Covid-19 August Check-in

CDC Sees Shadow: Six More Weeks of Masking

It feels a bit like Groundhog Day. Here we are in the heat of August again. Covid cases are high again. The largest rises are in the southern states again. And we’re just weeks from the start of school again. Decisionmakers staring at Delta’s epidemic curves were no doubt already wrestling with the question of “what next?” even before the CDC last week updated its recommendations. The two CDC recommendations that have garnered the most attention are:

[1] Added a recommendation for fully vaccinated people to wear a mask in public indoor settings in areas of substantial or high transmission;

[2] CDC recommends universal indoor masking for all teachers, staff, students, and visitors to schools, regardless of vaccination status.

And the poked beehive that is Twitter—especially Covid Twitter—has been locked in battle ever since. And by that, I mean they have continued a battle that’s been raging for the better part of 18 months and morphed as new data and variables have entered the mix. It’s a battle that includes considerations like risk, policy, data, tradeoffs, efficacy, second-order effects, and more. I’m looking forward to these battles winding down at some point. But as Aragorn said, “It is not this day.”

In the meantime, I have collected some thoughts on a few Covid topics as we enter August. Thank you for taking the time to read.

---

The Summer Wave Is Bigger (And Closer to Breaking) Than It Appears

The Delta variant of Covid-19 has helped spur a larger summer wave than many of us predicted. We’re not talking about 50-year storm at Bells Beach big (RIP Bodhi), but certainly larger than I thought we’d see—at least in terms of case counts. During summer 2020, the US 7-day average for cases peaked at 68,708 (per CDC data). We surpassed our 2020 total this past Friday, July 30.¹ I anticipate we’ll surpass 2020’s 7-day average for peak hospitalization census of 39,769 shortly as well. Deaths so far continue to remain much lower than last year, but it’s likely that our peak deaths from this wave won’t materialize until around the end of the month.

The states showing the most vertical case rises right now are in fact even more vertical than they appear. Many on Twitter have been using the UK’s epidemic curve timeline as a blueprint, but the states driving the rise in the US have dwarfed the UK in terms of their rate of increase in actual infections when you factor in testing. From May 15 through its reported case peak on July 21, the UK conducted approximately 1 test for every single inhabitant. For the US to get to that point by late August, we’d need to increase our daily tests by roughly 17x, to over 10 million tests per day. Not happening. Even if we double testing from July to August (certainly possible as school ramps up), we would end up with about 1 test for every 6 inhabitants during the same ~9-10 week rise in the epi curve.

What does this mean? It means the US has had a metric shit ton more infections in the past 4-6 weeks than show up on our epi curve. It means that our week-over-week rate of increase in infections (not just captured cases) has hit rates that blow the UK out of the water. It means that, though we gave the UK a 5- to 6-week head start in terms of being saturated with Delta, the fact that our entire country has been essentially wide-open this summer as Delta became saturated has likely caused us to close the gap. I talked about this with specific reference to Florida in my last post:

So it’s likely the underlying cases over the past several weeks have increased at a much higher clip than the 90+% peak shown above. These types of numbers harken back more to the “cases are 10x” multiplier from early in the pandemic than the “cases are 2.5-5x” at other times (depending on the spread and positive testing percentage).

Much like a wave at the beach, our Covid-19 wave has grown while the amount of water between it and the ocean floor below continues to shrink. And while the visible portion of the US wave is smaller than the UK’s was at this point in its Delta rise, the true US wave size is much closer on the UK’s heels, and so too is the inevitable break.

---

The Battle Over Masks

I’m not going to delve too deeply into the mask debate, because it is by far the most maddening of all Twitter battles in which I find myself sometimes engaged. I suspect it is the single biggest point of contention in all of Covidium right now. I have gone through more literature and studies on the efficacy of masks in terms of helping curb Covid/respiratory viruses than I care to admit. If one were so inclined, one can find giant Twitter threads citing numerous studies coming to contrary conclusions. I will say this: I was more bullish on masks as acting (at a minimum) as a “might as well”-type precaution that in the aggregate helps curb Covid spread before my trip around the sun (and before seeing so many Covid epi curves moving irrespective of mandates or compliance when compared to others within relative geographic proximity).

My circuitous, data-driven path to question the effect of masks on stunting the top of the Covid epi curve—especially given the types of masks and actual on-the-ground usage by the general public—pales in comparison to the absolute certitude of those who believe that masks are the most important tool in the battle against Covid, potentially excepting vaccines (some may even put masks ahead of vaccines given the innumerate and irresponsible coverage of Delta breakthroughs; more on that later). I mean, the folks who are absolutely certain of cloth mask efficacy against Covid are really certain—we’re talking gravity-level certain. I could convince the most anti-mask people on Twitter to wear seven masks and a face shield to bed while living alone before I could convince some of the most staunch mask supporters to concede even the possibility that masks provide at most a negligible benefit that doesn’t translate into a material change in the epi curve.

Of course, they are equally certain that masking very young children for hours daily or requiring weeks of quarantine for a kindergartener (starting real school for the first time—how exciting!) based on a potential classroom contact are no-questions-asked absolute policy slam dunks that easily pass any cost-benefit analysis. No data on Covid’s effect on children or transmission rates in schools or how European countries have approached schooling matters. In fact, any potential data-based argument will be condensed to things those “whiny” adults think. Kids under 10 years old are as hard as wrought iron, we’re told, and there is nothing at all to worry about in terms of any deleterious effect for a child to remain masked for the better part of two school years (if they’ve been in school in person at all since spring 2020). I happen to think that—at a minimum—it’s worth discussing whether the juice is worth the squeeze. But that’s about as far as I’m willing to go down the rabbit hole on masks, lest I be lumped in as a Covid denier for saying, “Hey, maybe we should be as deliberate with our decisions as some of the European countries.”

---

Hospitalized “From” or “With” Covid?

One thing that still irks me as someone who enjoys parsing data is our apparent inability to collect more detailed hospitalization data at admission. The first and perhaps most important piece of data to collect is why that person showed up at the hospital that day. Was it due to Covid-like illness or not?

A positive Covid test for someone admitted for a broken femur or to deliver a baby is still helpful data, but as the background infection rate in a particular area rises dramatically (as is the case now), it will increase incidental positives. Not only that, but we’re likely to capture a much higher percentage of admitted infections as cases than we can in the population at large. So while cases may triple in a state (for example), a rapidly rising positive testing percentage indicates an even higher rate of increase in underlying infections. Hospitals will better capture infections because they’re swabbing damn near everyone, which in turn increases the percentage of cases that appear to require hospitalization.

On top of separating out incidental Covid, it would be nice to further break those Covid-19 positive patients (admitted because of Covid-like illness) into groups based on the presence of antibodies and vaccination status, including date of last jab. This may become more important as we continue to add months to the time someone has been since becoming full vaccinated. Basically this:

Craig @TheLawyerCraig

Something simple like this would be fine:

---

The Media’s Vaccine Innumeracy Is Bad. It’s Going To Get Worse.

Imagine for a moment a country in which 100% of people 55 years old and older are fully vaccinated and 0% of people under 55 have been vaccinated. You might read headlines like, “More Covid deaths in vaccinated than unvaccinated.” Or worse, “Despite smaller percentage of population, most of Covid deaths in vaccinated.”

I haven’t drilled down the percentages on breakthroughs and age-based mortality, but the back of the envelope math works out in the hypothetical country above such that both headlines would be entirely accurate—and entirely misleading. And the media would not hesitate for a second to blast those headlines and rack up those fear-induced clicks and shares. There is no question that statistically the chance of an unvaccinated 25-year-old dying of Covid is lower than that of a vaccinated 85-year-old (in fact, around 100x lower based on the charts I’ve seen that run the math). Guess what? That cold, hard math does nothing to call into question the vaccine’s effectiveness.

Now take the world in which we live: we don’t have a 100% v. 0% breakdown like the above example, but vaccination percentages are much higher in the older cohort and generally go down with age. We also have several states with 75%+ of adults fully vaccinated and 85%+ of seniors vaccinated. Eventually, hospitalizations and deaths in vaccinated individuals will outpace the unvaccinated. In fact, I would be very surprised if that isn’t the case this fall in the Northeast, given the high vaccination rates there. Some in the media will provide context, perhaps even put in some effort to practice so they can actually demonstrate the type of command of the data required to distill it down into bite-size morsels for consumption. Many others will simply act like Rene Russo’s character in Nightcrawler and either put on horse blinders or knowingly seek out an inaccurate framing in order to best elevate themselves via fear-based clicks and shares.

---

Other Tidbits

The Vaccines Are Effective. The data on the vaccine shows me that it is remarkably effective. Virginia stratifies outcomes based on vaccination status. Looking solely at July 1-30, the hospitalizations break down 413 unvaccinated to 7 vaccinated. Deaths fall 18 to 1. In July, more than 50% of Virginians were fully vaccinated, but much more than 50% of the cohort most at risk are vaccinated, so it’s even better performance than a pure vax/unvax comparison (see the “Innumeracy” section above for why).

Early Signals Of Peaks? I noted last week on Twitter that Greene County, MO—one of the first Delta-infested areas—peaked on July 20. As of today, average daily cases are down 29% from the peak. Missouri as a whole is also reporting a decline in week-over-week confirmed cases as of this morning. Though its reported cases tend to be volatile (and can backfill), to get to an entire WoW decline usually means you’re on the backside of the peak. I’m interested in seeing where Arkansas, Nevada, Utah, Florida, Texas, California (especially LA), and the Gulf Coast trend this week—most of which have seen their rate of increase slow, and some dramatically so.

Play Date? I’m curious how many people would not allow their 12- to 13-year-old child to hang out outside of school with an unvaccinated friend. Or, imagine your child’s very close friend’s entire family was not vaccinated, but your (healthy) child was. Or, imagine it’s your 11-year-old child who cannot currently be vaccinated. Would you let your child hang out at that family’s place—maybe for a pool party? Go out to dinner with them? Given the reams of data on children and Covid, I’d be an unqualified “Yes”. In fact, I can’t even imagine asking the question to find out in such a situation. And if they volunteered it, the last thing I think I could ever do is pull some kind of “Eww, no” before saying our kids couldn’t hang out. Swap out the 11-year-old for an 85-year-old parent? Completely different calculus, obviously.

“Passport, Please”. Can we please retire the (seemingly intentional) talking past each other regarding whatever the hell people mean by “vaccine passports” or “proof of vaccination” or whatever we want to call it? Just say at the outset whether you’re talking about the school/international travel bucket or the grocery store/restaurant/gas station bucket (or some other bucket) so others can at least opt to engage you in good faith.

---

If you liked this post, feel free to share it.

And if you don’t yet subscribe to The Issue, please click on the button below. It’s free.

1. Recent CDC data is often revised, and it’s possible that this date could shift by a day.