neil ferguson “throw spaghetti code against the wall and hope it sticks” stuff.

baseline for these calculations used 2015-19 deaths by state and adjusted it for population growth and changes in age stratification. methodological discussion can be seen in in detail HERE.

this can get us a nice plot of excess deaths by state. this still has a bit of bias in it as some states have had 2021 peak season and some (notably california) have not, but it’s a useful set nonetheless and unlike so much in covid, it’s very high quality data. (for this reason, south carolina and alaska are omitted. they seem to have data issues here and materially incomplete reporting)

now comes the fun part: what can we get this to correlate to?

everyone always wants to know about covid restrictions, so let’s start there. i used this data and compared it to the wallethub freedom index because it’s pretty accurate and provides a nice, simple number that includes mitigations like masking and school policy that are absent from sets like google mobility data. is it perfect? no. is it good enough that if there’s a strong signal here we should see it? yes, i think so.

but there is no signal.

what you did to try to stop covid appears to have had no effect on all cause deaths. this is a result i pulled out dozens of times using google mobility data vs covid deaths as well back when internet felines were still permitted over at casa del bluebird. it looks to hold even now.

none of this policy worked.

you can have correlation without causality, but material causality without correlation is extremely implausible to the point of more or less requiring special pleading.

one could argue the issue here is the restrictions dataset, but it looks pretty sound at least in terms of relative state positions to me and lines up with other datasets i’ve used. (that said, if someone has another idea on a measure of such restrictions i can run this against that has some sort of temporal continuity, i’d be interested to hear about it.)

but this is not really that interesting a result. we knew this. it was known for generations and “lockdowns, restrictions, and masking don’t work” was mainstream epidemiological canon pre 2020. (great survey HERE)

but health policy is about more than reaction, it ought to be about prevention and the best prevention is being healthy before a pandemic starts. that’s why this data is so interesting:

obesity in a state is quite a strong predictor of all cause death outcomes. (R2 of 0.21 is a quite a high number in a system this complex and seasonal. 21% of overall outcome predicted is A LOT. this is not a faint signal. this is an air horn in a library. alarms should be going off.)

this signal makes a great deal of sense as obesity is a significant risk factor for covid and maps to many other risk factors as well such as diabetes and heart disease.

so, what we see here is that lock down and mask up is vastly outperformed by having eaten well and worked out. being lean and fit is a strong mitigator of covid and lacking these comorbidities is a 93% death risk reducer overall and a 99% reducer among the oldest. (lots of data HERE)

this keeps bringing me back to “so did closing the gyms and packing massive excess weight on americans by locking them up and terrifying them into stress eating really represent a good health policy choice?”

because it sure looks like it didn’t. (lots of data HERE in the second half of the sweden piece)

guys it’s really, really simple:

public health is about keeping the public healthy.

period.

the best way to do that is prevention, not pharma treatment of the already ill.

and the best prevention is a healthy body with a strong immune system.

we’re fat and it’s making us sick.

it’s the greatest health risk and crisis of our time, and nothing else is even close.

and the solution is well known and has nearly 100% positive side effects apart from an occasional and persistent loss of friends by never shutting up about crossfit.

if you will not change your life and take care of yourself, what possible moral standing do you have to demand that others upend their lives to take care of you?

“my health is important enough to oppress you over but not to put in any effort myself” is not much of a mantra, is it?

“more burpee, less fauci” is the way forward to a holistic health policy for america.

let’s get on it.