Excess Mortality Data for the US and Texas, Open Source

6,659 Views | 36 Replies | Last: 4 yr ago by PJYoung
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
12:33p, 7/22/20
AG
I wanted to do my own independent research into the excess mortality data available for the Texas and US, since I felt that the analysis publicly available was not comprehensive enough and how it was analyzed often told something about a person's biases.

Data source: https://gis.cdc.gov/grasp/fluview/mortality.html

Link to my spreadsheet data where you can see my data and analysis: https://drive.google.com/file/d/14NB6hA5cgChcVccShqsoFxxGkaTaU83z/view?usp=sharing

One thing to note, I did all of the analysis two ways -- one with raw numbers, one population adjusted. The population adjusted number seems more appropriate, since the population of the US and Texas has been going up over time. I tried to cut the data a few different ways, so that you can see how the different assumptions change the calculations. I think that focusing on March-June is more appropriate, as that is when CV was present in the US. This year had a more average flu season in terms of deaths so when you include that date range in the numbers, the excess mortality looks more favorable but is still significant.

One big caveat is that the all-cause data in the US is only complete until the end of June, so it would have not captured any potential spike in deaths in July. I can update this in about a month to share.

I think the conclusion here is that all-cause mortality is up. At the end of June, Worldometer had 130,050 deaths attributed to coronavirus for the US and 2,496 deaths in Texas. Both appear to be in the right ball-park, indicating it is not likely there is a gross overestimation of official coronavirus death counts.


UNITED STATES:




TEXAS:


Reply Quote
15
PJYoung
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
12:37p, 7/22/20
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
12:45p, 7/22/20
AG
I realized there was one more way to cut this -- the first 39 weeks in the season. From October to June.

UNITED STATES


TEXAS
Reply Quote
2
Complete Idiot
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:11p, 7/22/20
Very nice, confirms what my takeaway had been from looking at CDC data in the past. I certainly didn't analyze as thoroughly as you did but by looking at excess deaths it certainly appeared the published Covid death totals were in the same order of magnitude +/- 10K. I hadn't looked at the Texas numbers at all though, my thoughts in looking at your data is that Texas is under reporting, if anything.
Reply Quote
Complete Idiot
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:13p, 7/22/20
Certainly doing this from March makes way more sense than the November start date shared in a tweet earlier today, unless someone is claiming it's been in the US in statistically significant numbers since November and they are able to distinguish from influenza deaths.
Reply Quote
fig96
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:20p, 7/22/20
AG
Super interesting analysis, thanks for the post. Mind if I share this?
Reply Quote
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to fig96 • 1:21p, 7/22/20
AG
fig96 said:
Super interesting analysis, thanks for the post. Mind if I share this?
Be my guest
Reply Quote
1
Complete Idiot
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:26p, 7/22/20
I've just looked at https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm throughout the past few months, but they show it as percent of expected deaths.
Reply Quote
Duncan Idaho
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:27p, 7/22/20
Am I right in dumbing this down as:
The "they died with Corona and they didn't die from Corona" argument is wishful thinking that isn't supported by the data"?
Reply Quote
8
1 edit
fig96
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Duncan Idaho • 1:33p, 7/22/20
AG
Duncan Idaho said:
Am I right in dumbing this down as:
The "they died with Corona and they didn't die from Corona" argument is wishful thinking that isn't supported by the data"?
I'm taking it as "there are occasional anomalies, but overall the COVID death count is likely pretty accurate".
Reply Quote
12
Beat40
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:36p, 7/22/20
Hotard, for my education, what is the point of this analysis and the resulting use of it?

Are we trying to prove if COVID death reporting is accurate?
Reply Quote
1 edit
Honky Tonk Hero
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:41p, 7/22/20
AG
Interesting that for full US data set - the curve has now rejoined previous years - no longer showing 'excess deaths' across US population. Of course this could differ by location.
Reply Quote
1
Keegan99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Duncan Idaho • 1:53p, 7/22/20
AG
Duncan Idaho said:
Am I right in dumbing this down as:
The "they died with Corona and they didn't die from Corona" argument is wishful thinking that isn't supported by the data"?


Part of this will be answered in the fall and next flu season.

That is, we know COVID time-shifted deaths, since a person can only die once, but how was that shift distributed?

If the bulk was in nursing homes, where residents have minimal remaining lifespans, then all-cause mortality should drop below usual in the coming months.
Reply Quote
18
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Beat40 • 1:53p, 7/22/20
AG
Beat40 said:
Hotard, for my education, what is the point of this analysis and the resulting use of it?

Are we trying to prove if COVID death reporting is accurate?
Basically.

There is a lot of concern about the validity of the deaths data for coronavirus, whether it's overestimating deaths due to people being coded for dying with coronavirus for some other spurious cause (motorcycle accident) or people not being coded as dying with coronavirus because they died of a stroke in their home and were never hospitalized or tested. There are also people concerned that those dying are people who would have died shortly anyways, for example the person with stage 4 cancer who gets it.

The all-cause death is a reasonable way to sanity check, although of course even it has it's flaws. It will be also mixing factors for such things as lower auto deaths, higher suicide rates, etc.

Some people have used the all-cause morality to show how big the problem is. Those with that narrative use non-population adjusted numbers and take the most aggressive time frame.

People who wish to use all-cause morality to downplay the extend of the problem will use a larger time frame and use population adjustments to reduce the extent.

I just wanted to cut the data all possible ways to help people draw their own conclusions.
Reply Quote
9
1 edit
OldArmy71
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
2:05p, 7/22/20
AG
I have no idea what I am supposed to be seeing in these charts.

What do they mean?
Reply Quote
1
tysker
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 2:05p, 7/22/20
AG
Keegan99 said:
Duncan Idaho said:
Am I right in dumbing this down as:
The "they died with Corona and they didn't die from Corona" argument is wishful thinking that isn't supported by the data"?


Part of this will be answered in the fall and next flu season.

That is, we know COVID time-shifted deaths, since a person can only die once, but how was that shift distributed?

If the bulk was in nursing homes, where residents have minimal remaining lifespans, then all-cause mortality should drop below usual in the coming months.
Don't forget the shutdown 'saved' lives of those under 18 mostly due to fewer accidental deaths. One would expect those numbers to mean revert but maybe not if schooling is done primarily at home for an extended period.
Reply Quote
SouthTex99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
2:20p, 7/22/20
AG
Reply Quote
1 edit
RandyAg98
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to HotardAg07 • 2:37p, 7/22/20
AG
Good work!
Reply Quote
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to SouthTex99 • 3:04p, 7/22/20
AG
SouthTex99 said:
Excellent report thank you for sharing this. Cuts through all the agendas.

Mortality is no doubt elevated. It's not "off the scale" elevated but there's no denying that there's something going on to cause it. So has the government response mitigated the rate of mortality in the United States? Or would this have been the case no matter what?
Hard to answer. I would like to think an advanced civilization with infinite data available could run an advanced regression model to determine what were the influencing factors on the virus spread and deaths and it would look something like:

10% of the variation explained my government policy
20% of the variation explained by human adaptation/behavior -- mask wearing, compliance to distancing, travel/mixing, super spreader events, etc
20% of the variation explained by geographical and location factors -- climate, temperature, humidity, population density, reliance on public transportation
10% of the variation explained by societal factors -- multi family homes vs single persons, reliance on "front line jobs", access to healthcare, and the general health of the populace -- obesity, heart issues, diabetes, etc.
10% of the variation explained by the healthcare efficacy in detection, tracing, treatment, etc.

Rest - luck and randomness

Reply Quote
2
2 edits
RoyVal
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
3:05p, 7/22/20
AG
this is great info and very nicely and visually put together. thank you for sharing!
Reply Quote
1
HumpitPuryear
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
3:13p, 7/22/20
AG
There's a seasonal spike around week 15. I'm guessing that's peak flu season. Looking at 2017-18 curve for your Texas adjusted by population it's very dramatic. Did you dig into that at all? Any explanation? Was that a particularly bad flu season or was something else going on?
Reply Quote
HotardAg07
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to HumpitPuryear • 3:17p, 7/22/20
AG
HumpitPuryear said:
There's a seasonal spike around week 15. I'm guessing that's peak flu season. Looking at 2017-18 curve for your Texas adjusted by population it's very dramatic. Did you dig into that at all? Any explanation? Was that a particularly bad flu season or was something else going on?
It appears as if it was a particularly bad flu season. If you look at the source website, they show the % of deaths attributable to flu/pnuemonia and there was a corresponding spike in 17-18.

I believe that it was the flu season commonly referenced by people where 80,000 people died. They have since revised down the number to 69,000 for that season.

https://gis.cdc.gov/grasp/fluview/mortality.html
Reply Quote
1
HumpitPuryear
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
3:46p, 7/22/20
AG
Yes we lost roughly 5000 more Texans in a 10 week period in 2018 than we lost in the 10 worst weeks of COVID. That was not true for the US overall however. It's easy to see why a lot of people in Texas might say COVID is no worse than the flu while someone in NY would have a different opinion.

Look out if COVID has a seasonal peak like the flu does. If it's seasonal peak aligns with regular flu, February of 2021 is going to be a *****.
Reply Quote
1
Ciboag96
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
4:25p, 7/22/20
Take out New York from your US chart and see what happens.
Reply Quote
2
fig96
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Ciboag96 • 4:48p, 7/22/20
AG
We would have the same range less the number of deaths from New York?
Reply Quote
buffalo chip
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Ciboag96 • 9:45p, 7/22/20
S
My first reaction to that... The Ags would have beaten Alabama if you ignore the three PIC 6 throws by Kyle Allen...

After thinking about it, not a bad idea since NY performance was so much worse than every other state (so far).
Unless you listen to Dr Fauci...
Reply Quote
1
PJYoung
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to HumpitPuryear • 9:33a, 7/23/20
AG
HumpitPuryear said:
Yes we lost roughly 5000 more Texans in a 10 week period in 2018 than we lost in the 10 worst weeks of COVID. That was not true for the US overall however. It's easy to see why a lot of people in Texas might say COVID is no worse than the flu while someone in NY would have a different opinion.

Look out if COVID has a seasonal peak like the flu does. If it's seasonal peak aligns with regular flu, February of 2021 is going to be a *****.

I agree with you but I would also say we appeared to have just started our worst 10 week period in Texas a couple of weeks ago.

Reply Quote
3 edits
HumpitPuryear
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to PJYoung • 1:32p, 7/23/20
AG
Yeah I'm not implying this is over. In fact I said that if COVID follows a seasonal cycle like the flu than we could be headed for a rough start in 2021. Let's hope it mutates into something less serious. I'm just pointing out that, at least to Texans that are familiar with sharp increases in very deadly seasonal flu, that COVID so far has not surpassed our worst flu epidemics in terms of total deaths or stress to medical resources.
Reply Quote
Aggie95
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
1:55p, 7/23/20
AG
has anyone seen what the excess death count is for particular age group? Say those <50 and those >50?
Reply Quote
2
AustinAg2K
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
2:04p, 7/23/20
Great information. At this point, I don't think there's any reason to believe the numbers are fabricated. There is certainly anecdotal evidence of one or two cases where they died with Covid, as opposed to from Covid, but in general, Covid is the major cause of the deaths.

Logically, it makes sense, too. People keep saying that hospitals are increasing their numbers because they get more money from the CDC. However, the extra money pales in comparison to the amount of money they are losing from not having elective surgeries. Also, there is nothing we've seen that is particularly different than what has happened across the rest of the world. To believe we are fabricating or numbers is to believe the entire world is, too.

Now, if you want to make an argument that the number of infections/deaths does not justify the reaction, I think that's a fair argument to have, but I see no reason to not believe the published numbers.
Reply Quote
6
P.U.T.U
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Aggie95 • 2:24p, 7/23/20
AG
If someone could break it out by age that would be very telling.
Reply Quote
2
MaxPower
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
5:18p, 7/23/20
Interesting there's always an uptick in deaths in December. Any reason 2013 was so bad?
Reply Quote
GE
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Duncan Idaho • 5:32p, 7/23/20
AG
Duncan Idaho said:
Am I right in dumbing this down as:
The "they died with Corona and they didn't die from Corona" argument is wishful thinking that isn't supported by the data"?
Will know the answer to that if and when there is a Rona cure and vaccine. If true then the next few future years will have less deaths than average because Rona would have brought some forward from future years.
Reply Quote
1 edit
Dr. Not Yet Dr. Ag
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to HotardAg07 • 11:33p, 7/23/20
HotardAg07 said:
HumpitPuryear said:
There's a seasonal spike around week 15. I'm guessing that's peak flu season. Looking at 2017-18 curve for your Texas adjusted by population it's very dramatic. Did you dig into that at all? Any explanation? Was that a particularly bad flu season or was something else going on?
It appears as if it was a particularly bad flu season. If you look at the source website, they show the % of deaths attributable to flu/pnuemonia and there was a corresponding spike in 17-18.

I believe that it was the flu season commonly referenced by people where 80,000 people died. They have since revised down the number to 69,000 for that season.

https://gis.cdc.gov/grasp/fluview/mortality.html

2017-18 Flu season was awful. Easily the worst season I've experienced by far as a physician. I always found it a bit disingenuous when people compared COVID to the 17-18 flu season (using it as a measure of the typical flu season). That was the worst flu season the US has experienced in several decades, and yet from my experience as well as the numbers, COVID has been significantly worse.
No material on this site is intended to be a substitute for professional medical advice, diagnosis or treatment. See full Medical Disclaimer.
Reply Quote
8
1 edit
HumpitPuryear
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Dr. Not Yet Dr. Ag • 9:42a, 7/24/20
AG
The numbers don't support that. 2014, 2015, and 2018 peak mortality rate in Texas exceeded the peak mortality rate in 2020. I'm not disputing that COVID or those peak flu seasons are/were not serious health events. The difference is that we did not put millions of people out of work and spend billions in federal benefits to flatten the flu curve in 2018. Plus if COVID ends up being seasonal and hits hard in early 2021 on top of peak flu season we are going to wish we had more immunity in the population.
Reply Quote
3
Page 1 of 2 ▼
 
×
subscribe Verify your student status
See Subscription Benefits
Trial only available to users who have never subscribed or participated in a previous trial.