Post #1238: COVID-19 trend to 9/7/2021, pardon my artifact.

 

The Labor Day holiday left a big reporting “artifact” in the data.  The numbers nose-dive, purely due to the extra day of no or minimal case reporting.

But if I crudely remove that, it sure looks like the Delta wave has peaked, for the U.S. as a whole.

Best to wait a few days, and let any true impact of the holidays work through the data.  For example, right now, it’s obvious that pretty much the entire Midwest under-reported their cases a bit, based on the shapes of the lines.  So this might spring up a bit over the next few days.  That said, I’m not seeing any remaining artifacts that look big enough eliminate what appears to be a 9/1/2021 U.S. peak for this wave. Continue reading Post #1238: COVID-19 trend to 9/7/2021, pardon my artifact.

Post #1237: The COVID-19 situation at William and Mary

I’m writing this post to provide a little perspective on the COVID-19 situation at William and Mary.  Yesterday they updated their dashboard to reveal that there have been 115 known cases on campus.

To reiterate what the William and Mary administration said about that, almost all the cases are mild or asymptomatic breakthrough infections.  (That they are almost all breakthrough is unsurprising, given that 93 percent of students report that they have been vaccinated.)

My main takeaways are that:

  • Many Virginia universities and colleges are having a problem with COVID-19 in the 2021 fall semester.
  • The rate of new cases showing up in Virginia universities seems roughly similar to the Virginia 18-24 population as a whole.
  • William and Mary appears to have a much-worse-than-average problem this semester, based on a comparison to four other Virginia universities and to the current rate among all Virginia residents age 18 to 24.

How does William and Mary compare to other Virginia universities? 

In a situation like this, it’s sometimes helpful to look around at the rest of the state (as in Post #788, last year).  Below I’ve quickly tabulated the numbers for a handful of Virginia universities.

This is a rough cut, in that some of these combine students and faculty, and some of the enrollments my overstate the actual residential enrollment to which the testing applies.  There’s nothing systematic about which universities I chose.  They were just the first ones to come to mind in this context.

The other unknown is the effective start date of the school year.  The longer they’ve been back on campus, presumably the more time they’ve had to identify infected students.  Radford, for example, began classes on 8/23/2021.  U Va started on 8/24/2021.  W&M began classes on 8/19/2021. 9/1/2021.  (I looked at the wrong calendar).  So W&M has had fewer person-days on campus in which to find infected individuals (or to have cases develop).

That said, at first blush, none of the numbers for William and Mary look good.  Every Virginia university is having some sort of a problem, but ours looks somewhat worse than the comparison group. A combination of fewer tests and more cases particularly stands out.

For the high level of active cases, that must be due to the fact that these are almost all newly-reported tests.  Which is consistent with the late date for start-of-classes.  By definition, I think cases are considered active for 14 days following positive diagnosis, unless symptoms persist beyond that point.  In other words, if almost all of those tests took place less than 14 days ago, then by definition, almost all of the cases will be considered “active”.

Not shown — and to me, by far the oddest aspect of the W&M counts — is that W&M found more than 100 cases by testing fewer than 400 people. It’s on-order-of a one-third positivity rate (infections to persons tested).

I can’t quite grasp how they could have done that.  I surely hope that’s due to testing symptomatic persons only, then following up with contact tracing.  That “target rich” testing pool could plausibly yield such a high positivity rate.  To be crystal clear, it’s not even remotely plausible that the high positivity rate shown reflects a cross-section of the campus population.

In any case, the main takeaway is that everybody is having a problem with COVID-19 in the Fall 2010 semester.  (I was going to say more of a problem than last year, but Radford had a terrible time last year.  Now they have just about the best numbers on the table.)

The plain reading of the numbers makes it look like more of a problem at William and Mary than elsewhere.  But the oddity of the testing numbers (100 positives out of 400 persons) means that William and Mary’s testing has been driven by symptomatic cases (and possibly contact tracing).  But everybody tests symptomatic individuals, so that, by itself, would not explain the higher numbers at William and Mary compared to Virginia peers.

 


What are the rates of new COVID-19 cases in the college-age population of Virginia?

Data source for this graph of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 9/4/2021, 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.

It’s no secret that the overall rate of new cases is higher now than it was last fall. Above, you see the rate of new cases per 100,000 population per day for Virginia, in the first year of the pandemic and the second year of the pandemic.

It’s tough to compare the COVID-19 rates on these campuses to the COVID-19 infection rate in the community.  The problem is that we’re seeing cumulatives of all cases over some period of time, for the college data.  For the community, we’re seeing daily cases.

In short, I have to guess the number of days of exposure that have occurred so far, at these college campuses.  For example, I’d have to use a longer period to make the comparison to the colleges with early start dates, and a shorter period for comparison to colleges with more recent start dates.  Based on class start date for William and Mary (and arrival on campus prior to start of classes), I’m just going to use nice, round three and two week periods.  So for this next bit, I’m combining Census data showing 802,000 Virginia residents age 18-24, and Virginia Department of Health counts of new COVID-19 cases for that 18-24 population for the past three weeks, and then for the past two weeks.

And, for the past three weeks, for the Virginia age 18-24 population:

  • 0.8% were diagnosed with COVID-19.
  • 0.01% were hospitalized for COVID-19.
  • 0.001% died from COVID-19.

For the past two weeks, for the Virginia age 18-24 population:

  • 0.6% were diagnosed with COVID-19.
  • 0.007% were hospitalized for COVID-19.
  • 0.0006% died from COVID-19.

You have to be careful when comparing this to the college numbers because testing regimens may differ.  I would guess that asymptomatic cases are rarely diagnosed in the community.  They would only be found via contact tracing.  Colleges and universities, by contrast, may have a more systematic testing program in place, and so find a larger portion of cases.

That said, the all-Virginia estimate for this young-adult population is close to what we’re seeing in the schools above.  In some sense, that’s reassuring.  It’s not as if going off to college hugely increased risk, relative to staying home.

But William and Mary’s numbers again stand out as higher than expected.  Given the late start date, the William and Mary 1.7% number should be compared to the 0.6% observed in the college-age Virginia population over the past two weeks.

Honestly, I’m still puzzling over the fact that about a third of the students tested at William and Mary were positive.  And that the number of students that tested positive — so far — is running at a rate that’s about three times the community rate.  I’m wondering if they’re getting some false positives, as occurred during one episode last year.

But in another sense, it’s not a huge surprise that the average Virginia university has numbers close to the Virginia average for college-age persons.  To a large degree, we have to be looking at literally the same people.  Virginia colleges have about 500,000 enrolled, excluding one university claiming large virtual enrollment, versus 800,000 in the 18-24 age group.

Post #1236: What mask should I wear?

 

This post is my briefest answer to the question “what mask should I wear?”.  My wife was asked this question yesterday in the context of the high levels of COVID-19 currently circulating at the College of William and Mary.

If you don’t feel like reading this, then my shortest possible answer is “3M Aura masks, available in the paint department at Home Depot”.  Read on for other options.

This is a timely question for me, because I just sent a “mask sampler” to my daughter.  I sent boxes of seven different high-quality masks.  She will try one of each, find what suits her best, and pass the remaining unused masks to friends, who may then do the same.  Faces vary, people vary, and what works best for one person may not work well for another.

To be clear, everything I recommend in this post, I’ve bought for myself or my family.  And I’m answering this question under the assumption that you are as serious about avoiding COVID-19 as I am.

I’m going to make some specific recommendations, first, including options for people with small faces.  Then I’ll get this posted.  And later today I may add information on (e.g.) N95, National Institute for Occupational Safety and Health (NIOSH)-certified, when to throw a mask away, and so on.  Undoubtedly more than you ever wanted to know about masks.

In general, you want a NIOSH-certified N95 respirator (mask).  If you won’t wear one of those, I think your next best bet is a name-brand, made-in-Korea KF94 mask.  In either case, you want something with a snug fit, ideally an air-tight fit, against your face.

If you don’t like how they look, wear a cheap thin loose cloth mask over them.  You can get quality KF94s in black, if that fits your style better.

If you won’t wear either an N95 or name-brand, made-in-Korea KF94, I have no advice for you, other than to suggest that you read the final sections of this post, once they are written.

Surgical masks, even proper ones with both a BFE and a PFE rating, do not work as well as N95 respirators.  And if you buy “procedure masks” — cheap surgical-style masks — you have no idea what you’re getting.

If you insist on wearing those cheap blue procedure masks, at least learn the “tied-and-tucked” method for wearing them.  Read the article in the Journal of the American Medical Association (JAMA), or see one of the authors of that article explain it in this YouTube video.

Otherwise?  “KN95” has no legal meaning in the U.S.  Anything can be sold as a “KN95” mask.  Cloth masks are a total gamble.  Some work almost as well as a surgical mask.  Others don’t.  Double-masking is a gamble, because it increases back-pressure and so increases the likelihood that you breathe around the mask, rather than through it.

During the U.S. N95 shortage, when citizens couldn’t buy an N95 through legitimate retail channels, people had to make do.  Any mask was better than no mask.  But now?  When name-brand N95s are on the shelf at every hardware and drug store in the U.S.?  When you have your pick of sizes and types of fit?  Now you have no excuse not to wear a proper N95.


1: An expedient and low-risk N95 option

Go to the paint department at your local Home Depot and buy a box of 3M 8210 masks or 3M Aura masksClick the links to see pictures of them on the Home Depot website.  (You can also get curbside pickup, or (for a fee, in most places) have them delivered to your house, from your local store.  You could also get either if these from other vendors, via Amazon).

The 3M 8210 is a traditional “cup style” mask.  It’s a soft cup, with a foam-padded flexible metal nosepiece (for fit), and two thick elastic straps to hold it tightly against your face.

The Aura is a “flat fold” mask.  The Aura is probably a little easier on your face, and, in general, it’s just a lighter-weight mask.  It’s more flexible than an 8210, useful if you plan to do a lot of talking.  As with the 8210, it has a bendable metal nosepiece (for fit), and straps that go behind the head to hold it tightly against the face.

Beyond that, it’s all about how well it fits your face.  Note that, formally, these products are respirators, not masks.  A respirator is designed to seal tightly against your face.  Masks are not.  That’s a big advantage for these products over (say) surgical masks.  But if you don’t fit them right — if they leak around the edges — you don’t get the full N95 level of filtration.

Whatever you buy, try it on after you’ve read the NIOSH instructions on how to put one of these on.  While you’re at it, you might want to read what 3M says about wearing one of these properly (.pdf).  E.g., you are not supposed to pinch the metal nosepiece with one hand, to shape it to your face.  Use both hands, press it into shape, and you will get a better fit with less chance of a leak.

There’s a little trick to putting one of these on easily.  You hold it in your palm, let the straps dangle underneath your hand, put your palm to your face, then pull the straps over your head.

I use the 8210 and similar 3M “cup-style” products.  But I’ve been using them for decades (as dust masks), so I’m used to them.  They seal against my face well, and I find that semi-rigid style easier to put on than the flat-fold style.  So I’ve stuck with the product that I know.  The elastic straps start off quite tight, but stretch with use.

That said, 3M advertises the ease-of-fit of the Aura mask, and its ability to fit a wider range of faces.  For this reason, if you’re unsure of which to buy, you’re probably better off with the Aura.  I don’t think it gives quite as good a seal as the 8210 in an ideal case.  But it will work across a broader range of face shapes.  It’s the choice with  a lower risk of failure.

Buying them at Home Depot eliminates your risk of buying counterfeit masks.  Presumably, those masks were shipped from 3M to Home Depot, and that’s a reasonably secure supply chain.  When N95 masks were in short supply, that was a real problem.  (Check out the numbers on the 3M respirator fraud page.)   I’m not sure how much of a risk remains.  But I’m pretty sure that risk is minimal if I buy from a major retailer such as Home Depot.

The only hard thing about any of this is remembering that they are in the paint department, not the tool section.   Home Depot has a separate display of protective equipment, including masks, with the tools.  But the ones that you want are with paint.  Hence the picture.


2:  Other NIOSH-certified N95 masks I have recently bought.

NIOSH-certified respirators come in a variety of types and sizes.  If the first ones you buy don’t work out well for you, there are other options.  Here are four that I have bought — two for my daughter’s mask sampler, one that turns out to be my wife’s preferred mask, and one that fits over a small beard.

If you have a small face, #1:  3M 8110S.  This is a smaller version of the 8210 “cup-style” mask described above.  This is a specialty item, and you’re going to have to buy it over the internet and pay a modest premium for it.  The listings on Amazon keep changing, which is a little unsettling.  That said, it’s such an oddball item that I think risk of counterfeit is low.  I bought these via Amazon, from OZ Medical, which is an actual bricks-and-mortar full-service supplier of medical goods.  That particular listing is no longer up, and I’m not sure how long the current Amazon listing will remain.

If you have a small face, #2:  Patriot cup mask in size small.  This is another NIOSH-certified N95 cup-style mask.  This is made by a U.S. startup, which then has the advantage of being so obscure that nobody could make money counterfeiting their masks.  These are expensive, and with shipping, worked out to be $5 per mask.  I thought it was worth a try because a) it comes in size small, and b) the shape of the mask and seal differs from the 3M product.

A lightweight, breathable N95 option:  Kimberly-Clark duckbill.  This is a lightweight mask with weak straps and a thin metal nosepiece.  Oddly, that combination works well for me and gives an excellent seal around my face.  These masks are also spacious and flexible, which makes them comfortable for wearing for long periods of time.  This is my wife’s preferred mask.  These are also available in size small through specialty medical suppliers on-line.  The only real drawbacks I can see are that they look a bit odd, and because they are thin, they don’t last as many hours as the 3M cup mask above.

If you wear a goatee, you might want to try a Magid mask.  In general, beards prevent N95 masks from sealing.   They are a bad idea during a pandemic.  This was the only N95 I found that seemed to work well with a goatee.  It’s an odd mask with a thin, flexible silicone seal all the way around the circumference of the mask.  And it is quite large.  That combination means I can put it on over my goatee and still have it seal properly.  But be warned:  I find that about one-in-three of these masks will not seal.  You just have to accept the waste as part of it.  This will not fit a person with a small face.


3:  Ear loop masks, the caveats.

Ultimately, the question isn’t “what’s the best mask”.  The question is, “what’s the best mask that a person is willing to wear“.  That’s why I added three packages of ear-loop style masks to my daughter’s mask sampler.  They don’t filter as well as N95s, but they are more convenient and more stylish.

First, there’s a reason that all NIOSH-certified N95 respirators have behind-the-head straps, not ear loops.  You cannot get enough pressure from over-the-ear elastic loops to hold the respirator against the face with sufficient force to make an air-tight seal.  If you go with an ear-loop mask, you should understand that it’s not going to filter as well as a NIOSH-certified N95.

A second pitfall is that there are no U.S. standards for how well U.S. ear-loop masks actually function.  There are standards for the filtration efficiency of the cloth they are made from (typically expressed as BFE and PFE), but there’s nothing equivalent to actual on-the-face test that NIOSH-certified N95s must pass.  And that’s because, unlike an N95 respirator, you can’t get a compete seal from an ear-loop-style mask.  Unless you tape it to your face, it will leak to some degree.

Third, there are many masks sold as in the U.S. as “KN95” masks, but that has no legal meaning in the U.S.  (It does in China, but not the U.S.)  You can sell anything as a KN95 mask, and in the U.S., “KN95” really boils down to the flat-fold style typically found on actual, for-medical-use, Chinese-made KN95 masks.

KF94 is a Korean standard, and so, like the Chinese KN95 standard, it has no legal standing in the U.S.  You can and will see just about anything sold as a “KF94” mask. That said, because KF94 is less-well-know, and Korean supply channels are less polluted by fakes, that may not be as much of a problem as it is with “KN95” masks.

Fourth, you can buy and use a standard 20 cent blue ear-loop procedure mask.  But why?  I see those hanging off people’s faces all the time.  I always wonder why they think that does much good, when you can (e.g.) see their mouth through the gaps at the edge of the mask.  Even if worn correctly, in the standard fashion (not tucked-and-tied), these offer minimal protection.  (Don’t take my word for it. Read this mask-testing article in the Journal of the American Medical Association (JAMA).

Finally, if you really want to understand why N95s are better than cheap blue ear-loop procedure masks, read Post #938.  It’s not what the mask stops that counts.  It’s what the mask lets through.  And those masks let through a lot more virus than a properly-fitted respirator.  Again, see the JAMA article.  My graph (below) is an illustration of the difference between letting 70% of airborne particles through (typical ear-loop procedure mask) and letting 5% through (properly-fitting N95 respirator).


4:  Ear loop masks, my choices.

After working through all of that, I decided that my best choice for the ear-loop portion of my daughter’s mask sampler was to buy top-dollar, brand-name, Korean-made KF94 masks.  With adjustable ear loops.  I thought that provided the least chance for knock-offs, and the best level of filtration for an ear-loop-style mask.

I bought the following, all via Amazon.  Based on my research, these are three different well-known (in Korea) brands of Korean-made KF94 masks.  The cloth in these should filter almost exactly as well as the cloth in an N95 mask.  So, as with N95 respirators, it’s all about the fit.

An added bonus is that some come in black, and some come in small sizes.

Blauna, black, adult-sized mask.

BOTN, white, medium (small adult-sized mask)

LG, black, adult-sized mask.

These were relatively expensive, up to $2.50 per mask.

There were plenty of offers on Amazon for cheaper “KF94” masks.  Uniformly, of the ones I looked at, those much-cheaper offers tended tended to be a) made in China, and b) poorly made, based on user comments.

In the end, I took my own advice:  Post #935, If you have ten-cent lungs, by all means wear a ten-cent mask. I bought the expensive ones for my daughter.


5:  Placeholder for everything you never wanted to know about masks.

In theory, I’m going to come back to this and fill in all the details I have learned in the past half-year, regarding masks.  For now, let me just offer a few practical bits, all of which I have documented, at some time, in prior posts here.

How long does an N95 respirator last?  This depends on a lot of things, including how dusty your environment is.  In a clean environment, a mask such as a 3M 8210 will filter at an N95 level for hundreds of hours.

Typically, for the 3M N95s, the elastic is what limits the life of the mask, not the filter medium.  It gets stretched, and the mask gets too loose to seal properly.

That said, the filter medium will eventually clog.  As the filter medium gets near the end of its life, it gets harder to breathe through.  I experienced this first hand, early in the pandemic, when I couldn’t buy an N95 and so replaced the elastic on my 3M mask and kept wearing it.  Eventually, it will get so hard to breathe through that you will notice it.  Particularly if you have both a new and a used copy of the mask, to compare the back pressure.

How do you sterilize a mask after you’ve used it?.  The simple answer is, don’t.  If you are worried about the mask being contaminated, wear three masks in a three-day rotation.  Just letting the mask sit out in the air for a couple of days between uses is sufficient to reduce an viral contaminant on the mask surface to a negligible amount.

(And, in fact, last time I checked, there had not been even a documented case of spread of COVID-19 via fomites.  That is, via contaminated inanimate objects.  And that’s the main reason that you don’t get nagged about washing your hands much, any more.  Early on, CDC was worried about the potential for fomite transmission.  But I think experience has shown that if it occurs, it is vanishingly rare.

I’m a pretty cautious guy, as you might guess from this post.  But I don’t flinch at (e.g.) touching the cash register at the supermarket any more.  In fact, I’ve forgotten to wash my hands after my last N supermarket trips.  The whole “contaminated surfaces” thing was just another aspect of this that the CDC got wrong early on and never issued any type of clear statement correcting their initial position.)

What’s the difference between NIOSH-certified N95 respirator and FDA-certified-for-medical-use N95 respirator?  As far as the non-medical user is concerned, nothing.  Both filter out airborne particles to the exact same extent.  But medically-certified masks and respirators also have to stop a splash of liquid (e.g., spurting blood).  So the medically-certified ones have a waterproof factor that’s not required for (e.g.) food service or industrial use.

What are N95, BFE and PFE, HEPA, MERV, PM2.5, and so on?  These are filtration standards for masks, air filters, and the like.  I summarized that in the “Filtration Standards” section of Post #593.

 

Post #1235: Fairfax County K-12 COVID-19 cases

 

In a recent series of posts, I’ve started some analysis of COVID-19 cases in K-12 schools in Virginia.  Having worked with data all my life, the first thing I’d like to know is how reliable the data sources are.  So in this post, I first compare what my local school district reports to what the Commonwealth of Virginia reports, for what should be equivalent school-age populations.  And then take it from there. Continue reading Post #1235: Fairfax County K-12 COVID-19 cases

Post #1234: Redoing my baseline for analysis of COVID-19 spread in Virginia K-12 schools.

That headline above showed up on Google News this morning.  Their conclusion is consistent with what I thought I saw late last week, described in Post #1229.  Cases ticked up, but hospitalizations ticked down.  That’s what you’d get if the latest influx of cases was school-aged kids, as hospitalization for COVID is extremely rare in that age group. Continue reading Post #1234: Redoing my baseline for analysis of COVID-19 spread in Virginia K-12 schools.

Post #1233: Introduction to Virginia data on school outbreaks of COVID-19, a quick note.

 

 

This post looks at Virginia Department of Health information on outbreaks of COVID-19 in K-12 schools in Virginia.  Virginia not only tracks total outbreaks in schools, in most cases, it allows you to check on-line regarding the progress of an outbreak in an individual school.  Links given below. Continue reading Post #1233: Introduction to Virginia data on school outbreaks of COVID-19, a quick note.

Post #1232: COVID-19 trend to 9/3/2021

The U.S. stands at 50.7 new cases per 100,000 population per day.  That’s down a bit from yesterday (50.9) and the day before yesterday (51.3).

Accordingly, we had a little peak in the curve on September 1.  We’ll only know in hindsight if that holds up as the peak of the Delta wave.

We won’t know anything one way or the other until the middle-to-end of next week.  That’s due to the lack of data reporting over the weekend, coupled with this weekend being the Labor Day holiday. Continue reading Post #1232: COVID-19 trend to 9/3/2021

Post #1231: Setting up a test of school opening impact using Virginia data

This is just a brief announcement of something I intend to do.

In Virginia, school districts choose their own opening dates, subject to some restrictions imposed by the state.  As a consequence, the opening days for Virginia public schools traditionally span more than a month.   Anything from early August through mid-September.  These opening dates also tend to be quite “sticky” from year to year.  In other words, school districts that traditionally open earlier or later tend to do that year after year.

This seems like a reasonably good “natural experiment” for looking for the impact of school openings on the spread of the Delta variant.  It’s not a true experiment, because individual student’s start dates aren’t randomized.  Start dates for entire school districts aren’t randomized, so I don’t think it qualifies as a “quasi-experiment” either.

It’s a pretty good natural experiment, because those start dates were essentially determined years in advance.  They can certainly be correlated with other factors, if, say, school traditions in rural and urban areas differ.  Or if adjacent counties try to stay in sync with each other.

But, presumably, any COVID-19 related conditions in a county, up to the start of schools, are what they are.  And what I’ll be looking for is a change in trend, for the school-age population, starting two weeks after the school opening.

There’s a further complication, in that Virginia releases the age x day x region COVID-19 case counts only at the level of Virginia health districts.  So I have to crosswalk school districts to Virginia health districts, and aggregate the start dates by health district.

Finally, the consolidated school calendars of all the Virginia school districts has not yet been published by the Virginia Department of Education for the 2021-2022 school year.  So I’m working from the dates on last year’s calendars.  That said, all I’m doing is assigning Virginia health districts to early, middle, and late typical school start dates.  The early and late date districts start schools more than two weeks apart, on average.

The statistical power of an analysis of this sort depends on how the data behave in a time-series.  If the share of COVID-19 cases in school-age children is rock-solid-steady prior to the opening of school, even a small impact of school openings will be readily apparent.  By contrast, if it’s jumping all over the place prior to the start of school, even a large impact will get lost in all that random time-series “noise”.

(I should note that most statisticians do their analysis of natural experiments incorrectly, using cross-sectional variation as the basis for their statistical tests.  That’s not right.  I don’t much care if schoolchildren’s share of cases varies from county to county.  What I care about is how it varies from day to day.  If it’s really steady, and then jumps up at the right time, I’m pretty sure I’ve captured the impact of school opening, regardless of how that share varies across regions on any given day.)

This analysis is something of a rough cut, in that my aggregation of counties into health districts is un-weighted (each county counts the same).  But in most cases, it looks like it will assign the health districts to the right group of early-middle-late school openings.  For one thing, really populous areas form their own health districts.  So for the bulk of Virginia’s cases, that issue of sloppy aggregate of individual school districts does not even arise.

Separately, I can also check the “outbreak” statistics captured by Virginia.  Any time you get multiple cases, at the same time, within a school, that should be separately flagged in Virginia’s data.  But “outbreaks” are a lot “grainier” than the case counts — they come in big, discrete lumps.  So I think I’ll have better statistical power if I stick with case counts.

Anyway, owing to the generally late opening of Virginia schools, this is just the warm-up exercise.  I’m setting up the programming and checking the stability of the numbers, so I can revisit this in a couple of weeks.

Here’s the preliminary result.

Source:  Calculated from Virginia Department of Health COVID-19 case counts by age, date, and health division, and 2020-2021 (sic) school district schedules from the Virginia Department of Education.

I’m not sure what to make of that.  I had hoped to see three parallel lines.  Instead, the health districts containing early-opening school systems already show an increase in the fraction of cases among individuals under age 20.

But, in fact, almost half the school districts, within the health districts where typical school opening date is early, should have opened early enough that school-spread cases should have worked their way into the system by 9/2/2021.  That is, it’s vaguely plausible that the rise of the blue line is an actual impact of school reopening.

So I’m not going to dismiss this analysis quite yet.  I’ll come back in a couple of weeks to see if the other two curves pull upward, in sequence, as the school year openings progress across the state.

Post #1230: COVID-19 trend to 9/2/2021

The U.S. stands at 50.9 new COVID-19 cases per 100,000 population per day, down a bit from yesterday.  The seven-day increase was 5%, also down a bit from yesterday.

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 9/3/2021, 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.

I’ve added a linear trend line to the chart of showing the growth rate in new cases.  This is so you can see where I’ve been coming from with the idea that we ought to see the peak of the Delta wave in early September.  Up until about a week ago, we were smoothly on track for that.  But that’s stalled this past week.

CDC is still showing that hospitalizations peaked for the U.S. as a whole.  That said, if I look at individual states, cases and hospitalizations appear to be in synch.

Here’s the U.S.

Source:  CDC Covid data tracker.

Here’s e.g., Texas, Florida, California.  There, the fall in hospitalizations matches the fall in new cases.

I’m still looking for direct evidence that the last little blip I saw was related to re-opening of schools.  It is, as I have noted before, tough to get any sort of U.S. breakout of new cases by age, in any sort of timely fashion.

I can get that information for Virginia, but many Virginia school districts start school after Labor Day.

The bottom line is that I’m still looking for any reasonable confirmation that the age mix of new cases showed some sudden shift toward school-age children.

As for the rate of testing, well, the CDC’s reported data have such long lags that it’s impossible to tell, yet, nationally, whether there was a clear increase in testing associated with school openings.

Post #1229: COVID-19 trend to 9/1/2021, back-to-school uptick

 

Yesterday, there was a little uptick in the numbers that I didn’t expect to see.  Today, the message in the tea leaves appears much clearer.  Enough so that I’m going to jump to a conclusion based on two days’ deviation from trend.

It’s starting to look like we can toss my prediction about the path of the Delta wave into the trash.  But the upside is that we can do that for a very specific reason: School is back in session.

Even if that’s right — if this current little uptick in cases is a back-to-school uptick  — it’s not yet clear the extent to which that reflects more actual infections, or merely more screening, testing, and discovery of infections.  That’s what I’ll try to look at next.

Continue reading Post #1229: COVID-19 trend to 9/1/2021, back-to-school uptick