Coronavirus general discussion and chit-chat

I don’t think that is accurate, the current average daily rate (7 day moving average) is somewhere around 200-300 for the USA.

I’m seeing a little over 500/day and rising in the US.

(Seven day moving average)

The source I got that from was a week out of date, it was around 300 or so last week and has ticked upwards since then (as your chart shows).

Still not the 2000 suggested in the message I responded to.

Incidentally I see that the chart you quote plots different countries on the same scale and the UK figures on there are for deaths “within 28 days of a positive test” and as such aren’t an accurate representation of deaths caused by Covid. I wonder if all other countries on there use that same definition because if not it is useless for making comparisons.

No one has a source of data that is perfectly consistent. He lists his source (maybe John’s Hopkins?) but each country reports differently. In fact, each US state reports differently.

I have figured out the discrepancy.

@John_DiFool was going off of @Morgyn 's May 4th post in the Breaking News thread. Like Snowboarder Bo before, Morgyn was simply relaying Worldometers running count for World and U.S. cases and deaths. Morgyn would take today’s figures, subtract yesterday’s, and arrive at a daily count.

Straightforward, right? Counterintuitively, not so. In brief: comparing counts day to day leads to overcounts, most notably for deaths.

The issue has come up in the Breaking News thread a few times – I’ve posted with Snowboarder Bo and Morgyn about it, and some other posters as well. We’d noticed that the daily counts in Worldometers’ graphs don’t match the daily counts derived from the cumulative counts.

At first, we just chalked it up to “Meh, different sources, or artifacts of timing, etc.” But that answer never quite satisfied because sometimes the discrepancies were quite large.

When Morgyn posted the today-minus-yesterday U.S. death count for May 4, 2022, it was posted as 1.932 deaths. However, on the same exact Worldometers webpage … you can scroll down to the Daily Deaths graph and see that a count of 448 was given on May 3 and 391 on May 4. So what gives?

The solution finally clicked into place from some COVID number tracking I’ve been doing on a different forum. Starting in mid-February, I’d been posting info from the Worldometers graphs, 7-day average of case counts and death counts every Monday. And kept the past numbers in the posts – for instance, I made one of those posts today and here are the 7-day average U.S. death count numbers as they stand today – based completely on the graphs, not on today-minus-yesterday:

Thu 02/17/2022 - 2,184
Tue 02/22/2022 - 1,938
Mon 02/28/2022 - 1,745
Mon 03/07/2022 - 1,295
Mon 03/14/2022 - 1,120
Tue 03/22/2022 - 805
Mon 03/28/2022 - 662
Mon 04/04/2022 - 592
Mon 04/11/2022 - 505
Mon 04/18/2022 - 383
Mon 04/25/2022 - 372
Mon 05/02/2022 - 348
Mon 05/09/2022 - 289

So – what’s going on? In fact, Worldometers cumulative count is based on when a case or death is reported (or makes it into Worldometers calculations). The cumulative count DOES NOT reflect when a case was first noted or when a death occured.

Back to that 1,932 number for U.S. COVID deaths on May 4, 2022. That didn’t mean that 1,932 Americans died of COVID on that day. Rather, it means something more like so:

  • ~400 died in the 24 hours prior to the point in time the May 4th cumulative number was pulled.
  • some number previously uncounted died on May 3rd
  • some number previously uncounted died on May 2nd
  • some number previously uncounted died on dozens of separate days in the recent past

How did I hit on this? After posting the weekly numbers from the graphs for a while, I started noticing that previous week’s numbers were changing when I went back to them.

For instance, when I posted the U.S. 7-day average death count for Monday May 2 from the Worldometers graph, it was 318. But look above – today I pulled the same figure for the same date from the same source – now it’s 348. About 210 deaths allocated to the 7 days ending on May 2.

And I’ve noticed that even on some fairly old data in the graphs. Individual days from summer and fall 2021 even change a bit here and there. Fairly large changes in the graph’s numbers are common for any date within the last two or three months.

So … the cumulative U.S. case and death counts on Worldometers are actually correct – they just can’t be used to accurately derive just WHEN the counted cases were detected or when the deaths occurred. For that, I rely on the Worldometers graphs and consider them as close to sacrosanct as we’re going to get.

You can look at either deaths by death date, or deaths by report date. The problem with “death date” is that values for recent dates will always be low, as not all of yesterday’s deaths will have been reported for quite some time. The problem with report date is that it’s lumpy. Sometimes it’s even negative, as data gets corrected or adjusted. Other times you get a bunch of old deaths reported in a bunch.

I tend to look at report data when I want recent trends, and average it over a week to smooth out the lumps a little. It’s not perfect, but it gives some sense of what is happening now. If you want to have a better idea of history, it’s better to look at death date.

Something else to consider when comparing old numbers to new:

Detroit Free Press from March 28: “Another change is that the (Michigan) health department will no longer report how many COVID-19 deaths were discovered during a review of vital records.”

(Maybe they are reporting them, but not breaking them out - it’s difficult to tell from the article.)

My vague memory was that the “reviewed” numbers earlier this year added about 10% to the daily count.

I think that from here on out, we’re not going to be able to get accurate numbers of positive tests and thus, of cases. Hospitalizations, sort of…if someone’s hospitalization is revealed to be because of COVID. Same with deaths. I don’t know how I feel about this…I like numbers. They tell us where we are.

I’ve had a nasty cold for about 6-7 days. I was miserable for two or three nights, and couldn’t sleep. And was exhausted during the day. All of us have it/have had it.

We’ve taken several rapid tests, but they keep coming up negative, so, just nasty cold, not repeat Covid.

Oddly, when I was looking up whether the newest sub-variant was evading testing, I came across an article differentiating cold symptoms and Covid symptoms. It said it’s unlikely to be Covid if you were sneezing and had a runny nose.

Does omicron not have those symptoms? When I had Covid (presumably delta, based on timing) I had a couple of days of intense sneezing episodes, and my nose ran like a faucet.

I’m just wondering if they’re way behind on the symptoms, or if omicron has different symptoms than delta had.

My son has covid. His rapid test yesterday was so positive that the test line turned red before fluid reached the control line. He has a runny nose.

I don’t think he’s mentioned sneezing.

In laws are home. Finally tested negative and flew home. No reported lasting conditions.

Equivocal science reporting from a scientist at the Africa Health Research Institute and from National Geographic:

New versions of Omicron (BA.4 & BA.5 - b) are again causing a surge of COVID-19 cases in South Africa, and studies show that these new subvariants are so different from the original version of Omicron that immunity generated from a previous infection may not provide much protection.

If you were unvaccinated (AND had a previous infection? Unclear - b), what you got is almost no immunity to BA.4 and BA.5,” says Alex Sigal, a virologist at the Africa Health Research Institute and at the University of KwaZulu-Natal. “There might be some immunity that may be enough to protect against severe disease, but not sufficient to protect against symptomatic infection.”

BA.4 and BA.5 dodge previous immunity

In the first study of BA4 and BA.5 on immunity, which has not yet been peer reviewed, scientists led by Sigal, of the Africa Health Research Institute, isolated live viruses from nasal swabs (in vitro? - b). The scientists then ran tests to see whether antibodies from unvaccinated and vaccinated people who had been infected with the original Omicron BA.1 strain just a few months ago were able to neutralize these new variants. Sigal’s team discovered that these antibodies weren’t able to protect against symptomatic infection.

The South African study does have some good news for vaccinated people: “We found that you get a lot of protection from vaccines, even if you got infected with Omicron despite being vaccinated—a lot more protection than if you weren’t vaccinated going forward,” says Sigal.

Researchers and the science media really, really, REALLY have to be super-specific about what they mean by " protect " and " protection ". Without hard, painfully precise definitions for "protect(ion) ", the paragraphs above make no sense at all when considered altogether as a piece.

What am I failing to parse here?

My read is they are saying previous infection is irrelevant when considering these new variants. No previous infection of any variant will offer any protection against them.

Vaccinations still offer strong protection against these new variants, or at least serious illness or death.

So only your vaccination status matters. A country like the US where half of our covid immunity is from previous infection instead of vaccination would be expected to be hit hard by these new variants, similar to how omicron went.

That’s my take.

EDIT: Rereading the final paragraph about the “good news”, it’s saying these new omicron strains are less vaccine-evading than the original omicron was.

In other words, vaccinated people now are in better position than just before the original omicron wave hit, while unvaccinated people are in worse position regardless how many times and with what variants they were previously infected with.

If there are any people who are still unvaccinated and have never had any covid, Lord help them I guess. (Actually, the whole point here is that they would be in the exact same boat as someone who had been previously infected with the original strain, delta, omicron, or even all three. None of those antibodies help against these specific new strains.)

If I’m reading this right, and all this is true, that’s some quality evolution right there.

That’s incredibly implausible, though. So my guess is they are just muddled about protection from severe disease vs. protection against symptomatic disease.

I’m not connecting those dots. Little help?

It’s incredibly implausible that a vaccine based on original-strain covid would provide more immunity to a new variant than an actual infection, especially an actual infection with a slightly more similar strain.

Seems reasonable to me. I don’t see anything implausible about a new variant completely rewriting much of its features to cloak it from previous immunity, but it still uses the same spike protein.

(I say “rewriting itself” only as short hand, not meant literally.)