PonyFans.comBoard IndexAround the HilltopFootballRecruitingBasketballOther Sports

Football this fall? No way

General discussion: anything you want to talk about!

Moderators: PonyPride, SmooPower

Re: Football this fall? No way

Postby JasonB » Wed Jul 29, 2020 9:25 am

Anyone catch the deaths added to the totals this week?
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

Re: Football this fall? No way

Postby mustangxc » Wed Jul 29, 2020 10:02 am

I heard that the Texas healthcare system reformulated its classification system and as a result there are an additional 600 deaths attributed to COVID-19. Not sure what changed in the classification system so I don't know what to think.
User avatar
mustangxc
PonyFans.com Super Legend
 
Posts: 7281
Joined: Tue Nov 14, 2006 3:57 pm

Re: Football this fall? No way

Postby ponyboy » Wed Jul 29, 2020 11:43 am

Yep, I see it, a net of about 600 new attributions on 7/27. It knocked the Texas IFR back to the level it was at about two weeks before. It resumed its normal decline the day after the new data.

I've got the IFR in Texas at .254% as of EOD yesterday -- as compared to about .232% had the new data not been added.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby ponyboy » Wed Jul 29, 2020 12:41 pm

By the way, per WebMD, "about 8 in 10 people who get COVID-19, the illness caused by the coronavirus called SARS-CoV-2, will get mildly sick." Most of those people with mild symptoms do not get any test whatsoever, which is why we're almost certainly way-over-estimating IFR. In my calcs I go with the CDC director's assumption that for every 11 people who get the virus, only 1 gets a test.

Speaking of which, I'm fairly certain that I got the virus about three or four weeks ago. It was kind of something I'd never experienced before -- low grade fever, dry cough, and lots of fatigue. But it just felt wierd, especially when I laid down. Like most folks, I just self-isolated and recovered at home. No official test.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby JasonB » Wed Jul 29, 2020 1:49 pm

ponyboy wrote:By the way, per WebMD, "about 8 in 10 people who get COVID-19, the illness caused by the coronavirus called SARS-CoV-2, will get mildly sick." Most of those people with mild symptoms do not get any test whatsoever, which is why we're almost certainly way-over-estimating IFR. In my calcs I go with the CDC director's assumption that for every 11 people who get the virus, only 1 gets a test.

Speaking of which, I'm fairly certain that I got the virus about three or four weeks ago. It was a wierd thing I'd never experienced before -- low grade fever, dry cough, and lots of fatigue. But it just felt wierd, especially when I laid down. Like most folks, I just self-isolated and recovered at home. No official test.


1) Keep in mind that happens with the flu also.
2) We are only at a 6-10% positivity rate on tests. And the people getting tested are either sick or have come into contact with someone who had corona. This means for the general population, if you randomly tested them, it would be much lower than the positivity rate than you see with people actually getting tested. The estimates and calculations for how to calculate a true IFT are all WAG right now. But just think about it this way - with the large, large number of flu cases we get annually, and with Covid currently infecting single digits percentage of the population in Texas... in 2017-2018, the CDC thinks 5M+ people in Texas got the flu and we had 11K people die that year of P+I (which is why I think the newspaper articles citing 9500 I deaths is bogus, because there is no way there were only 2K non-I pneumonia deaths that year. But I can't find any data tables that just show the I deaths a that high). Even if you stretch and think that 1M people in Texas have already had Covid, that means 5K*5 = 25K deaths with the same number of infections, which is significantly larger than the 11K P+I deaths of a very strong flu season. So it makes me think there is something wrong with the IFR calculation.
3) The reason for the added deaths is that they are going from what the doctor indicated the cause of death is. The coroners are too backloaded to do full investigations, so they had 600 bodies waiting for full investigation. When they saw that over 99% of the time, when a doctor said it was Corona and the Morgue agreed, they decided to skip the step and just declare them as covid deaths.
4) Ponyboy, I'd be interested in seeing three numbers from your data. First would be deaths/recoveries. Second is deaths/total number of positive tests from 4 weeks ago. Third is deaths/(number of positive tests from 4 weeks ago * 11). That will get all the data out, I woudl be interested in seeing those numbers.
5) Hospitalizations continue to go down, which hopefully (as long as it isn't due to reporting error) means we should hit peak death per day next week and then start dropping.
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

Re: Football this fall? No way

Postby ClassOf81 » Wed Jul 29, 2020 1:51 pm

mustangxc wrote:I heard that the Texas healthcare system reformulated its classification system and as a result there are an additional 600 deaths attributed to COVID-19. Not sure what changed in the classification system so I don't know what to think.

Meanwhile, hospitals have been told to bypass the CDC and send their info directly to a branch of the federal government, leading to fears that the numbers are going to be artificially reduced.

I know one person who got it and died, I know another who recently lost his mom to it and more and more people close to me have lost a friend or a family member or a co-worker. Are we at 150K yet? If not, it's coming right away.

I miss sports, but compared to this, they're not important.
User avatar
ClassOf81
Heisman
 
Posts: 1037
Joined: Wed Jul 05, 2000 3:01 am
Location: Highland Park, Texas

Re: Football this fall? No way

Postby ponyboy » Wed Jul 29, 2020 2:12 pm

JasonB wrote:Ponyboy, I'd be interested in seeing three numbers from your data. First would be deaths/recoveries. Second is deaths/total number of positive tests from 4 weeks ago. Third is deaths/(number of positive tests from 4 weeks ago * 11). That will get all the data out, I woudl be interested in seeing those numbers.


Here's all the data from 7/28:

Positive Tests -- Negative Tests -- Deaths -- Infected -- % of Population -- IFR
394,265--------3,082,661-----5,877----4,336,915------14.5%------0.254%

I'm currently using infections from 3 weeks ago/current deaths. If I change my IFR calculation to 4 weeks ago, the number is 0.334%. (Positive tests from 3 weeks ago = 2,316,435. 4 weeks ago = 1,759,846)

Does that give you what you're looking for? Or I can just email you my spreadsheet if you want to mess with the numbers. PM me if you'd like that.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby ponyboy » Wed Jul 29, 2020 2:38 pm

Ok, this is really funny. Just got an email from Baylor University that my son should have received his COVID-19 kit in the mail. He needs to complete and return today. If he tests positive, he can't come to school. If he tests negative, he can.

Um, if he tests positive today, he'll be over the thing before school begins. If he tests negative, he can still acquire COVID-19 before school begins -- and then proceed to infect all of the campus.

Bureaucracy: Making the world a better place for us all.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby JasonB » Wed Jul 29, 2020 3:24 pm

ponyboy wrote:Ok, this is really funny. Just got an email from Baylor University that my son should have received his COVID-19 kit in the mail. He needs to complete and return today. If he tests positive, he can't come to school. If he tests negative, he can.

Um, if he tests positive today, he'll be over the thing before school begins. If he tests negative, he can still acquire COVID-19 before school begins -- and then proceed to infect all of the campus.

Bureaucracy: Making the world a better place for us all.


Even the USVI requires a test three days before entry. That is absurd!
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

Re: Football this fall? No way

Postby ponyboy » Wed Jul 29, 2020 3:36 pm

JasonB wrote:
ponyboy wrote:Ok, this is really funny. Just got an email from Baylor University that my son should have received his COVID-19 kit in the mail. He needs to complete and return today. If he tests positive, he can't come to school. If he tests negative, he can.

Um, if he tests positive today, he'll be over the thing before school begins. If he tests negative, he can still acquire COVID-19 before school begins -- and then proceed to infect all of the campus.

Bureaucracy: Making the world a better place for us all.


Even the USVI requires a test three days before entry. That is absurd!


Luckily, I misread. They need confirmation of his address today so they’ll know where to send the test kits.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby JasonB » Wed Jul 29, 2020 3:51 pm

ponyboy wrote:
JasonB wrote:Ponyboy, I'd be interested in seeing three numbers from your data. First would be deaths/recoveries. Second is deaths/total number of positive tests from 4 weeks ago. Third is deaths/(number of positive tests from 4 weeks ago * 11). That will get all the data out, I woudl be interested in seeing those numbers.


Here's all the data from 7/28:

Positive Tests -- Negative Tests -- Deaths -- Infected -- % of Population -- IFR
394,265--------3,082,661-----5,877----4,336,915------14.5%------0.254%

I'm currently using infections from 3 weeks ago/current deaths. If I change my IFR calculation to 4 weeks ago, the number is 0.334%. (Positive tests from 3 weeks ago = 2,316,435. 4 weeks ago = 1,759,846)



Does that give you what you're looking for? Or I can just email you my spreadsheet if you want to mess with the numbers. PM me if you'd like that.


Okay, here is what I am looking for:

First is the deaths/recoveries. Texas says 5877/244449 = 2.4%. That is the number of deaths compared to the number of people we know who have actually recovered fully.

Second would be the extrapolation. We *think* that there are 11x people who have had the virus than actually got tested. So, that changes the above calculation to result in 0.22.

Now lets take deaths/positive tests from 4 weeks ago. That should be 5877/(some number around 300K). Then find 5877/(that number * 11), which would represent that extrapolation.

What I am guessing is that the number of people who tested positive in that last part is going to be around the same number as the recoveries in the first part. I would advise that, if you are going to calculate the IFR, that you should just use the recovery rate rather than the number of positive tests.

So, what I would suggest is that the data shows that the IFR is, in a best case scenario, 0.22. And in a worst case scenario, 2.4%. The 2.4% number, of course, is not accurate because there are certainly a lot of people who have had it than got tested. But I think it is important to point out the range, because the 0.22 result is based on a GIGANTIC assumption that 11x the cases exist out there. And nobody really knows that at all. So providing the range of possibilities is really important.

Objectively speaking, with the current number of hospitalizations, we are going to be around 9K deaths by the end of August (you can get there with either 20% of hospitalizations die or by the current death rate of 150 people per day for the next two weeks and then gradually slowing after that). A typical flu season in Texas has a lot more infections, and a lot fewer deaths. So I don't see how the IFR would be anywhere close to the flu. That part of the math doesn't really add up.

On top of that, the CDC still promotes the IFR range to be 0.8 to 0.5, and most likely 0.65. So I don't think the 11x multiplier is the right one. And that 11x multiplier would probably need to be estimated on a locality basis because the testing practices are so different from place to place right now. The multiplier for Texas is going to be way different than Florida for example, and even within Texas, the multiplier in Lubbock is probably way different than Dallas.

I found this article that suggests a 0.2x IFR: https://www.medrxiv.org/content/10.1101 ... 20101253v3

However, it hasn't been peer reviewed, and I would point out that it is based on seroprevelence. The problem with that is that those tests have been notoriously unreliable so far, and the labs are not good. There are two test producers who have what would be considered a decent specificity and sensitivity. And depending on what lab it gets sent to, results are all over the place.
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

Re: Football this fall? No way

Postby ponyboy » Thu Jul 30, 2020 9:08 am

Making the call on the multiplier if a tough one for sure. But you've got to do it -- it's a choice forced upon us. I've seen numbers from 11x to 35x and more. I chose the most conservative figure, one backed by the CDC.

By the way, I sent my data to an academic and professional who's intimately involved in all this, a man in his late '70s who has the experience of several pandemics over his time. I'm posting his response:

Beautifully done. Where did you get this data?

It is important to consider different effects on different groups; one size does not fit all. The controlling factor is the health of the immune system: a healthy system can fight off the virus while a compromised system struggles and occasionally fails, i.e., the patient dies.

An overwhelmed health care system will lead to more deaths, so spreading out the cases can help.

Also, even if a patient doesn’t die, how miserable will he be and what may be the long-term consequences of his illness? These consequences worry me and I’m scared to death of a respirator.

So it seems to me that we will continue for some time to need to strike an ever-changing balance between health (including mental health) and our economic needs. Statistics like yours, constantly updated, are vital to maintaining the balance (as best we can.)

People just need to remember that it is not much of a solution to survive the pandemic in abject poverty and mass starvation. That is going to be the fate of much of the world, I fear.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby JasonB » Thu Jul 30, 2020 10:01 am

As I pointed out above, the 11x multiplier doesn't correlate to the number of deaths we are seeing right now with such a small infection percentage in the population. In Probability terms, it is a Subjective Probability rather than Experimental Probability at this point.

It doesn't matter, because unfortunately the number of deaths in Texas is really high right now, unfortunately. If you combine the past two weeks with the next two weeks, we will equal the number of deaths in a one month period that we normally hit in a flu season in Texas.

Condolences to ECM. Herman Cain attended the Trump Rally in Tulsa, got Covid, and just passed away.
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

Re: Football this fall? No way

Postby ponyboy » Thu Jul 30, 2020 10:57 am

JasonB wrote:If you combine the past two weeks with the next two weeks, we will equal the number of deaths in a one month period that we normally hit in a flu season in Texas.

Yep. Unless something materially changes (e.g. reaching "herd immunity" or another some such barrier to the continued growth), I think you're right about this point, for two reasons:

1. As expected, the daily deaths curve took a sharp turn for the worse in early July -- from an average of about 30 deaths per day (a figure we'd consistently been seeing for the three months leading up to that date) to what looks to be about 300 deaths per day today. And looking at the curve, I do think we could easily find ourselves in the 500 deaths per day region by mid August.

2. Normal flu season is 6,500 deaths. We're at about 2,316 new deaths in the last two weeks. And I'd estimate we'll get another 5,250 in the next two weeks, given the direction of the curve right now. (Total deaths are currently 6,190).

If this happens, by mid August we'll be very to close my estimate from a few weeks back that total COVID-19 deaths in Texas would end up at about 10.5K when all is said and done. (And I predicted that this thing would be "played out" by August 31st). It'll be interesting to see whether we do get over that hump soon and start reversing course.
ponyboy
PonyFans.com Super Legend
 
Posts: 15134
Joined: Wed Mar 22, 2000 4:01 am
Location: University Park,TX US

Re: Football this fall? No way

Postby JasonB » Thu Jul 30, 2020 5:20 pm

ponyboy wrote:
JasonB wrote:If you combine the past two weeks with the next two weeks, we will equal the number of deaths in a one month period that we normally hit in a flu season in Texas.

Yep. Unless something materially changes (e.g. reaching "herd immunity" or another some such barrier to the continued growth), I think you're right about this point, for two reasons:

1. As expected, the daily deaths curve took a sharp turn for the worse in early July -- from an average of about 30 deaths per day (a figure we'd consistently been seeing for the three months leading up to that date) to what looks to be about 300 deaths per day today. And looking at the curve, I do think we could easily find ourselves in the 500 deaths per day region by mid August.

2. Normal flu season is 6,500 deaths. We're at about 2,316 new deaths in the last two weeks. And I'd estimate we'll get another 5,250 in the next two weeks, given the direction of the curve right now. (Total deaths are currently 6,190).

If this happens, by mid August we'll be very to close my estimate from a few weeks back that total COVID-19 deaths in Texas would end up at about 10.5K when all is said and done. (And I predicted that this thing would be "played out" by August 31st). It'll be interesting to see whether we do get over that hump soon and start reversing course.


We aren't close to herd immunity though... 10% of the population and 10K deaths is what we are looking at. Herd immunity is 50%+ of the population.

Flu on average kills 11.1 Texas per 100,000 residents. 30M people, 3,300K per year on average.
JasonB
PonyFans.com Super Legend
 
Posts: 7129
Joined: Mon Aug 27, 2001 3:01 am
Location: Allen, Tx, USA

PreviousNext

Return to Around the Hilltop

Who is online

Users browsing this forum: No registered users and 11 guests