Post #1176: COVID-19 trend to 7-14-2021

 

The sporadic nature of most states’ data reporting is adding some variation to the estimate of the trend.  The estimated increase in new cases for the past seven days is only 74% today, not the 92% that came out of yesterday’s calculation.

That said, there have been exactly two instances in this pandemic when the seven-day growth in U.S. new COVID-19 cases exceeded 74%.  Those were the seven-day periods ending yesterday, and the day before yesterday.

Just to say it plainly, for the U.S. as a whole, this rate of growth in new cases is unlike anything we’ve seen so far in the pandemic.

Continue reading Post #1176: COVID-19 trend to 7-14-2021

Post #1175: COVID-19 trend to 7-13-2021

 

New COVID-19 cases are up 92% compared to one week ago.  That increase was so hard to believe that I took the raw data and calculated the rates for a few states by hand.  Just to double-check.

What put it over the top, compared to yesterday, is that Florida finally reported fresh data, coming in with a seven-day moving average of over 27 new cases / 100,000 / day.  That’s higher than Missouri, which was, at one time, the epicenter of the U.S. fifth wave.

It really says something, I think, that we’re still setting new records more than a year into the pandemic.  I checked the data all the way back to April 2020.  That 92% increase is the largest one-week percentage change in new U.S. cases since the start of the pandemic.

And yet, we are still sleepwalking into this U.S. fifth wave of COVID-19.

Source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 7/14/2021, from https://github.com/nytimes/covid-19-data.”  The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

Continue reading Post #1175: COVID-19 trend to 7-13-2021

Post #1174: COVID-19 trend to 7/12/2021: Going vertical

 

On the plus side, most of the U.S. still has a low absolute number of new cases per day.

On the minus side, the increase in new cases is steep.

The most important thing to realize is that the infections shown at the end of the graph below occurred about 16 days ago, on average.  It takes about 12 days for the sequence of infection, symptoms, care-seeking, testing, and reporting to work out.  And then, this is a seven-day moving average of the reported data, so the numbers are, in effect, an average of another 3.5 days old on top of that.  The point is, you should add your best guess for 16 days’ additional spread in the community, beyond what you see here, as you consider any changes in your behavior in response to the fifth (Delta) wave of COVID-19 in the U.S.

Continue reading Post #1174: COVID-19 trend to 7/12/2021: Going vertical

Post #1172: U.S. COVID-19 fifth wave takes off.

 

The average number of new COVID-19 cases per day remains low for the U.S. as a whole.  That said, the trend is up, and it’s now showing a fairly rapid rate of growth.  Daily new COVID-19 cases rose 22 percent over the past seven days, as shown on the log-scale graph below.  In hindsight, we can date the end of the U.S. fourth wave and the start of the fifth wave to 6/20/2021.

Data source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 7/9/2021, from https://github.com/nytimes/covid-19-data.”  The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

This fifth U.S. is widely attributed to the emergence of the Delta (formerly Indian) variant of COVID-19.  And yet, except for the outbreak in southwestern Missouri, it’s hard to see that in the data.  Here are the top eight states, in terms of incidence of Delta variant, according to the most recent data from the CDC:

That said, this graph may be misleading because it looks at the at the number of new cases per day, not the growth trend.  All we can tell from this graph is that the only state where there’s a notable outbreak of COVID, right now, is Missouri. 

By contrast, if we graph that with a log scale, below — so that the slope of the line is now the growth rate — we can see that almost all of those lines have turned upward, and several have some fairly steep slopes.

If we take all 20 states where CDC has estimated the prevalence of Delta, and plot against the last month’s increase in cases, it does show a fairly large and statistically significant positive correlation.

The upshot is that, while the picture isn’t crystal clear, and while there’s really only one area seeing crisis levels of hospital admissions from Delta, there is some modest evidence that the uptick in cases across states is systematically related to the level of Delta in a community.

Possibly the picture is unclear because we still don’t have good estimates of the incidence of Delta.  The CDC data on which the graph above is based typically reflect just a few hundred COVID-19 samples per state.

Possibly the picture is being muddled by seasonality.  We had an outbreak in 2020, at this time of year, when none of these new variants were around.  That was a wave centered on the hot-climate, high-air-conditioning-use states.

We should not fall into the trap of thinking that there can be only one cause.  It’s entirely possible that the same seasonality that gave us the second wave of COVID-19 is once again at work.  And so what we’re seeing could be a mix of that seasonality and the impact of the more-infectious Delta variant.

In fact, when I correlate the recent trend with typical summertime temperature by state (my proxy for typical time spent in air conditioned spaces), I find a fairly nice-looking graph.  And when I put both factors into a regression (not shown), the case for seasonality as a cause of the current uptick is just as strong as the case for the spread of the Delta variant.

Finally, it almost goes without saying that we owe this wave of the pandemic to people who refuse to get vaccinated.  While vaccination is not the only way to acquire immunity, it’s the surest.  And so, if we plot recent month’s trend against the fraction of the population that has not been vaccinated, we get the expected result.  The greater the fraction not fully vaccinated, the higher the growth in new cases.  By eye, this factor is at least as strong as the other two.

Well, no matter how you slice it, case counts are rising again.  So far, there’s only one small region with crisis levels of hospitalizations.  Outside of that region, you could make equally strong cases that this latest rise is driven by the Delta variant, or that it’s just seasonality — a repeat of last year’s second wave.  Plausibly, the answer is that it’s a combination of the two.  Whatever the reason, we can chalk this fifth wave up to the significant portion of the population that refused to get vaccinated.

Post #1171, COVID-19 trend to 7-6-2021, Delta variant?

 

Data on new COVID-19 cases is now being reported for the majority of states again, following the July 4th holiday.  The only high-population state still missing is Florida.  As I explained in an earlier post, when that happens, I gap-fill with the seven-day moving average based on the most recent data on which the state reported some data.

Data source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 7/7/2021, from https://github.com/nytimes/covid-19-data.”  The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

The little notch at the end of the curve is the effect of July 4th on reported positive cases.  As we’ve seen in the past, that’s likely a combination of slow data reporting due to the holiday (which is then recouped later), and some permanent reduction in new positive cases, presumably due to people with marginal symptoms not bothering to get tested over the holiday.

Until the data settle further, I think the only important takeaway there is that we’ve returned to an upward trend.  But so far, not much of an upward trend.  How rapidly new cases are actually rising at this point is hard to say, given all the factors that are perturbing the data (as discussed in my last post).

Good news, Delta is already here!

And not much is happening (yet), other than in southwest Missouri.

The CDC has updated its estimate of the incidence of the Delta variant, and it’s right on track with what I reported out yesterday.  As of the two weeks ending 7/3/2021, Delta accounted for more than half of new cases.  Given the less-than-two-week doubling time, Delta plausibly now accounts for two-thirds of new U.S. cases, as I reported yesterday based on the Scripps analysis.

Source:  CDC

By region, Delta has plausibly saturated the market in the Mountain and Central regions.  For the US Health and Human Services (HHS) Mountain and Central regions, Delta already accounted for more than 75 percent of new cases as of the two weeks ending 7/3/2021.  Currently, Delta must account for nearly all new cases in those regions.  Those regions have already hit a Delta incidence rate similar to that in Great Britain.

Source:  CDC

And yet, if I simply plot the trends, you can see some possible impact of Delta, but not an overwhelming or consistent impact.  Region 7 — where the southwest Missouri outbreak is occurring — has a distinct upward trend.  Region 8 — where Delta is nearly as prevalent — does not.

Good news, Missouri is the new Michigan.

Similarly, CDC published an updated listing of the prevalence of Delta by state.  These numbers are, on average, close to a month out-of-date.  Given the doubling time, that means that in all likelihood, most of the states at the top of the listing now have an incidence of Delta in the 90 percents, similar to what is now seen in Great Britain.

And yet, if I simply plot the trends for the top states on that list, to look for a spike in cases, I find Missouri.  And only Missouri.

The numbers suggest we ought to have a fairly significant U.S. spike from Delta.  It’s just that much more contagious.

But instead, this is shaping up to be weirdly similar to the U.S. fourth wave, driven by the Alpha (U.K.) variant.  We don’t have any nationally-uniform impact.  What we have is some apparently mild and sporadic impact of the Delta variant on daily new cases.  And, as with the Alpha (U.K.) variant, we have one locale with a fairly significant outbreak.  For Alpha, that centered on Michigan and surrounding states.  For Delta, looks like that’s centering on southwest Missouri and surrounding states.

Basically, so far so good.

It’s worth restating that we owe this most recent round of COVID-19 to the people who won’t get vaccinated, full stop. In the past, we could try to place blame on any number of possible bad actors, from bad advice from the CDC to bad planning on the part of the Federal government, to people who didn’t have the good sense to wear a mask during a deadly pandemic.  But no longer.  We have an effective vaccine.  We just have a large segment of the population that’s not smart enough, or cooperative enough, or caring enough to take it.

I still say that the numbers and the theory still suggest we ought to be having a larger outbreak.  The Delta variant is estimated to be that much more contagious.  So, maybe it’s just a matter of time.  On the other hand, maybe we have enough people immune via prior infection or vaccination that this fifth (Delta-driven) wave will never be more than a mild annoyance for most of the country.

That said, the bottom line is that there’s no rational reason to have had this fifth wave at all.

Afterword on separating fact from fiction.

As you can tell from reading this blog, I’m very much a science-based reasoner.  I did health services research of one sort or another for most of my career, so I’m fairly well armed in the battle to sort information from nonsense.

That said, I had another health-related question tossed at me this morning, and it really highlighted how straightforward it should be for the average person to do the same thing.

Here’s the giveaway:  Reality tends to be constant, simple, and internally consistent.  But bullshit comes in infinite variety and shifts almost daily from one inconsistent story to another.

I believe that’s because evolution applies as much to memes as to biology.  In effect, our social media keep trying out different varieties of bullshit to see if anything will stick.   Some of it — the bullshit with the perfectly crafted back-stories, and the elaborate and memorable details — is plausibly disinformation.  That is, stories deliberately created by adversaries to try to mess up Western civilization.  But you don’t need that hypothesis to explain what you see.  All you need is the notion of “stickiness”.  Some stories simply resonate with some people.

And so, much like the virus itself, the bullshit evolves to become ever-more-effective based not on some grand plan, but based on whatever succeeds.  Whatever sticks.

And, as air travel is to the spread of the more effective virus, social media is to the spread of the more effective bullshit.  Instead of containing a more virulent story to one area, any piece of bullshit that resonates with people gets rapidly splattered across the globe.  A virus or story that might otherwise have been isolated to a small group rapidly finds all the individuals who might potentially be infected with it.

But the easy way to spot the bullshit is that it does, in fact, change and evolve.

The real story is simple.  This is a somewhat deadly disease, the vaccines available for it are pretty good at preventing infection, and they are downright excellent at preventing the worst outcomes (hospitalization and death).  Near as I can tell, nobody in any mainstream public health organization anywhere in the world has said any differently.  Some details have changed — airborne spread is one, spread via fomites (inanimate contaminated objects) is another.  But the main story line has been consistent.

By contrast, if you stop and look back at it, the sheer variety of the bullshit is truly astounding.  Just off the top of my head, with no research:

  • It’s not real, it’s just a hoax by the World Health Organization.
  • It’s a hoax perpetrated by the Chinese to mess up the U.S. economy.
  • It’s a hoax, and if the Democrats win the election, you’ll never hear of it again.
  • It’s real, and was deliberately released by the Chinese to mess up the U.S. economy.
  • It’s real, but it’s no more serious than the flu.
  • It’s real, but if you’re healthy you won’t have any problem with it.
  • People aren’t dying from COVID, they are dying from other conditions.
  • Hospitals are pretending people have COVID to collect more reimbursement.
  • Physicians, ditto.
  • It’s all a conspiracy to prevent the free exercise of religion.
  • Mask requirements are the first step toward dictatorship.
  • Mask requirements are a way for the government to control your lives.
  • The vaccine will alter your DNA.
  • The vaccine contains a microchip designed by Bill Gates.
  • The vaccine will kill you.

I’ve probably missed a few dozen there, but I’m hardly going to take the time to look them up.

You’d think that by the time the average person saw the Nth piece of bullshit, they’d kind of figure out that all of that stuff was nonsense.  But that’s not the way it works.  Instead, each bullet point above has its own separate effect.  Each finds some fresh segment of the population that is happy to use that particular story as their excuse for ignoring mainstream public health advice.

And so, collectively, that process of mutation and spread of mis- and dis-information prolongs the pandemic.  At this point — where the only reason we still have a pandemic is that some people won’t get vaccinated — you almost have to ask which form of evolution has been more harmful.  You have to wonder whether the rapid evolution of all that bullshit has contributed more to the pandemic than the rather slower mutation of the virus itself.

Post #1170: Recent COVID case trend is unknown, Delta variant

I wish I knew what the recent U.S. trend in new COVID-19 cases was.  But the fact is that nobody knows.  Today’s data release, with data through 7/5/2021, had data from fewer than 20 states.  I don’t really think that’s enough for constructing a estimate of the current trend.

The fact is that between the states that no longer report on the weekends, the states that only report sporadically, and the July 5th holiday day, there’s really no current information.

In addition, we can probably expect the July 4th holiday to leave an artifact in the data similar to that seen for the Memorial Day holiday.  Some changes in testing behavior and speed of test processing, in addition to the plain fact that the states didn’t report information on new cases.

The upshot is that it may be well into next week before we get another clear fix on what the underlying trend is.


Delta variant incidence

I stumbled across a non-CDC source for information about the current incidence of the Delta variant.   This site is a Federally-funded project based at the Scripps Research Institute.  Based on their estimates, Delta currently accounts for about two-thirds of U.S. cases.

Source:  Outbreak.info  Julia L. Mullen, Ginger Tsueng, Alaa Abdel Latif, Manar Alkuzweny, Marco Cano, Emily Haag, Jerry Zhou, Mark Zeller, Emory Hufbauer, Nate Matteson, Kristian G. Andersen, Chunlei Wu, Andrew I. Su, Karthik Gangavarapu, Laura D. Hughes, and the Center for Viral Systems Biology outbreak.info. Available online: https://outbreak.info/ (2020)

The website has some technical problems, but the graph syncs up well with the older CDC data.  The CDC data showed that Delta accounted for 26% of cases, as of roughly June 14.  With a doubling time of about two weeks, that could easily result in the roughly 65% of cases shown currently, above.

To me, that’s something of a good news/bad news joke.  The bad news is, Delta is now the dominant COVID-19 variant in the U.S.  The good news is that outside of a few areas, such as southwestern Missouri, that’s not having a huge impact yet.  Case rates were starting to turn upward in a lot of places, but only modestly.

In any case, only 35 more percentage points to go, and it’ll be as bad as it can get.

Interestingly, the Delta-driven outbreak in Great Britain is continuing apace, and yet Britain is not postponing plans for removing their remaining COVID-19 restrictions.  Almost all new cases in Britain are the Delta variant.

The current rate of new cases in Great Britain is about 25,000 per day.  That would be equivalent to about 150,000 per day in the U.S., or more than ten times the current rate.

The mere presence of the Delta variant does not guarantee an outbreak.  Almost two weeks ago, Germany passed the point where half of new cases are the Delta variant, and they have seen no uptick so far:

I think the upshot of all of this is that we’ll get there when we get there.  Odds are still in favor of a major uptick in new cases, based on the increased infectiousness of Delta.  We were starting to see a broad upward movement in new case counts last week, but it didn’t appear to be centered on the states that the CDC said had the highest incidence of the Delta variant.

Maybe by the end of the week the picture will be clearer.  The US CDC is due to update its state-level estimates some time this week.  Presumably, most states will have reported current counts of new cases.  I’ll reassess at that time.

Post #1166: The U.S. fifth wave gathers speed, driven more by seasonality than by the Delta variant.

 

It surely looks like we’re well into the U.S. fifth wave of COVID-19.  Here’s the regional graph looks as of 7/2/2021.  New cases were up 18 percent in the last seven days.  That’s driven in part by a fairly large spike in cases reported by Florida.

Source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 7/3/2021, from https://github.com/nytimes/covid-19-data.”  The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

For sure, we can thank unvaccinated Americans for this wave.  But beyond that, it’s not exactly clear what’s driving the bulk of it, because we have two things going on in the U.S.:  Spread of the Delta variant, and summertime air conditioning season. Continue reading Post #1166: The U.S. fifth wave gathers speed, driven more by seasonality than by the Delta variant.

Post #1165: Upward trend in COVID-19 cases is now quite clear.

 

Changes in state data reporting have made it increasingly difficult to get a good estimate of the most recent trend in COVID-19 cases.  Starting today, I’ve made some imputations — described in the last section below — to account for the worst of the state data issues.

With those in place, the upturn in daily new COVID-19 cases in the U.S. is now quite clear.  As expected (Post 1160, June 15 2021), we now seem to be tracking along the same path that Great Britain took following the spread of the Delta variant. Continue reading Post #1165: Upward trend in COVID-19 cases is now quite clear.

Post #1164: COVID-19 trends, U.S. new case count begins to rise

 

The count of daily new COVID-19 cases in the U.S. has begun to rise.  This rise in new cases is an expected consequence of the spread of the new Delta COVID-19 variant (Post #1162, 6/23/2021).  It might also be helped along a bit by the return to living in indoor air-conditioned spaces in the southern U.S.

Here’s the regional/national graph, in logs (and so, the slope of the line reflects the rate of growth).  I’ve put Florida back into the analysis because they appear to be reporting data five days a week, which is adequate for this purpose.

Source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 6/28/2021, from https://github.com/nytimes/covid-19-data.”  The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.

As you can see, the upward movement is fairly broad-based.  The only region where cases declined in the past week was the Northeast.

Here’s the bigger picture, showing the five U.S. waves of COVID-19.

The initial wave of COVID-19 was largely confined to the New York City region and parts of the West Coast.

After that, starting at the end of June 2020, we had a summertime wave in hot-climate states, attributed to time spent in indoor air-conditioned spaces.  We may yet have a repeat of that, but it will be commingled with the impact of the Delta variant.

The third wave was the wintertime wave, with the high peak attributed to the natural seasonality of coronavirus.  As with (e.g.) flu, this virus appears to spread best in the winter months.

The fourth wave was due to the spread of the U.K. (now Alpha) variant.  Alpha was more infectious than the variants it replaced.  This wave was suppressed by a surge in vaccinations, with daily vaccination rates peaking in mid-April.  There, the numbers suggested that the rate of vaccination (plus existing post-infection) immunity should have been enough to suppress the surge due to the Alpha variant.

And now we’re in a fifth wave, due to an even-more-infectious Delta variant.  This time, however, new vaccinations have slowed, the vaccine is less effective against this variant.  The upshot is that we should not expect this wave to be terminated by the ongoing vaccinations.  This one is likely to stop when enough of the remaining un-vaccinated individuals have acquired immunity via infection.

It has reached the point where the CDC estimates of the Delta variant, by region and state, are now positively correlated with growth in new cases. 

For the 10 HHS regions, it looks like this.

And by state, it looks like this.  The state analysis is incomplete (CDC doesn’t have data for every state), and the data for that are quite old.

Based on the slopes of the linear trend lines, the state data appear to show a vastly higher impact of Delta on case growth, compared to the regional data. That’s an artifact.  The state-level estimates of Delta as a percent of cases are roughly six weeks old.  That’s about three doubling times, or a factor of eight or so.  Because the Delta-as-percent-of-cases is about 8 times too low, the trend line compensates for that by giving the line a slope that is roughly eight times higher.  So, in fact, despite the vastly different numbers, the slopes of the lines agree fairly well.  The difference in the numbers is due to the different ages of the data.

The bottom line of all of this is that we shouldn’t expect this most recent little uptick to be some short-lived fluke.  To the contrary, we should expect the Delta variant to run through the remaining un-vaccinated (or otherwise non-immune) population until something in excess of 90% of the population has immunity to it.  Given the current slow rate of vaccination (well under one million per day now), that’s probably going to occur through lots of infections within the un-vaccinated population.

That conclusion is based on the following:

  • The situation in Great Britain.  They have a total vaccination rate slightly higher than ours, and they remain in partial lockdown.  But they are a few weeks ahead of us on the spread of Delta, and new cases continue to rise there.  On a per-capita basis, Great Britain now has almost six times as many daily new COVID-19 cases as the U.S. does.  All other things equal, we should expect to see some like that level of increase.
  • The numbers, that is, the estimated “r-nought” or natural infectiousness of this new variant.  The r-nought for Delta is high enough (estimated as 5 or higher) that, practically speaking, the U.S. will never be able to achieve herd immunity via vaccination.  We’d need more than 90% of the population vaccinated.
  • The cross-sectional variation in case growth.  Missouri has by far the greatest proportion of new cases being the Delta variant, and Missouri has an ongoing surge.  Physicians in that state attribute the surge to the Delta variant.  And it’s just a matter of time before the rest of the states follow that path.  It now appears that process has started, based on the positive correlation between the incidence of the Delta variant and the growth in new cases.