COVID-19 Megathread 6: Return of the Omicron

[jimrtex]

AYY LMAO
https://twitter.com/thehill/status/1339701188086906882

The actual ruling applied to restaurants offering eat-in dining in general, saying that the state failed to demonstrate a risk factor.

[jimrtex]

The US is vaccinating people way too slowly. A top doctor says the federal government is to blame.

Dr. Ashish K. Jha, a top US doctor and the dean of Brown University School of Public Health, on Tuesday shared in a Twitter thread why he believed the rollout of COVID-19 vaccines in the US was flawed, and he said the issue begins with the federal government.

The US Food and Drug Administration in December authorised two different vaccines for COVID-19 — one created by Moderna and the National Institutes of Health, and another created by Pfizer and BioNTech — for emergency use in the US.

While people across the US have already begun to receive the vaccine, a limited supply means the vaccine won’t be widely available to all who need it well into 2021, prolonging the pandemic that has so far killed more than 336,000 people in the US, according to data from Johns Hopkins University.

According to an analysis published Tuesday by NBC News, at the current pace, it could take the US nearly a decade to vaccinate enough Americans to meaningfully bring the pandemic under control. The White House previously said it aimed to vaccinate 80% of Americans by the end of June, which would require more than 3 million vaccinations per day, according to the report. So far, the US has vaccinated just about 2 million people in 16 days.

More Trump administration incompetence, or a deliberate effort to sabotage the Biden administration at the cost of American lives? Given the last four years, either (or even both) seem plausible.

Flawed analysis by NBC.

It is unlikely that the aim was to vaccinate 80% of the US population by June, but rather 80% of the adult population, or around 200 million persons.

They also appear to be counting doses rather than persons vaccinated.

In essence they are projecting that initial doses will be administered for 21 days, and then for the next 21 days no first doses will be administered, as they go back and administer the second dose.

[jimrtex]

The updated numbers for COVID-19 in the U.S. are in for 12/30 per: https://www.worldometers.info/coronavirus/country/us/

I'm keeping track of these updates daily and updating at the end of the day, whenever all states finish reporting for that day.

ΔW Change: Comparisons of Weekly Day-to-day Growth or Decline of COVID-19 Spread/Deaths.

• IE: Comparing the numbers to the same day of last week, are we flattening the curve enough?

Σ Increase: A day's contribution to overall percentage growth of COVID-19 cases/deaths.

• IE: What's the overall change in the total?

Older Numbers (Hidden in spoiler mode to make the post more compact)

Spoiler alert! Click Show to show the content.

Show

<Last Numbers for 3/26-3/28 in this Post>
<Last Numbers for 3/29-4/4 in this Post>
<Last Numbers for 4/5-4/11 in this Post>
<Last Numbers for 4/12-4/18 in this Post>
<Last Numbers for 4/19-4/25 in this post>
<Last Numbers for 4/26-5/2 in this Post>
<Last Numbers for 5/3-5/9 in this Post>
<Last Numbers for 5/10-5/16 in this Post>
<Last Numbers for 5/17-5/23 in this Post>
<Last Numbers for 5/24-5/30 in this Post>
<Last Numbers for 5/31-6/6 in this Post>
<Last Numbers for 6/7-6/13 in this Post>
<Last Numbers for 6/14-6/20 in this Post>
<Last Numbers for 6/21-6/27 in this Post>
<Last Numbers for 6/28-7/4 in this Post>
<Last Numbers from 7/5-7/11 in this Post>
<Last Numbers from 7/12-7/18 in this Post>
<Last Numbers from 7/19-7/25 in this Post>
<Last Numbers from 7/26-8/1 in this Post>
<Last Numbers from 8/2-8/8 in this Post>
<Last Numbers from 8/9-8/15 in this Post>
<Last Numbers from 8/16-8/22 in this Post>
<Last Numbers from 8/23-8/29 in this Post>
<Last Numbers from 8/30-9/5 in this Post>
<Last Numbers from 9/6-9/12 in this Post>
<Last Numbers from 9/13-9/19 in this Post>
<Last Numbers from 9/20-9/26 in this Post>
<Last Numbers from 9/27-10/3 in this Post>
<Last Numbers from 10/4-10/10 in this Post>
<Last Numbers from 10/11-10/17 in this Post>
<Last Numbers from 10/18-10/24 in this Post>
<Last Numbers from 10/25-10/31 in this Post>
<Last Numbers from 11/1-11/7 in this Post>
<Last Numbers from 11/8-11/14 in this Post>
<Last Numbers from 11/15-11/21 in this Post>
<Last Numbers from 11/22-11/28 in this Post>
<Last Numbers from 11/29-12/5 in this Post>
<Last Numbers from 12/6-12/12 in this Post>
<Last Numbers from 12/13-12/19 in this Post>

12/20: <Sunday>

• Cases: 18,267,579 (+189,811 | ΔW Change: ↑1.02% | Σ Increase: ↑1.06%)
• Deaths: 324,869 (+1,468 | ΔW Change: ↑6.61% | Σ Increase: ↑0.80%)

12/21: <M>

• Cases: 18,473,716 (+206,137 | ΔW Change: ↑0.28% | Σ Increase: ↑1.13%)
• Deaths: 326,772 (+1,903 | ΔW Change: ↑16.75% | Σ Increase: ↑0.59%)

12/22: <T>

• Cases: 18,684,628 (+210,912 | ΔW Change: ↑4.87% | Σ Increase: ↑1.14%)
• Deaths: 330,824 (+4,052 | ΔW Change: ↑35.79% | Σ Increase: ↑1.24%)

12/23: <W>

• Cases: 18,917,152 (+232,524 | ΔW Change: ↓6.50% | Σ Increase: ↑1.24%)
• Deaths: 334,218 (+3,394 | ΔW Change: ↓3.14% | Σ Increase: ↑1.06%)

12/24 (Holiday): <Þ>

• Cases: 19,111,326 (+194,174 | ΔW Change: ↓17.07% | Σ Increase: ↑1.03%)
• Deaths: 337,066 (+2,848 | ΔW Change: ↓15.01% | Σ Increase: ↑0.85%)

12/25 (Holiday): <F>

• Cases: 19,210,166 (+98.840 | ΔW Change: ↓62.21% | Σ Increase: ↑0.52%)
• Deaths: 338,263 (+1,197 | ΔW Change: ↓58.96% | Σ Increase: ↑0.36%)

12/26: <S>

• Cases: 19,433,847 (+223,681 | ΔW Change: ↑18.09% | Σ Increase: ↑1.16%)
• Deaths: 339,921 (+1,658 | ΔW Change: ↓35.39% | Σ Increase: ↑0.49%)

12/27: <Sunday>

• Cases: 19,573,847 (+140,000 | ΔW Change: ↓26.24% | Σ Increase: ↑0.72%)
• Deaths: 341,138 (+1,217 | ΔW Change: ↓17.10% | Σ Increase: ↑0.36%)

12/28: <M>

• Cases: 19,781,624 (+207,777 | ΔW Change: ↑0.80% | Σ Increase: ↑1.06%)
• Deaths: 343,182 (+2,044 | ΔW Change: ↑7.41% | Σ Increase: ↑0.60%)

12/29 (Yesterday): <T>

• Cases: 19,977,704 (+196,080 | ΔW Change: ↓7.03% | Σ Increase: ↑0.99%)
• Deaths: 346,579 (+3,397 | ΔW Change: ↓16.16% | Σ Increase: ↑0.99%)

12/30 (Today): <W>

• Cases: 20,216,991 (+239,287 | ΔW Change: ↑2.91% | Σ Increase: ↑1.20%)
• Deaths: 350,778 (+4,199 | ΔW Change: ↑23.72% | Σ Increase: ↑1.21%)

You might want to incorporate this data in your reports.

https://covid.cdc.gov/covid-data-tracker/#vaccinations

This only updates MWF, and they will skip holidays - at least according to notes that were present before Wednesday's update. The map and results are new with the Wednesday update. so this is probably a work in progress. The early notes also said this likely lagged actual vaccinations by a few days.

[jimrtex]

The US is vaccinating people way too slowly. A top doctor says the federal government is to blame.

Dr. Ashish K. Jha, a top US doctor and the dean of Brown University School of Public Health, on Tuesday shared in a Twitter thread why he believed the rollout of COVID-19 vaccines in the US was flawed, and he said the issue begins with the federal government.

The US Food and Drug Administration in December authorised two different vaccines for COVID-19 — one created by Moderna and the National Institutes of Health, and another created by Pfizer and BioNTech — for emergency use in the US.

While people across the US have already begun to receive the vaccine, a limited supply means the vaccine won’t be widely available to all who need it well into 2021, prolonging the pandemic that has so far killed more than 336,000 people in the US, according to data from Johns Hopkins University.

According to an analysis published Tuesday by NBC News, at the current pace, it could take the US nearly a decade to vaccinate enough Americans to meaningfully bring the pandemic under control. The White House previously said it aimed to vaccinate 80% of Americans by the end of June, which would require more than 3 million vaccinations per day, according to the report. So far, the US has vaccinated just about 2 million people in 16 days.

More Trump administration incompetence, or a deliberate effort to sabotage the Biden administration at the cost of American lives? Given the last four years, either (or even both) seem plausible.

Flawed analysis by NBC.

It is unlikely that the aim was to vaccinate 80% of the US population by June, but rather 80% of the adult population, or around 200 million persons.

They also appear to be counting doses rather than persons vaccinated.

In essence they are projecting that initial doses will be administered for 21 days, and then for the next 21 days no first doses will be administered, as they go back and administer the second dose.

My impression was that the 80% of the population as a whole was a little unrealistic. 80% of the adult population sounds more reasonable.

We won't be achieving herd immunity for some time, but if we can get the rate of infection and death down, we might be able to return to a semi-normal.

The number of first doses doubled in the next five days.

[jimrtex]

What evidence do we have that the dose that is being administered is the most efficient dosage? 
(The mere fact that it was the only dosage fully tested is not, by itself, evidence of that.)

https://www.goodrx.com/blog/why-does-the-covid-19-vaccine-need-two-shots/

This suggests that one dose of the Pfizer vaccine is around 50% effective, but acknowledges that analysis is flawed.

In the Phase 3 trials they gave a bunch of people the vaccine, and an equal number a placebo. The study participants were demographically.

Only a very small percentage of those who were given the placebo contracted COVID-19. The tests weren't that long and most people have not contracted COVID-19 (perhaps 7% in the USA over 10 months). There would be no reason to encourage participants to be promiscuous. It is possible that if the risks were explained to participants they would be more cautious than non-participants. There may well have been frequent testing.

There were 43,000 participants in the Pfizer Phase 3 trials, presumably evenly split between the two groups. 162 placebo recipients contracted COVID-19 (0.75%). That is 99.25% of the placebo recipients remained uninfected.

But only 8 of the vaccinated recipients contracted COVID-19 (0.04%). The 95% efficacy came from this difference 162 vs. 8. There is some good/bad luck here.

If they did another study with 21,500 persons in each group, would they get the same results? Quite, unlikely. There might be only 6 persons in the vaccine group, or 10 persons. But it is quite unlikely that it is an extreme outlier and there would be 105.

Likewise in the placebo group it might be 173 or 157.

Now back to one-dose results. They didn't give 21,500 persons one dose. Some number only got one dose for some reason. They might have been in a severe auto accident. They might have had an illness unrelated to the vaccine (which they might have attributed to the vaccine). "I don't remember when I had diarrhea this bad", doesn't mean they don't recall, but simply that they did not keep a journal of their bowel movements. "June 27, 2017 ..." Oh yeah, that was worse. Some might not liked the weekly test, or the shot did make them feel bad. I have vaccinations where I'd noticed the soreness in my shoulder, and others where I didn't remember, except when I reached over and felt the bandaid. I think the COVID-19 is more like the former. It is not symptomless.

Some might have only got one dose because they were infected before the second dose. There must of been some screening. They would not want study participants who already were heath compromised or drug addicts. They probably wanted those who had good medical records, etc. They wanted persons who were likely to show up for the weekly COVID-19 test, and would come back in 3 weeks for the second dose.

So lets say 95% did get the second dose. That would mean 2150 did not. Apparently, 0.75%/2 of these or 8 were infected. This might or note be very accurate due to the small sample size. But what if 98% did get the second dose. Then 430 only got one dose. If 0.75%/2 of those were infected, only two would be infected. The sample is quite small, and conclusions can not be made about the efficacy.

If infection occurred before the second dose could be administered, then the sample is biased. It is not expected that immunity is instantaneous. If the reason for no second dose is that you were already infected, it is not the same as those who received only one dose and would not have been infected during the next five years because they had become immune.

There must be some science behind developing a vaccine that requires two doses vs. one that requires one dose, just as there are reasons that require the vaccine to be kept at extremely cold temperatures. Different companies developed different strategies. They may have even been encouraged to do so by the CDC.

"Let's all do the same thing. If we get lucky we will be able to have lots of production in 6 months. If we don't it might be 2-1/2 years."

"If ours come in first, we will be rich! If not the government will at least pay our development costs."

[jimrtex]

I agree we don’t have the same certainty about a half dosage as we do about a full dosage.  But the data seems to suggest that a half dosage is more than half as effective as a full dose.  And in that case, under these circumstances, we should be willing to tolerate some more uncertainly in exchange for what is likely much more efficiency.

This is not accurate. The data in this case is not robust enough to be making statistical conclusions.

Instead of the effective rate, look at the ineffective rate.

Compared to non-immunized (placebo) participants, the placebo participants are 20 times as likely to be infected. Meanwhile those with one dose are 10 times a likely to be infected.

[jimrtex]

Friday's report from the CDC is the first in which new vaccinations exceeded newly distributed doses.

https://covid.cdc.gov/covid-data-tracker/#vaccinations

This would indicate that any delays are not due to the federal government holding back on distribution, but rather on injecting them. But only 30% of distributed doses have been injected.

24 States have now exceed vaccination to 2% of their populations.

16 of 27 states (59%) with a Republican governor have exceeded 2%.
8 of 23 states (35%) with a Democratic governor have exceeded 2%.

[jimrtex]

I agree we don’t have the same certainty about a half dosage as we do about a full dosage.  But the data seems to suggest that a half dosage is more than half as effective as a full dose.  And in that case, under these circumstances, we should be willing to tolerate some more uncertainly in exchange for what is likely much more efficiency.

This is not accurate. The data in this case is not robust enough to be making statistical conclusions.

Instead of the effective rate, look at the ineffective rate.

Compared to non-immunized (placebo) participants, the placebo participants are 20 times as likely to be infected. Meanwhile those with one dose are 10 times a likely to be infected.

What I’m saying is that just by making statements like “robust enough to be making statistical conclusions”, you are buying into the outdated frequentist orthodoxy.  Modern Bayesian statistics doesn’t think like that.  The question shouldn’t be a binary “ is this enough to change our conclusion”?  All additional data we get should change our conclusion, the only question is how much it should change it.

And the people who got one does are only 10 times more likely to be infected when you measure them within the first few days after getting the vaccine.  When measured 14-21 days after getting the first dose, they are less than twice as likely to get infected.

There were 39 infections between the 1st and 2nd doses (1.857 per day)
There were 2 infections within 7 days of 2nd dose (0.286 per day).
There were 9 infections within the (average) next 41 days (0.220 per day)

The rate within 7 days after the the 2nd dose may be close to that for the 7 days before the 2nd dose, but perhaps not. It may be possible that the second dose begins to take effect quicker than the first if the immune system is already triggered.

Among the placebo group:

There were 82 infections between the 1st and 2nd doses (3.905 per day)
There were 21 infections within 7 days of the 2nd dose (3.000 per day)
There were 172 infections within the (average) next 39 days (4.410 per day*).

*It appears that there was variation in the infection rate over the surveillance period coincident with increased community prevalence (if there are more people who may transmit the infection to you, you are more likely to become infected).

I agree with your conclusion that there was little difference in infections in the first eleven days or so between the placebo group and the vaccine group.

We can agree that one dose is more effective than one or two placebo doses.

But it is not clear that one dose is as effective as two doses, or if there is variation among population groups. 20 μg might be enough in some persons, and 60 μg might be barely enough in others.

And only 30% of doses that have been distributed to states has actually been injected into arms (there may be some reporting lag as well as distribution lag).

[jimrtex]

How did Israel get access to so much vaccine?  They’ve already vaccinated 20% of their population; not other major country is over 2%.  Even if we had actually used all our supply, we’d only be at 6%.

Prime Minister Netanyahu?

Pfizer has agreed to speed distribution to Israel in exchange for records of adverse reactions. Israel keeps more records of their citizens, uses paper ballots, requires ID to vote, and has walls to keep out illegal immigrants.

[jimrtex]

How did Israel get access to so much vaccine?  They’ve already vaccinated 20% of their population; not other major country is over 2%.  Even if we had actually used all our supply, we’d only be at 6%.

Prime Minister Netanyahu?

Pfizer has agreed to speed distribution to Israel in exchange for records of adverse reactions. Israel keeps more records of their citizens, uses paper ballots, requires ID to vote, and has walls to keep out illegal immigrants.

Ah, so build the wall so we can all get vaccinated. Ok.

All the MAGA cultist anti vaxxers in the US are dragging down the numbers.

It's shocking what a disingenuous, belligerent nutcase jimrtex is when discussing the actual issues. He should really stick to the redistricting app.

Why do you think states with Republican governors are doing better with actual application of vaccinations?

For example, there are 16 States with more than 3% vaccinated. 11 of those 15 have Republican governors:

WV, SD, ND, VT, AK, NE, NH, MT, TN, IA, OK

Only five have Democrat governors.

ME, CT, CO, NM, RI

President Trump should get more doses to the Dakotas so that they don't run out.

[jimrtex]

How did Israel get access to so much vaccine?  They’ve already vaccinated 20% of their population; not other major country is over 2%.  Even if we had actually used all our supply, we’d only be at 6%.

Prime Minister Netanyahu?

Pfizer has agreed to speed distribution to Israel in exchange for records of adverse reactions. Israel keeps more records of their citizens, uses paper ballots, requires ID to vote, and has walls to keep out illegal immigrants.

Ah, so build the wall so we can all get vaccinated. Ok.

All the MAGA cultist anti vaxxers in the US are dragging down the numbers.

It's shocking what a disingenuous, belligerent nutcase jimrtex is when discussing the actual issues. He should really stick to the redistricting app.

Why do you think states with Republican governors are doing better with actual application of vaccinations?

For example, there are 16 States with more than 3% vaccinated. 11 of those 15 have Republican governors:

WV, SD, ND, VT, AK, NE, NH, MT, TN, IA, OK

Only five have Democrat governors.

ME, CT, CO, NM, RI

President Trump should get more doses to the Dakotas so that they don't run out.

It seems like vaccine distribution right now is heavily correlated with how white a state is.  Likely much more another example of institutional racism than anything else.

Here are the worst states by % of shots given:

Jurisdiction Doses distributed Doses administered % shots used
Alabama 326,850 76,528 23.4
Georgia 770,625 183,870 23.9
North Carolina 820,825 211,572 25.8
Alaska 132,350 35,027 26.5
Arizona 563,025 151,844 27.0
California 2,833,400 782,638 27.6
Nevada 205,200 58,651 28.6
Mississippi 192,750 55,399 28.7

Hmm, what do AL, GA, NC, and MS have in common?  How about CA, NV, and AZ?

Hmm, and that is distinct from LA and TX?

I can't think of anything.

p.s. I think injections per 100K is a better metric.

[jimrtex]

According to the OurWorldInData vaccine tracker (https://ourworldindata.org/covid-vaccinations), today is the first day in which the US recorded vaccinating over 1 million people. (I'm not sure what the lag in this data is.)

And we are now at a 7-day average of almost 800k doses, compared with just 400k a week ago.  It does seem like we are getting to where we need to be here.

This is likely source for US injections.

https://covid.cdc.gov/covid-data-tracker/#vaccinations

The notes say that those performing the injections have 72 hours to report injections to state and local officials, and there may be further delays between that receipt and passing it along to the CDC.

If I were someone performing injections, I would be wanting to report daily after the close of business. It seems unduly difficult to be reporting on Thursday what you did on Monday. On the other hand, I can see the person making the report on the next day. If the clinic, etc. closes at 4 PM or 5 PM, etc. are they really going to have someone entering the records overnight? They will be doing it the next AM.

The CDC data is as of 6 AM. So the CDC data for 1 million+ doses is from reports made between 6 AM Thursday and 6 AM Friday. The CDC is not issuing their updates until late in the afternoon.

I doubt that many states reported at 5 AM on Friday, more likely at 4 PM or 5 PM on Thursday (local time). This would give them as much time to input data from clinics.

There was an inflection last Tuesday (see ourworldindata chart, zap out Israel, UAE, etc. to expand vertical scale).

If we assume that was for Sunday injections reported on Monday to state authorities and forwarded to CDC late on Monday, then there is a two-day lag, in part because the CDC report is so early in the morning (originally the CDC report was based on 9 AM and they probably became annoyed at east coast states dumping data on them just before the deadline).

So this means that 1 million plus happened on Wednesday. It might not be possible to confirm this since the CDC is only reporting MTWTF (originally it was MWF) and this is the first week that they have done 5-per-week.

The next report on Monday would be for Thursday-Friday-Saturday injections. Saturday will be less than Thursday-Friday, but greater than Sunday. So the Monday report may be just shy of three million.

The CDC has just begin differentiating between
(1) Doses injected.
(2) Persons injected with 1 or more doses.
(3) Persons injected with 2 doses.

About 1/4 of the Wednesday(?) doses were second doses, with the total up to 1.6M. Those who have had only one dose is 9.0 million.

Type of does is 58% Pfizer-42% Moderna. Moderna should lag a bit because they were approved later, and the period between the two doses is 28 for Moderna vs. 21 Pfizer.

I suspect that at least initially, most second doses will happen precisely 3 weeks after the first. If you have done one, you are not going to be waiting around 5 weeks before making an appointment, and you may be getting the call back.

[jimrtex]

https://twitter.com/piyushmittal/status/1350868085129953281
Particularly chilling to me is this part:

Jones announced Saturday on Twitter that she learned of the warrant and plans to turn herself in on Sunday. The Florida Department of Law Enforcement confirmed there is a warrant for Jones’ arrest but said it cannot disclose what charges she faces until she is in custody.

I was not previously aware that premeditated "arrest first, charges afterward" was a thing that could happen in the United States.

This is and has always been a nothing story.  Do some basic research and fact-checking.

How come she has Florida backwards in her picture?

[jimrtex]

https://twitter.com/piyushmittal/status/1350868085129953281
Particularly chilling to me is this part:

Jones announced Saturday on Twitter that she learned of the warrant and plans to turn herself in on Sunday. The Florida Department of Law Enforcement confirmed there is a warrant for Jones’ arrest but said it cannot disclose what charges she faces until she is in custody.

I was not previously aware that premeditated "arrest first, charges afterward" was a thing that could happen in the United States.

This is and has always been a nothing story.  Do some basic research and fact-checking.

How come she has Florida backwards in her picture?

It's a thing with some selfie-cameras.  A mirror effect is created.  Trying to comprehend why makes my brain hurt.

She should have fixed it before she posted it. She probably didn't notice.

[jimrtex]

2) Jones is not a coronavirus "scientist" and describing her as such, like NPR and MSNBC did, is extremely generous and construed to give the story more legs than it deserves.  She is a doctoral student of geography at FSU.  Her skill set, as applicable to COVID-19, is in mapping and GIS applications.  She has no education or special skills in epidemiology, biology or public health.  While calling her a "(data) scientist" may be technically correct, media outlets owe their readers a fuller explanation of her role and educational background to make clear she is not a medical professional. 

Is she still a doctoral student at FSU? I thought she had dropped out.

[jimrtex]

https://twitter.com/piyushmittal/status/1350868085129953281
Particularly chilling to me is this part:

Jones announced Saturday on Twitter that she learned of the warrant and plans to turn herself in on Sunday. The Florida Department of Law Enforcement confirmed there is a warrant for Jones’ arrest but said it cannot disclose what charges she faces until she is in custody.

I was not previously aware that premeditated "arrest first, charges afterward" was a thing that could happen in the United States.

I assume they mean, "cannot disclose to the media," what charges she faces. Which is entirely reasonable.

2) Jones is not a coronavirus "scientist" and describing her as such, like NPR and MSNBC did, is extremely generous and construed to give the story more legs than it deserves.  She is a doctoral student of geography at FSU.  Her skill set, as applicable to COVID-19, is in mapping and GIS applications.  She has no education or special skills in epidemiology, biology or public health.  While calling her a "(data) scientist" may be technically correct, media outlets owe their readers a fuller explanation of her role and educational background to make clear she is not a medical professional. 

Should NPR and MSNBC call her a, "Data Scientist, which is literally the job she had with the state, so not a 'real' scientist, kind of like Jill Biden is not a 'real' doctor, because if we call Dr. Biden a 'real' doctor, people might think she carries a stethoscope and one of those knee-mallet things everywhere she goes, and similarly people might think Jones works under a fume hood with safety glasses on while Erlenmeyers sit bubbling away on magnetic stirrers?"

Anyway, non-story is a non-story. Don't comment publicly on data that's not yours. Don't take data home with you. Use your head.

Her job description appears to be Geographic Information System Analyst.

[jimrtex]

According to the OurWorldInData vaccine tracker (https://ourworldindata.org/covid-vaccinations), today is the first day in which the US recorded vaccinating over 1 million people. (I'm not sure what the lag in this data is.)

And we are now at a 7-day average of almost 800k doses, compared with just 400k a week ago.  It does seem like we are getting to where we need to be here.

This is likely source for US injections.

https://covid.cdc.gov/covid-data-tracker/#vaccinations

The notes say that those performing the injections have 72 hours to report injections to state and local officials, and there may be further delays between that receipt and passing it along to the CDC.

If I were someone performing injections, I would be wanting to report daily after the close of business. It seems unduly difficult to be reporting on Thursday what you did on Monday. On the other hand, I can see the person making the report on the next day. If the clinic, etc. closes at 4 PM or 5 PM, etc. are they really going to have someone entering the records overnight? They will be doing it the next AM.

The CDC data is as of 6 AM. So the CDC data for 1 million+ doses is from reports made between 6 AM Thursday and 6 AM Friday. The CDC is not issuing their updates until late in the afternoon.

I doubt that many states reported at 5 AM on Friday, more likely at 4 PM or 5 PM on Thursday (local time). This would give them as much time to input data from clinics.

There was an inflection last Tuesday (see ourworldindata chart, zap out Israel, UAE, etc. to expand vertical scale).

If we assume that was for Sunday injections reported on Monday to state authorities and forwarded to CDC late on Monday, then there is a two-day lag, in part because the CDC report is so early in the morning (originally the CDC report was based on 9 AM and they probably became annoyed at east coast states dumping data on them just before the deadline).

So this means that 1 million plus happened on Wednesday. It might not be possible to confirm this since the CDC is only reporting MTWTF (originally it was MWF) and this is the first week that they have done 5-per-week.

The next report on Monday would be for Thursday-Friday-Saturday injections. Saturday will be less than Thursday-Friday, but greater than Sunday. So the Monday report may be just shy of three million.

The CDC has just begin differentiating between
(1) Doses injected.
(2) Persons injected with 1 or more doses.
(3) Persons injected with 2 doses.

About 1/4 of the Wednesday(?) doses were second doses, with the total up to 1.6M. Those who have had only one dose is 9.0 million.

Type of does is 58% Pfizer-42% Moderna. Moderna should lag a bit because they were approved later, and the period between the two doses is 28 for Moderna vs. 21 Pfizer.

I suspect that at least initially, most second doses will happen precisely 3 weeks after the first. If you have done one, you are not going to be waiting around 5 weeks before making an appointment, and you may be getting the call back.

There was no report on Monday January 18 because of the holiday, so the report for Tuesday January 19 covered 4 days. Though the snapshot time was 6 AM (EST), the reports have not been issued until much later in the day (evening).

The Tuesday report showed 3.428M injections over 4 days. If there is a two-day lag, this would be consistent with 1M per day, and 500K per day on Sunday. There may also have been a further reporting delay due to state officials in charge of reporting not working on Monday (people actually doing injections probably continued to do so, but at reduced pace).

The Wednesday report showed 818K injections which likely reflects a drop for Monday. The Thursday report showed 1.021M injections.

[jimrtex]

https://twitter.com/politicalmath/status/1353821818449129473?s=20

I've been following these numbers too and have been very confused by them. I don't doubt that some parts of NY are running out of vaccines (Erie County has had to cancel appointments)... I'm wondering if the figures he's citing about the state running out of vaccines might just be ones administered by the state or by counties, and the vaccines that aren't being used are being administered by hospitals or pharmacies.

Here is the daily report on vaccination.

https://covid.cdc.gov/covid-data-tracker/#vaccinations

Note the report is daily as of 6 AM. Hospitals and pharmacies are required to report to the state or other health authorities within 72 hours of administering. I'd guess that in NY, that NYS and NYC would be separate, and there might also be regional authorities.

The CDC report was originally supposed to be MWF. But that then switched to MTWTF, but they now appear to be reporting MTWTFSS. An odd artifact is that the number distributed has not bumped the last two weekends, while administration has increased daily. Maybe the vaccine is not being distributed on weekends because of the requirement for cold storage.

Here is a report from West Virginia that shows the rest of that shown on Twitter. Note the explanation for the 100% administration.

https://dhhr.wv.gov/COVID-19/Pages/default.aspx

It appears that West Virginia is administering their vaccinations at mass events in each county, though Cabell (Huntington) has three scheduled, and that may be true for Kanawha (Charleston) as well. Note the strong weekly pattern with peaks on Wednesday-Thursday each week.

Texas has an app that shows where you can get vaccinations.

https://tdem.maps.arcgis.com/apps/webappviewer/index.html?id=3700a84845c5470cb0dc3ddace5c376b

There are 100s of locations. But if you switch to first dose availability in the upper right most disappear. I think initially they tried to distribute to every pharmacy and clinic, but only were able to do a few 100 for each, and it might have been hit or miss if people found  out about them. I think Texas is switching more to mass events.

There was a case in Houston where a doctor was doing mass vaccinations at a park, when he opened a Moderna vial (10-11 doses per vial, Pfizer is 6 or 7). An opened Moderna vial has a viability of about 6 hours. It was dark, the event had not been well publicized and the computers were not working. The doctor asked if any of the other health professionals or police doing traffic control had been inoculated. They had. So he started calling around trying to find people who qualified in groups 1A and 1B. He found a couple who were 96 and 88 and gave them each a shot. He used all but one shot, and when the last person did not show up, he gave it to his wife.

The next morning he went to work and logged all the shots. He also returned all the unopened vials. The event was not well publicized.

A week later he was fired. A week after that he was charged and this was carried by national media. He was apparently told he should have thrown the unused doses away.

[jimrtex]

I hate to say it but they should be aiming at 4 to 5 million vaccinations a day, at least. Negotiating between 1 million and 1.5 million a day is too slow. The more time you give this virus to be out there, the faster new variants will develop. That's why you need "Shock and Awe" to beat it back as fast as possible.

I'm still confused about whether Biden's goal is 100 million people vaccinated in 100 days or 100 doses (i.e. 50 million people) in 100 days.  

If it's the latter, that is way too slow, and we are already exceeding that number.  The former number should probably be our realistic goal, at least until more vaccines are approved in the US.

I believe we have contracts with Pfizer and Moderna for 200 million doses each to be fulfilled by July in regular increments.  This works out to about 200 million people in 200 days, which should be a large enough portion of the adult population to achieve herd immunity.  

I'm not sure it is possible for Moderna or Pfizer to produce more than they are already doing, and even they could do this, I'm not sure it would make sense to prioritize selling additional doses to the US compared to other parts of the world anyway.

Over the last 22 days, 1.317M/day have been distributed, vs. 0.862N/day administered.

It may be that distribution is being delayed due to cold storage requirements. There is currently a 19-day supply that has been distributed based on the average injection rate over the past 7 days.

On the glass half full version, 11 states have now administered at least a first shot to more persons than the total case load over the past year).

[jimrtex]

Legit good story:

Oregon health workers stuck in snow give other drivers vaccine

There was a similar situation in Houston, where the doctor who administered the vaccine was fired and had been charged with theft.

Opened vials of the Moderna vaccine have a 6-hour shelf life. The vaccination event at a park was badly organized and the computers were not working, so attendance was weak. After it was dark, the doctor offered the vaccine to health care workers and police working security. When that failed, he started calling around to find other people, none of which he knew, including a 96-YO man and his 88-YO wife. He had administered 9 of the 10 remaining doses. When the final person did not show up, he gave the shot to his wife.

He had basically worked from 5 AM to 11 PM that day. He went to work the next day, and as required, logged the doses into the computer.

He was fired a week later, and charged a week after that.

[jimrtex]

The CDC is tracking vaccinations here

https://covid.cdc.gov/covid-data-tracker/#vaccinations

and here

https://covid.cdc.gov/covid-data-tracker/#vaccination-trends

Click on "Data Table for COVID-19 Vaccinations in the United States" midway down the first page to get state-level statistics.

If you look at the cumulative totals on the second page, it appears that there is about a two day lag in reporting.

The daily report is posted in the afternoon/evening but is based on 6 AM, which means that the reports from the state likely come in the day before. If you have to get your reports in by 6 AM, you probably don't set a 5 AM soft deadline and bring in clerks overnight. Hospitals, pharmacies and other administrators are likely to report at the end of the day, or perhaps the next morning after they have a chance to double check records. The state may then spend the day compiling data before sending it on to the CDC sometime in the afternoon.

So if a shot was administered on Thursday, the state would receive the report that evening, or on Friday morning, and pass it on to the CDC Friday afternoon or evening, depending on how long they give hospitals, etc. to report. The CDC then spends the morning and into on the afternoon on Saturday preparing their report.

There is some additional lag for a minority of injections. The Saturday CDC report increased the reported totals for second shots for Wednesday and Tuesday by a little short of five percent and for Monday and Sunday by about 1%. So maybe 90% are being reported in 2 days, with almost all within 4 days.

A strong indicator of state performance in terms of injections per capita, is the number of distributed doses per capita. That appears to benefit states with older populations. States with small populations tend to have older population. The reason they have smaller populations is that people in their 20s left the state and had their children elsewhere, where they have established new lives. In some cases, they return to their home state in retirement. Someone might have left West Virginia for a job in Ohio and Michigan in 1970, and now have sold their homes and returned to West Virginia. These are the top states with respect to doses distributed per capita.

Alaska	33969
District of Columbia	23656
Connecticut	20764
West Virginia	19920
New Hampshire	18953
Maine	18937
Vermont	18602
Hawaii	18575
Oklahoma	18565
South Dakota	18216
Rhode Island	18152
Massachusetts	18101
Pennsylvania	17907
Nebraska	17906
Arkansas	17895
Florida	17655

It is not clear how they are counting doses distributed through the VA, the DOD, and the Indian Health Service. If a state, the VA would be the 8th largest state. The DOD would be 21st. The IHS would be 37th, but would serve relatively large shares of the population in AK, MT, ND, SD, OK, NM, and AZ. where the AIAN population is a considerable share of the total population. The VA population will skew older, while the DOD will be unbalanced. AK has the largest military population as a share of the population.

It may be that they are double-counting, in which case the national total is over-reported by about 3%. But if they are not counting them in the state totals, then state totals are being under-reported. If two pallets of vaccine are loaded on planes, one sent to Elmendorf AFB, and another to ANC, and those sent to Elmendorf AFB are administered at Elmendorf-Richardson, and those sent to ANC are stuck into arms through civilian channels in Anchorage, all are being administered to people counted as being in Alaska.

The lowest states tend to skew toward younger populations.

Wisconsin	15909
Georgia	15899
Washington	15872
Missouri	15667
Utah	15394
Montana	15314
Texas	15182
Idaho	14911
Nevada	13889
South Carolina	13298

South Carolina may show up near the bottom of the doses/100K (42 of 51), but is 6th among the ratio of administered to distributed.

Some of the best states for administration / distribution may be beginning to run into supply limitations. I think the facility that administers the first dose is responsible for administering the second, and tracking the patients. Once that first dose is administered, a 21-day (Pfizer) or 28-day (Moderna) timer starts ticking. The federal government will be distributing second doses when these come due, and the facilities will be calling the patients with reminders or making appointments.

Right now, the lag between first doses and second doses is around 27 days. This is about what would be expected. Pfizer had a modest dominance, so the minimal possible delay would be 24 days (average of 21 and 28 is 24.5, so I dropped 0.5 to account for Pfizer dominance). But not all persons will get the second shot precisely on schedule, and some (few?) won't get the second shot.

Right now there are about 21 million people who will need a second dose in the next 21-28 days, or close to 900,000 per day. The maximum administration so far has been around 1.4 million per day, so most will be going for second doses. Any extra will be used for first doses, but this is starting to get tight in some states.

For North Dakota if you take the number of doses in stock (distributed but not administered), and the number who will need a second dose each day over the next 24.5 days, the supply is down to 5.7 days. Of course those second doses won't be distributed evenly, there will be more towards the end of the month, and there will be more doses distributed. But there is a tightening of the variance among states for the first dose.

Relative standard deviation for first dose is 14%, for second dose is 31%.

[jimrtex]

https://twitter.com/isaacstanbecker/status/1359184376563720193

Thank you President Biden!!

This is likely irrelevant.

Second doses are lagging first doses by between 24 and 27 days, just what you would expect given that the minimum delay is 21 or 28 days. Some persons won't get the second dose precisely 21 or 28 days after the first, but that doesn't mean that they won't.

The only reason it would change would be if the Biden administration if falling down on record keeping.

[jimrtex]

No more red on the charts!

Winning!

Thank you President Biden!

Thank you President Trump and Governors Burgum, Dunleavy, Justice, Lujan Grisham, and Noem!!!

Most of the recent increase is due to second doses being administered, and is due to planning by President Trump administration to not exhaust supplies on first doses.

[jimrtex]

US life expectancy dropped a full year in first half of 2020, according to CDC

Life expectancy in the US dropped a full year in the first half of 2020, according to a report published Thursday by the US Centers for Disease Control and Prevention's National Center for Health Statistics. Experts say that Covid-19 was a significant factor contributing to the decline.

The life expectancy for the entire US population fell to 77.8 years, similar to what it was in 2006, CDC data shows.

Changes to life expectancy also widened racial and ethnic inequities. Compared to 2019, life expectancy for non-Hispanic Black people in the US fell about three times what it did for non-Hispanic White people, by 2.7 years. It fell by twice as much for Hispanic people, by 1.9 years.

https://www.cnn.com/2021/02/18/health/life-expectancy-fell-pandemic/index.html

I read this article, and I can’t understand where this calculation is coming from. 

Covid has killed about 0.15% of Americans.  If the deaths were equally spread across all age groups, whis would result in an overall decline in life expectancy of about 0.05 years.  And since the deaths were heavily weighted toward older people, this number is actually much, much lower. 

The article says it also accounts for things like drug overdoses and alcoholism as a result of the shut downs, but these would have to be like 100x more deadly than the actual virus for the total to make sense.

The actual CDC articles is here:

https://www.cdc.gov/nchs/data/vsrr/VSRR10-508.pdf

The key is the Technical Notes on Page 7, column 3. Notice how the age-specific death rate is calculated: Deaths (for age group)/Population(for age group) * 2. The multiplication by two is to account for the deaths occurring between January and June 2021 being annualized.

The age-specific death rate is propagated (e.g. L₀ = 100,000 age 0, L_{1 =} L₀ (1-r₀) will be living at age one;
L₂ = L₁ (1-r₁); etc.

You can calculate the total number of years lived by any cohort and divide by the initial number to calculate the average (mean) lived. This is the life expectancy they are using.

(Bad) Assumptions being made: Those born in 2020 will experience the death rate for 75 YO in the first half of 2020, in 2095 when they themselves are 2095, and then when they are 76 in 2096 they will replicate the experience of those who were 76 in 2020, etc. and this will actually start around age 55 to 60 when people begin to die in noticeable numbers.

If you keep applying an elevated death rate year after year, it begins to add up.

I don't know how life expectancy is ordinarily calculated. I would assume that you would use age-specific death rates over several years (10?) and average them, or even better to statistically weight them to reject outliers.

[jimrtex]

Georgia is opening vaccine eligibility to all adults (age 16 and up) beginning on Thursday.

So is Texas on Monday

Alaska made sense.  But Georgia and Texas are both rank way at the bottom in terms of vaccination progress (Georgia is dead last among all states).  I’m confused as to how they can be among the first states to completely open eligibility.

Texas is at 71.3% of over 65 receiving at least one dose, which is above the national average, and actually above Alaska.

Texas is a younger state. If someone moved to Texas in 1990 at age 25, they are now 55 and have children in 20s to 30s, and some grandchildren. The parents they left behind are now in their 70s.

When I signed up, I was just placed on a wait list with no idea when I would come up. About a week before the first shot, they gave me an opportunity to make an appointment, choosing a location and an hour. The hours were somewhat picked over.

My doctor's medical practice announced in January that they would be getting vaccines - but if you could get it elsewhere, to do so (i.e. they would not have enough for all their patients). It appears that the supply is now broadening out (Texas has a tracking site for available locations). So more pharmacies and doctors offices may have a small supply.

They can easily put those over 65 at the front of the wait list, but increasingly those under 65 will be able to come up. There is also regional balancing to consider. In some areas of Texas more than 71.3% of over 65 have received one dose.