Last March, I strongly advocated strict stay-at-home orders (aka “lockdowns”) throughout most of the United States. When they said “two weeks to bend the curve,” I immediately demanded extensions. When those extensions were granted, I cheered. When states began reopening in May, before meeting the benchmarks I (and the CDC) had set, I was profoundly disappointed.

Here is the basic pro-lockdown argument that I made:

1. Left unchecked, covid would have infected and killed something like 1 million Americans.
2. Lockdowns lasting 2-5 months (long enough to suppress the virus) would drastically reduce covid’s spread, saving several hundred thousand Americans.
3. The likely benefit of saving several hundred thousand Americans (even mostly elderly Americans) was greater than the (likely catastrophic) economic cost.
4. Therefore, we should impose lockdowns.

Many people disagreed with me.

At the time, at least in my Facebook circles, there were two main arguments against lockdowns:

~1. (Not-1) Covid isn’t very deadly, and will not kill more than 100,000 Americans in total. (The exact number used varied, a lot, but was most often between 50,000 and 100,000, because flu kills around that many Americans in a severe flu season.)

~3. (Not-3) Even if covid is going to kill a huge number of Americans, the economic costs of saving all those Americans will be so high that it won’t be worth it.

~3a. (Not-3, variant) Even if covid is going to kill a huge number of Americans, lockdowns will also cause a huge number of Americans to die (from suicide, untreated medical crises, et cetera), erasing the benefits of suppressing the virus.

Basically: covid isn’t that bad to begin with, and, even if it is, the lockdown cure is worse than the disease.

I am comfortable saying today that that those counter-arguments were wrong.

Covid turned out to be pretty deadly. In March 2020, I wrote, “Under ideal circumstances, the covid virus appears to kill about 0.6% of the people it infects.” I cited evidence that covid was likely to infect between 20% and 60% of Americans before natural herd immunity kicked in, causing between 400,000 and 1,200,000 deaths.

That seems to have been about right. The best estimates today now place covid’s fatality rate between 0.5% and 0.8%. Covid has killed 500,000 Americans so far, and it looks like it will kill a grand total somewhere between 600,000 and 700,000 when all’s said and done. The end, when it comes, won’t be from natural herd immunity, but from vaccines, which the Trump Administration and American industry produced with miraculous speed. If not for Trump’s warp-speed vaccine development, we probably would be staring down the barrel of a million or so dead — and even that’s pretty close to ideal, because covid turned out to have a higher herd immunity threshold. Somehow, our medical system was able to hold together and never quite ran out of hospital beds (which would have caused far more deaths).

Meanwhile, the economic damage was bad… but it wasn’t nearly as bad as I expected. I wrote last year:

Let’s accept that, regardless of what we do, pandemic makes major recession inevitable. People will die, supply chains collapse, consumer spending will vanish, it’s a nightmare. So let’s say, with… moderate measures… we lose 10% of GDP next year. (That’s several times larger than the Great Recession.) $2 trillion vanished in goods and services, taken from all of us, especially the poorest.

Let’s then assume that taking extreme measures makes things twice as bad. We lose 20% of GDP. It’s three years of the Great Depression all packed into one hell year. $4 trillion gone.

If those assumptions are true (and they’re not — they are at best in the ballpark), then it means that our additional safety measures need to save at least ($2 trillion / $9 million =) 222,000 American lives in order to be economically “worth it”.

None of that happened. My (deliberately pessimistic) expectations were not in the ballpark.

In Q2 2020, the economy shrank by an oh-my-God-my-stomach-just-dropped-into-the-second-subbasement rate of 7.85% (31% annualized). At that moment, it did look like the lockdowns had perhaps brought about an economic apocalypse. There had never been a GDP number that bad ever. Even the Great Depression didn’t hit that hard in a single quarter. But, in Q3 2020, the economy rebounded, growing 8.35% (33% annualized) — the best GDP growth number in history.

The United States ended up coming through 2020 with a 3.5% loss in GDP. That’s still a terrible economic shock — not as bad as the Great Recession, but worse than anything else since the Oil Crisis Recession. The way our bailouts favored large corporate consolidation while ma-and-pa stores got wiped out was a disaster we’ll spend years fixing — too late to bring back the ma-and-pa stores ruined in the process. But it was not a depression-level event. Here’s U.S. unemployment (both the “official” U-3 and the “comprehensive” U-6 measures):

Unemployment of 6.4% is, again, real bad. I think Republicans were very wrong to accuse Democrats of making the economic pain worse to harm President Trump… but, if I looked at this chart without knowing anything else, I might very well ask God Almighty why He decided to destroy the economy at precisely the right moment to weaken President Trump’s re-election campaign. Nevertheless, after months of recovery, unemployment today is where it was in 2013 and most of the 1980s — not great, not terrible.

Could we have done better if we hadn’t had a national wave of lockdowns? Maybe a little better, but probably not. Lockdown opponents rallied around Sweden as a (relatively) shining example of how to stay open in a pandemic, arguing that Sweden would build up herd immunity and come out the other side with a much better economy than all of us lockdowners.

But Sweden ended up down 2.2% of GDP over the course of 2020. That’s a bit better than us, but not enormously, and worse than some countries with tighter lockdowns than we had. Lithuania’s lockdown was much tighter and more comprehensive than the average American lockdown, but their economy lost only 1% of GDP in 2020.

Here’s a chart whose x-axis measures something pretty similar, for a broad range of countries:

Do you see a strong correlation between “tightness of lockdown” and “amount of economic damage”? Just eyeballing it, I don’t. And more thorough analysis (like the post this graph is from) doesn’t uncover many clear trends here, either — other than the obvious “being an island nation can really pay off, except the UK for some reason.”

This chart, from here, I don’t like quite as much, but it draws out the GDP growth / lockdown correlation a bit more clearly:

Shorter lockdowns don’t have any obvious GDP benefit.

Meanwhile, the epidemic of “deaths of despair” that was forecast by some lockdown skeptics just haven’t materialized. Lyman Stone at the American Enterprise Institute has been fanatically tracking excess death counts for months, and what he keeps finding is (1) the vast proportion of excess deaths in 2020 were covid or undiagnosed covid or misdiagnosed covid, (2) the other excess deaths are correlated with covid outbreaks but not with lockdowns, and (3) an awful lot of people want to make the contrarian narrative work, but it just doesn’t hold up to scrutiny. Some handy (if long) Twitter threads of his: 1, 2, 3, 4

So: covid did kill a whole bunch of people and — although the pandemic caused a global recession — lockdowns specifically seem to have only made the economy and non-covid mortality a little worse, if that.

Not bad! But this post isn’t called “What I Got Right About Lockdowns.” It’s called “What I Got Wrong About Lockdowns.” And it was a doozy.

The lockdowns don’t seem to have saved very many lives.

This is unexpected. Pretty much the one thing everyone on all sides seemed to agree on back in March 2020 was that lockdowns were a brutal but effective method for controlling the virus. It’s such a simple, obvious theory (so simple a 14-year-old girl thought of it!): keep people at home, they can’t go around spreading germs, the virus dies out… or at least is slowed drastically. Who could argue against it? None of the lockdown skeptics I was talking to. In my 3-pronged argument for lockdowns, my other premises came under heavy fire, but I never had serious arguments about this one.

Yet, by early May, I was developing some doubts. I wrote then:

While we know quite a bit about the characteristics of the virus itself, we DON’T really know very much about the social dynamics of mitigation, lockdown, and so forth. Much of what we do know is derived from not-entirely-analogous situations from the flu pandemic of a hundred years ago — not the best data set.

Consider the famous Imperial College model… [It made] assumptions [about] how people would behave. 

For example, they assumed that, if infected households were asked to quarantine, those households would reduce community contacts by 75% and that 50% of households would comply. They assumed that, under a broad stay-at-home order, everyone would reduce external contacts by 75% (while increasing internal contacts by 25%).

Are these numbers reasonable? They seem reasonable to me. Are they correct? …how could we possibly know? We’ve never done anything remotely like this in modern times. The paper certainly cites no sources. These are educated guesses about human behavior. And if you tweak these parameters just a twinge, it can make the difference…

So it’s quite possible that government mitigation efforts combined with people’s natural desire to stay home in the midst of an epidemic does a better job of reducing disease spread than the experts think. It’s also possible that the added benefits of legally-binding stay-at-home are smaller than we think they are. Either way, moving to a “Sweden model,” where the government issues strong recommendations and tightly enforces social distancing in public spaces, but does not actually make you stay home, might lead to fewer deaths than one would expect based on the virus’s lethality.

Lockdowns, according to the theory, save lives by reducing transmission. You impose a lockdown, then, a week, two weeks, maaaaybe four weeks later, with the transmission links cut off, positive tests start to fall. If lockdowns don’t reduce cases in a significant, measurable way, then they can’t reduce deaths. If they don’t reduce deaths, then they are pointless.

Here’s a chart of Minnesota’s (estimated) cases over a time period of exactly 14 weeks. (The data is from my covid weather reports last year.)

Look at this chart. Based on the theory above, when do you think Minnesota’s lockdown started? When did it end?

Point at spots on the chart that make sense.

Remember: this chart covers 14 weeks. I added hash marks to the bottom to help you out.

You got a spot picked out? Good. So do I. Here’s the answer:

(No, seriously, you got some spots picked out? No peeking!)

Okay, here it is:

Lockdowns started on March 28th. Two to four weeks later, cases were not falling into suppression; they were skyrocketing. Then, we plateaued for four weeks, with a modest dip in the first week of May. Near the end of that plateau, we ended the lockdown… and then, instead of going through the roof as we expected, cases started to plummet instead.

This graph makes the whole idea of lockdowns look like a joke. I hate saying that, because I really can’t overemphasize how intensely I favored lockdowns during precisely this time window. Given the same information, I’d probably even make the same decision again.

But we now have lots and lots of information not available to us in March 2020, and it does not make lockdowns look good.

The central problem is, Minnesota is not a fluke. You can make this basic chart with a lot of states. Sometimes, cases start falling before lockdowns could possibly have had an effect; sometimes, they start falling long after any lockdown effect was priced in. Either way, it’s clear that, whatever caused case numbers to start falling, it wasn’t lockdowns.

Marginally Compelling has examined a couple of paradigm cases — Florida and New York — in some detail, and, while the full post is subscribers-only, there’s a good summary on Twitter. Lyman Stone was showing how lockdowns fell short in Europe way back in late April 2020. The passionately anti-lockdown (and not always reliable) AIER has a roundup of scientific studies that indicate lockdowns were ineffective, and/or that pro-lockdown studies made incorrect assumptions.

Another, easier way of looking at it is to just compare states that did lockdowns to states that did not do lockdowns, or did shorter lockdowns. Is there a consistent, meaningful difference between how more restrictive states weathered any given wave of covid, compared to their demographically-and-geographically-similar neighbors who were less restrictive? Is there any apparent change in direction in any state based on its lockdown dates? Can you (correctly) pick out start and end dates for lockdowns in, say, half of them? If lockdown theory is even slightly right, the answer for all three should be an easy, obvious Yes.

But the answer is No:

That chart is from Marginally Compelling’s very handy “Your State’s Covid Numbers In Context,” July Edition, a monthly newsletter where he just charts every state against every other state in each region. So if you didn’t believe me from this chart, you are free to jump over there and look at 56 more charts (8 per month for 7 months) until you are convinced that lockdowns have no visible impact on cases or deaths — and are therefore useless.

As Marginally wrote, lockdown represents “the SimCity version of government, where dialing up the lockdown policies dials down the viral infections.” It feels like it should work. It feels like doing something against a horrifying pandemic that has claimed far too many lives. But it doesn’t appear to actually help, and wishing really hard that your state’s covid lockdown will work if you personally just comply hard enough with it… well, that doesn’t help, either.

This doesn’t mean that all covid interventions are useless. You should wear a mask, regardless of whether anyone forces you to or not. Masks seem to work. (But state & municipal mask mandates evidently do not: check those state charts again.) Regulations limiting restaurant capacity still strike me as probably a good idea, although I haven’t looked at the data in months and could be wrong. Large, dense, indoor gatherings of strangers, like concerts, are still a terrible idea, and states are still right to ban them. (Churches can and should open with low density and appropriate precautions.) When hospital capacity was near its limit, lockdowns themselves don’t seem to have done much to control spread, but a closely related factor — fear — appears to have helped. Public authorities on the brink of a hospital crisis are probably smart to spread fear of what that would look like.

Yet the best interventions — the ones that can actually get case loads down to Taiwan levels so everyone can return to normal — appear to be (1) closing your borders to prevent initial epidemic “seeding” (ideally, you want to be an island nation), and then (2) aggressively testing, tracing, and isolating cases so the disease never gets out of control.

Once the disease does get out of control, your best bet seems to be internal border controls, where you deploy troops to seal off out-of-control parts of the country to control the damage — something we, unfortunately, never seriously contemplated in the United States, aside from a handful of unserious attempts imposed to score political points, far too late to control the virus. But Finland did it, with some apparent success.

Once you’ve let the disease get out of control in an area, though, there seems to be no way back except (3) Wuhan-style lockdowns, where you weld people into their homes and disappear the violators, or (4) herd immunity. At least, I can’t think of a free country that successfully moved from epidemic back to containment.

(“But New Zealand!” New Zealand closed its borders when there were 28 total cases, zero deaths, and test positivity rates low enough to be confident there actually were only 28 cases in the country. Australia was only a couple ticks behind them. Covid was never out of control in either country. Compare: by the time the United States started locking down, on March 19th, we had an estimated half-million cases in circulation, our testing was still trash, and thousands of people who would ultimately die of the virus already had it.)

The disease got out of control in the United States way the heck back in February 2020, back when our entire testing apparatus was on fire, then-candidate Biden was calling President Trump a xenophobe for Trump’s (good but overly timid) Wuhan travel ban, and we were being told to worry about anti-Asian racism and the flu even more than we were being told to worry about the virus.

Once that happened, it seems to be the case that most of our half-million-plus deaths were more or less inevitable. Our saving grace has been Operation Warp Speed and the most rapidly-deployed vaccines in history. (It is sheer lunacy that we delayed approval of the Johnson & Johnson vaccine until nearly March so the FDA could dot all the i’s on the application form, and absolute crazy-pants that we haven’t approved AstraZeneca… but that’s another blog post.) The vaccines will end the pandemic and save hundreds of thousands of Americans’ lives over the course of 2021.

Which, unfortunately, is a lot more than we can say for the lockdowns of 2020.

Minnesota is in the midst of a very dangerous outbreak of covid-19, the likes of which we have not seen before. The situation is much worse today than it was at any time during the lockdown.

DETAIL:

I put the data in, ran this chart, and swore:

This data lags by seven days. “LTC” is short for “long-term care” (basically, nursing homes).

There are a lot of cases circulating right now. My rough estimate is that there are about 75,000 Minnesotans right now who are either pre-symptomatic or early symptomatic — probably walking around and potentially infecting you.

To put that in perspective: our previous high was 24,000 (estimated) walking infectious on May 9th. From June 2nd to September 19th, the number never went above 10,000. I wouldn’t swear by the exact numbers here, but the sheer scale of how big this outbreak has gotten in the past couple of weeks is clear.

To put that in further perspective: if you see 100 Minnesotans today, this estimate says that 1 or 2 of them has covid and is infectious… and doesn’t know it yet.

For obvious reasons, that’s very dangerous, especially if you’re an older person or are otherwise at risk.

If there’s good news this week, it’s that cases are not rising nearly as quickly in nursing homes as in the general population:

Cases are still rising in long-term care homes — and that spike has killed a bunch of old people — but we’ve managed so far to keep it from getting out of control in nursing homes (unlike in April).

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.) This is crude enough that, when positivity is significantly above 2% (and it is right now!), the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Our testing capacity continues to grow, but, in the past two weeks, the huge growth in cases has outpaced the growth in testing:

We are way outside the 2% “safe zone.” We are now also outside the 5% “good enough zone.” We’re past the 10% “we’re kind of flying blind” zone. And that’s not for lack of testing! We’re running over 30,000 PCR tests per day right now!

(We’re also running a few thousand antigen tests per day, but I’ve excluded those because of their high positivity rates and need for a confirmatory PCR test. So this is how bad the numbers still are after I nudge them down a little bit with my choice of data.)

That means we are missing more cases. My case estimates at the top attempt to account for that, but, as positivity rises, we become increasingly blind to the spread of the infection, and my estimate (like everyone else’s) becomes less certain.

As I noted at the top, all this data lags by 7 days.

At first glance, hospitalizations don’t look… that bad:

So, according to state data, we aren’t doing too badly on hospital capacity. (We also have around 3,000 free non-ICU beds, too many to even put on this chart.)

But I’ve learned that state data isn’t telling the full story. Both media reports and firsthand accounts indicate that, while ICU beds are sitting out there unused, qualified ICU staff are in the middle of a severe shortage. This is partly because, hey, the economy needs more nurses, but also partly because a whole lot of nursing and medical staff are currently home, either infected with covid or quarantined due to exposure.

Basically, we’re quickly running out of ICU capacity, and it’s unclear how close we are to our actual “wall.”

UPDATE: This isn’t data, but was sent to me by a friend tonight, who is himself friends with some medical workers, and he says he is hearing the same thing from “multiple sources.” (Also, I know you guys skip to the pictures before reading the words, so here’s a picture.)

And the tsunami is not letting up:

Both hospitalization charts mentions “new data.” That is because of last month’s switch in how Minnesota reports hospitalization data. As you can see, there’s a significant gap between how many beds were counted as occupied under the old method compared to the new method. But, however you slice it, it appears that our hospitals are under significant strain right now.

If there’s any comfort here, it’s that deaths are still relatively low, but… if several months of reading Marginally Compelling‘s state-by-state monthly covid updates have taught me anything, it’s that a big wave of cases like this is followed by a wave of deaths later:

I have some hope based on recent studies which show covid mortality rates falling (presumably due to better treatments, although there are other possible causes). I don’t think it’s going to get as bad as the Minnesota Model projected, but it’s not going to be fun, either.

HOW BAD WILL THIS GET?

What I’ve learned from other states is that waves like this tend to last 6-10 weeks or so, then subside. By my count, we may be in Week 7 of the current wave. (We may also be a little earlier than that.) So this could begin to reverse course soon.

But, frankly, I think these things tend to end largely because people get scared and change their behavior, and I see a lot of people too exhausted by the pandemic (and distracted by the ongoing presidential election) to get properly scared of this thing again.

WHAT TO DO

I don’t expect another lockdown, because the first lockdown was not effective. Plus, a lot of it was pointless. People don’t seem to be catching covid from socially distanced shopping at Ace Hardware with only fleeting contacts with carriers. They almost certainly aren’t getting covid from playing on the playground outdoors.

However, I fully expect Minnesota to close bars, restaurants, and theaters (again), for everything but takeout, within the next week or two. The only thing that gives me pause is… I don’t know how those businesses will survive without immediate financial relief. If Gov. Walz can’t find a way to get these businesses money, he may leave them open despite the death toll.

You, however, should choose to avoid dining out anywhere.

Schools that are open have bent over backwards to stay open, and may still be able to shutter themselves on an outbreak-by-outbreak basis, but… I’m nervous. My daughter’s school (in-person, 5 days/week) had zero cases from August until November 1st. Then it had one case in the preschool, followed the next day by one in 4th grade, and another in 5th grade a day later. There’s no sign yet of transmission within the school building (these were all picked up outside school, it seems), but social distancing protocols can only get you so far when there’s this much covid in general circulation. I just don’t know.

As for you: you should try to avoid getting this. If you’re young, you probably won’t die, but you might need a hospital bed, and we’re running short on those. Please don’t take one up unnecessarily.

And if you’re old or sick or both, you might very well die. Sure, you might have only a 20% chance of dying in your age bracket… but we just sat through an election that went great for Republicans, where President Trump almost came from behind to win (but he fell short; expect an official race call for Biden soon). What was the likelihood of that? As it happens… about 20%. Don’t risk it!

Pod up with a small number of other people — 1 or 2 other households, perhaps — monitor symptoms and exposures, and minimize contacts with people outside your pod. That’s all we can do, really.

FOR MORE FREQUENT UPDATES

Frankly, I’m sometimes not sure why I even bother continuing to do these. Early in the pandemic, lots of people were still getting a handle on the data, but now we have David Montgomery slamming out solid stats every day. He’s my go-to for daily updates, as are his MPR covid reports. If you’re looking for data updated more often than my roughly fortnightly updates, that’s where I would go.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago. And I finally uploaded my giant Excel worksheet to Google Sheets so you can see the raw numbers and calculations I am doing. Fair warning, though: I did absolutely nothing to clean it up. It’s a mess. Please enjoy Reinhart-n-Rogoffing me.

After more than one false alarm, Minnesota is officially in its second wave of covid-19.

I told you many times over the summer not to worry about media-hyped “record-breaking case counts”; case counts were increasing because testing was increasing. Not this time. This one’s real.

Minnesota likely has more actual infections than it ever has before. It definitely has more hospitalizations. It may soon have more deaths.

You should take appropriate precautions. Tighten your social graph back up. Reduce shopping trips. If you started meeting with friends indoors who are not in your bubble, time to stop. If you started working from the office again and can return to working from home, make an excuse to do so for the next few weeks.

More on this at the end. First, let’s survey the situation.

DETAIL:

Basically, we have a lot of cases, even after you account for increased testing. Like so:

This data lags by seven days. “LTC” is short for “long-term care” (basically, nursing homes).

If there’s good news this week, it’s that cases are not rising nearly as quickly in nursing homes as in the general population:

Cases are still rising in long-term care homes — and that spike has killed a bunch of old people — but we’ve managed so far to keep it from getting out of control in nursing homes (unlike in April).

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.) This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Our testing capacity continues to grow, but, in the past two weeks, the huge growth in cases has outpaced the growth in testing:

We are way outside the 2% “safe zone.” We are now also outside the 5% “good enough zone.” That means we are missing more cases. My case estimates at the top attempt to account for that, but, as positivity rises, we become increasingly blind to the spread of the infection, and my estimate (like everyone else’s) becomes less certain.

As I noted at the top, all this data lags by 7 days.

“But, James, that’s just cases. Lots more young people are getting it now and having mild cases that don’t require hospitalization! It’s okay!”

There’s some truth to that — but not enough. Hospitalizations are at an all-epidemic high:

The good news: we do still have plenty of hospital beds. Even at our current rapid clip, even if things get somewhat worse than they are now, we are still at least several months away from exhausting our ICU capacity. (And we have around 3,000 free non-ICU beds, too many to even put on this chart.)

The bad news: they’ve never been this full, and they’re still getting fuller at a very quick clip:

Both hospitalization charts mentions “new data.” That is because of last month’s switch in how Minnesota reports hospitalization data. As you can see, there’s a significant gap between how many beds were counted as occupied under the old method compared to the new method. But, however you slice it, it appears that our hospitals are under significant strain right now.

And, at last, we’re starting to see this wave show up in the death stats:

Even that relatively small outbreak in nursing homes is killing a lot of old people. Meanwhile, for everyone else, deaths have never been higher. We may reasonably expect this to get worse before it gets better.

I have some hope based on recent studies which show covid mortality rates falling (presumably due to better treatments, although there are other possible causes). I don’t think it’s going to get as bad as the Minnesota Model projected, but it’s not going to be fun, either.

WHY A SURGE?

I talked about this last time, and I stand by what I said then, with one small addition: back when summer started, I offered the suggestion that covid was going into remission because it transmitted poorly in summer weather, especially with most Minnesotans interacting outdoors.

I can’t help noticing that our spike arrived just as fall settled in, the weather chilled, and people started to move inside.

WHAT TO DO

These waves do not last forever. That’s one thing we’ve learned in the past few months. Arizona had a horrible six weeks early in summer… and then it ended. Louisiana got it really bad later in summer, again for about six weeks… and then it got better. California’s ugly period lasted about eight weeks… but now it’s better. So you don’t need to cancel your life, but it’s time to take extra precautions.

In other “it won’t last forever” news, we are getting close to a vaccine. And, sure, yeah, you, Joe Public, aren’t going to be offered the covid vaccine for a long, long time, because production will be slow… but, once the high-risk people get it, the worst risks from this pandemic will be over. That day seems to be only a few months away. Yes, covid is around six times deadlier than the flu, but, for most of us outside high-risk groups, six times basically nothing is still basically nothing. For this reason, I don’t anticipate a third wave of covid in Minnesota; I expect the pandemic to be (for all intents and purposes) over before another wave has time to build.

But, here and now, this thing is still raging through a population full of high-risk people, faster than it ever has before. You should take sensible precautions, especially if you or someone close to you is high-risk. If you were planning a big family Thanksgiving with 50 people, this is a good time to figure out how to get that number down to 10. You should keep socializing outdoors when possible, even as it gets chilly. (Maybe a good time to buy an outdoor heater?) We’re still planning to do trick-or-treating, but I’d be remiss if I didn’t tell you that the CDC advises against it. And we’ll be masked up and maintaining distance from people — our spooky candy chute is all ready to go, so we don’t have to get close to the little tykes.

I’ve been taking my family to (short, daily, songless) Mass lately, and advising others in my life to do the same. Now that we’re in a wave, I think I will advise higher-risk people to head back to TV Mass for a few weeks, and others to redouble their social distancing efforts.

I won’t pull my daughter out of school — they’ve done a wonderful job with social distancing, and the costs of remote learning are very large — but, as soon as the school gets a case (which seems inevitable at this rate), I’ll cut ties with my parents and everybody else for a couple of weeks.

Basically, I’m taking everything from the “Acceptably Risky” column in my risk table from a few months ago and moving it to the “Too Risky” column until this wave is over.

The government may very well take additional action, and that action may very well be justified. I wouldn’t eat indoors at a restaurant right now if you paid me. I would not be at all surprised to see eateries closed again for indoor seating, especially if the federal government passes (much-needed) debt-financed relief for small businesses. Continue supporting your local restaurants, especially non-chains (but many chains are hurting, too!) through generous use of the take-out window and ridiculously nice tips.

I do not anticipate another lockdown, because evidence does not seem to indicate that lockdowns are very effective at slowing the spread. (I’ll blog about that later, possibly when this is all over. As an April/May lockdown supporter, I need to publicly come to grips with what I was wrong about.) Broad-spectrum lockdowns, at this point, appear to be more about political theatre than about public health. I don’t see Gov. Walz doing that in Minnesota at this time. This is a blue state, but it ain’t New York.

SELECTED GROWTH RATES

Average week-over-week growth in estimated cases, July 30th-Aug 6th: 3%

Sep 4th-10th: -9%

Sep 11th-17th: -15%

Sep 18th-24th: 50%

Sep 25th-Oct 1st: 29%

Oct 2nd-Oct 8th: -6%

Oct 9th-Oct 15th: 26%

Oct 16th-Oct 22nd (preliminary; will grow): 29%

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago. And I finally uploaded my giant Excel worksheet to Google Sheets so you can see the raw numbers and calculations I am doing. Fair warning, though: I did absolutely nothing to clean it up. It’s a mess. Please enjoy Reinhart-n-Rogoffing me.

After three months of stability, a significant covid surge is underway in Minnesota. We are a long way from our hospitalization capacity, and it may get itself under control in another week or two. Nevertheless, public health officials are undoubtedly trying to figure out how to mitigate it, because the surge could just as easily continue.

DETAIL:

The chart really tells the story here. A surge that looked like lots of other false alarms started in the third week of September… but then it kept going into the fourth week of September and the first couple days of October, with a small (but by no means certain) downswing just as the data runs out:

This data lags by six days. “LTC” is short for “long-term care” (basically, nursing homes). The above estimate is for non-LTC cases. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.)

So, for people who don’t live in long term care, case counts are approaching the heights they reached in May, at our first peak.

Fortunately, LTCs still seem to be doing fine. Since the most vulnerable people live in LTCs and the most devastating spring outbreaks happened there, this is very good news.

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Despite that giant jump in case counts, test positivity has held pretty close to 5%:

This is because our testing has increased as cases have increased. In the past 14 days, Minnesota has conducted about 23,000 tests per day, on average. For those of you who remember back in April/early May, when they couldn’t break the 10k barrier, it’s pretty cool that we have this many tests now.

Lower positivity rates = fewer undetected cases = more accurate and higher official counts = lower actual cases. The fact that we’ve been able to hold the line on testing even in the middle of a surge is really quite helpful. On the other hand, it has to be worrying to the Minnesota Department of Health that a testing surge this big was only barely large enough to hold the line.

As I noted at the top, all this data lags by 6 days.

The start of the surge three weeks ago coincided with a surge in hospitalizations. This is somewhat surprising, because you would normally expect to see a surge in hospitalizations a week or two after the surge in cases. Instead, the surge starts in both datasets around September 17th:

The header on that hospitalization chart mentions “new data.” That is because, this week, Minnesota switched to a new data reporting system for COVID hospitalizations. It’s pretty slick.

The main new feature (for me) is new information about non-covid cases in hospitals, which allows us to calculate a daily surge capacity (instead of using the crude surge capacity estimates I was using before).

On the other hand, all the numbers are slightly different, probably due to differences in reporting methods. A number of ICU hospitalizations have apparently been reclassified as non-ICU hospitalizations.

However you slice it, though, it’s clear hospitalizations are up quite sharply:

We really do appear to suddenly be back in the vicinity of Minnesota’s first peak, back in early May… except that we aren’t seeing a whole lot of people drop dead.

The reason we aren’t seeing as many deaths seems to be that we’ve done a better job protecting the elderly, particularly those in nursing homes. We should also give some credit to improved covid treatments (thanks, doctors, nurses, and Big Pharma!).

Nevertheless, deaths are ticking up a bit. Hopefully that will soon reverse. However, a big surge in hospitalizations doesn’t bode well.

WHY A SURGE?

Oh, heck if I know. I can fiddle with charts, but covid has defied smarter men than me plenty of times.

Statewide voluntary social distancing levels have returned to pre-covid levels outside the Metro. That could certainly help it spread, and might explain why the Health Department is talking so much about rural hotspots. But could this alone cause that drastic increase? Why the late-September inflection point? And the Health Department is reporting cases rising statewide.

I hear lots of states around us are having bad times of it. We have a lot of commerce and other intercourse with our neighboring states, so, if they get sick, we’re likely to catch it from them. But I lack the werewithal to parse through all the journalism and find out just how bad it is in Wisconsin, Iowa, and the Dakotas; tracking Minnesota is hard enough.

The timing of the surge makes me look at school openings with a lot of suspicion. That bums me out, since I’ve been high on school openings. And I mostly still am: there seems to be not much evidence of schools causing surges like this one, school-linked Minnesota covid outbreaks are few and far between, and the National COVID-19 School Response Dashboard (run by School Principals associations) shows no significant differences in infection rates between distance-learning and in-person schools (nationally).

Still: cases started surging a few weeks after the school year started. We’d be silly not to examine whether school is a contributing factor.

Whatever the cause of the surge, it’s happening, and there doesn’t seem to be much we can do about it at the moment other than keep ourselves safe and don’t take dumb risks! Pretty much the usual!

SELECTED GROWTH RATES

Average week-over-week growth in estimated cases, July 30th-Aug 6th: 3%

Sep 4th-10th: -9%

Sep 11th-17th: -15%

Sep 18th-24th: 50%

Sep 25th-Oct 1st: 28%

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago. And I finally uploaded my giant Excel worksheet to Google Sheets so you can see the raw numbers and calculations I am doing. Fair warning, though: I did absolutely nothing to clean it up. It’s a mess. Please enjoy Reinhart-n-Rogoffing me.

As we enter a third straight month of a stable, low plateau in covid cases, it seems safe to say that what Minnesotans are doing to combat covid-19 is working, and we should continue doing it. We should also use this period of stability to experiment with reopening the state in more ways — starting with the return to school.

THE RETURN TO SCHOOL:

It is not clear to me why schools in my county have not resumed in-person instruction for at least elementary students, given the overall stability of the state’s case counts and the success other school systems have had with reopening (including our own local Catholic schools, now open for three weeks without disaster). Believe it or not, my six-year-old’s class has been really good about masking responsibly.

DETAIL:

The fluctuations from two weeks ago did, indeed, revert to the average. Case levels are stable. This data lags by (in this case) six days.

“LTC” is short for “long-term care” (basically, nursing homes). The above estimate is for non-LTC cases. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.) And here is the flip side. (Note that daily figures are sometimes negative, as the state occasionally reclassifies LTC cases as non-LTC cases.)

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Some media makes hay whenever Minnesota registers its “all-time highest” daily case numbers. However, a large portion of that is thanks to our expanded testing capacity, which is way, way, way above what we had during our first peak in April/May. That means that while we may be detecting as many — or even more! — cases than we detected during the first wave, there are almost certainly far fewer actual cases around the state. You can see that expanded testing capacity doing its job here:

Lower positivity rates = fewer undetected cases = more accurate and higher official counts = lower actual cases.

As I noted at the top, all this data lags by 6 days.

How are the hospitals doing? They’re working hard, but keeping their heads above water. This data is current as of yesterday at 4 PM:

For our part, we aren’t exactly filling up hospital beds with new patients, either, which is consistent with my estimate that our overall case numbers are holding pretty steady.

Deaths are my least favorite indicator right now, but still pretty flat:

Statewide voluntary social distancing levels remain stable, and our masking compliance seems okay. (However, I still don’t see any movement in the data that seems directly attributable to the mask mandate.)

I will be doing a blog post at some point here revisiting the lockdowns in light of what we know now, but, as a little teaser: I think the principal error we made during the lockdown was overestimating the size of the role government policy can play in social distancing. Solidarity and (largely appropriate) fear appear to have provided most of the benefits of the lockdown period, not the lockdown itself, and misunderstanding this distorted the debate over the lockdownon all sides.

I GOT TESTED

I had a cough last week, so, in accordance with MDH guidelines, I got myself a covid test.

I can now strongly recommend that you not try to get a test from the so-called Urgency Room. I went there because their website promised a 20-minute wait time, whereas my more local clinic said it would be 90 minutes and my local CVS had no appointments until the following day. The Urgency Room’s “20 minute wait time” actually meant that it took 45 minutes to check me in, then they sent me to my car to wait for another 45 minutes to be seen, all while the website still reported a 20-minute wait time. Heck with that.

I got tested at CVS the next day instead. It was relatively straightforward, although I had to give the test to myself, which made me nervous, because I am not good at doing things correctly with my hands. Pretty sure I did it right, though.

My results were negative, so I lowered the state’s testing positivity rate that day. You’re welcome!

SELECTED GROWTH RATES

Average week-over-week growth in estimated cases, July 30th-Aug 6th: 3%

Aug 7th-13th: -3%

Aug 14th-20th: -6%

Aug 21st-27th: -2%

Aug 28th-Sep 3rd: 18% (that brief surge I talked about last time did turn out to be real!)

Sep 4th-10th: -9%

Date I expect Minnesota’s second peak to surpass its first peak: [not currently trending in that direction].

Please note that predictions I have made about this pandemic have had some hits and some very notable misses.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago.

But recent news reports have indicated a covid “surge” underway in Minnesota, and this seemed like a good reason to spend an hour updating my spreadsheets to see where we stand.

OVERALL:

There is no evident surge in Minnesota covid cases right now. Upward trends in confirmed case counts are largely due to increased testing, which is actually good news. Estimated cases are flat, hospitalizations are flat, deaths are flat.

Don’t get careless, because people are still catching covid and dying of it, and carelessness is how we get superspreading and actual spikes. And don’t assume these numbers will stay flat forever. But my recommendations largely haven’t changed.

THE RETURN TO SCHOOL:

I sent my eldest daughter to First Grade this week, full-time in-person, at a local Catholic school. They have been very creative about ways to ensure social distancing. I feel quite safe. However, as a precaution, I have stopped seeing my at-risk parents for the next few weeks while we see whether we actually are safe.

I’m much less confident that colleges and universities will be able to stay open for long, because their educational model simply doesn’t allow the level of social distancing my daughter’s First Grade class has achieved. Local colleges and universities have been ludicrously innovative in some ways (frustratingly pig-headed in others), but I’m not sure anything can keep them running. It will certainly be interesting to see which colleges succeed at in-person education over the next ten weeks or so.

DETAIL:

Our estimated case count remains largely flat. There’s a hint at a small rise right at the end of this data, and preliminary data for next week suggests that this rise is real and we’ll be seeing it next week. However, it’s no “second wave,” and we’ve had plenty of fluctuations of this sort over the past six weeks which soon reverted to the mean. We’ll keep a close eye on this, but, as of right now, it shouldn’t be a major concern for average Minnesotans. This data lags by (in this case) six days.

“LTC” is short for “long-term care” (basically, nursing homes). The above estimate is for non-LTC cases.

Nursing homes have seen a definite rise in cases over the past few weeks, possibly as residents and staff begin to relax tight restrictions from earlier in the pandemic… but they’re still in quite good shape compared to, say, May. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.)

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Some media continues to make hay about the fact that Minnesota is now registering its all-time highest daily case numbers. That is true! But a large portion of that is thanks to our expanded testing capacity, which is way, way, way above what we had during our first peak in April/May. That means that while we may be detecting as many — or even more! — cases than we detected during the first wave, there are almost certainly far fewer actual cases around the state. You can see that expanded testing capacity doing its job here:

Lower positivity rates = fewer undetected cases = more accurate and higher official counts = lower actual cases.

As I noted at the top, all this data lags by 6 days. Preliminary data for the most recent 6 days looks somewhat worse, but not worse enough to justify alarm or school closures or anything of that nature.

How are the hospitals doing? They’re working hard, but keeping their heads above water. This data is current as of yesterday at 4 PM:

For our part, we aren’t exactly filling up hospital beds with new patients, either, which is consistent with my estimate that our overall case numbers are holding pretty steady.

I finally redid my deaths chart to show data and trends better, and… well, it shows deaths are pretty flat, with the expected fluctuations when dealing with small numbers. Not sure what else to tell you here:

I’ve been toying with a theory for a few of these updates that Minnesota’s plateau is partly thanks to a spontaneous rise in social distancing, as suggested by the state’s “change in distance traveled” figures. I told you in August that the data lagged by so long that we wouldn’t know the truth until September. Did my theory bear out? Eh, kinda. As our case growth plateaued, distance traveled fell… then ended up rising again, then falling again, and ended up more or less plateauing.

It’s plausible that Gov. Walz’s mask mandate is contributing to this, as it’s now had enough time to “work,” but I don’t see any movement in the data that seems directly attributable to the mandate.

AGAINST CONVENIENT LIES

Nationwide deaths from covid-19: 176,624, +/- maybe 20% due to under- or over-reporting. There are definitely some people who died of car crashes (and under other, less black-and-white, circumstances) who got counted as covid deaths. There are also definitely some people who died of covid and didn’t get noticed, but who show up in the excess death figures (some other excess deaths were people killed indirectly by the lockdowns). Those undercounts and overcounts likely cancel each other out to some extent.

Please do not buy into the ridiculous fake news this week that “only 6% of those people ACTUALLY died” of covid; it’s silly and falls apart the second you open the CDC report the claim comes from and read the related data table in context. I’m going to resort to linking an Instagram doctor who can debunk this one visually, but, seriously, read the report yourself; it’s written in ordinary English. There’s no need for anyone to be dumb about this. Being dumb about it is a choice.

Meanwhile, please stop worrying (read: concern trolling) about Sturgis. It was always difficult to make the case that Sturgis was going to be more dangerous than the mass protests and rioting. Those riots yielded (1) no discernible increase in cases or deaths, and (2) widespread establishment denunciation of the suggestion that the protests could possibly have any bad effects ever (even when those effects were visible on camera). Suddenly rule #2 went away for Sturgis, apparently because it was going to be populated by all those mean, nasty conservatives, and some argued, idiosyncratically, that Sturgis-goers weren’t social distancing as well as looters in a burning Target.

Sturgis ended two and a half weeks ago. If it were going to cause a spike in Minnesota, we’d have seen it by now. We haven’t. If you only condemn or “worry about” mass gatherings with people from the Red Tribe, while giving a pass to or participating in Blue Tribe mass gatherings (I SEE YOU STEVE SACK, YOU JERK), you’re not only making America’s covid problem worse, but you’re one of the causes of that civil war I was writing about for the past couple weeks.

SELECTED GROWTH RATES

Average week-over-week growth in estimated cases, July 30th-Aug 6th: 3%

Aug 7th-13th: -3%

Aug 14th-20th: -6%

Aug 21st-27th: -2%

Date I expect Minnesota’s second peak to surpass its first peak: [not currently trending in that direction].

Please note that predictions I have made about this pandemic have had some hits and some very notable misses.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago.

Reminder: I break these out between long-term care (LTC) and non-LTC residents where possible. Furthermore, I focus on non-LTC because most of the people reading this are not LTC residents, and most of my advice is not applicable to them.

OVERALL:

Minnesota’s covid epidemic is not getting better. But it is also not getting worse. Given what a good position we’ve been in for the past few months, “not getting worse” is really, really good. Continue enjoying your summer safely.

DETAIL:

Our estimated case count over the past almost-month has been remarkably stable. It’s not going down like in, say, Sweden. It’s not going up like in, say, France. (Yeah, Sweden’s getting better, France worse.) Our numbers are just holding steady. This data lags by (in this case) six days.

“LTC” is short for “long-term care” (basically, nursing homes). The above estimate is for non-LTC cases.

But nursing homes also remain stable, with covid nearly but not quite eradicated in LTCFs. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.)

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but it should be accurate enough for us to trust the trendline.

Here is the raw, official state data I use to build these estimates:

Some media has made some hay about this chart. They have observed that Minnesota is now registering its all-time highest daily case numbers. That is true! But a large portion of that (not all, but lots) is thanks to our expanded testing capacity, which is way, way, way above what we had during our first peak in April/May. You can see that expanded capacity doing its job here:

Lower positivity rates = fewer undetected cases = more accurate and higher official counts = lower actual cases.

As I noted at the top, all this data lags by 6 days. However, I’ve been eyeballing the preliminary data for this week (as I always do), and we seem to be on track for a continuing plateau.

How are the hospitals doing? They’re working hard, but keeping their heads above water. This data is current as of yesterday at 4 PM:

The new hospitalizations seem to be reaching their own plateau after a steady rise — exactly what we would expect to see a few weeks after new cases plateau. Critically, we are still far short of our emergency capacity, which I’ve added to the Total Hospitalized chart this week (thanks to the reader who suggested it). I’ve had to double the height of the chart just to fit the capacity lines:

I still don’t really know quite what non-LTC deaths are doing, but they have definitely gone up since the start of August. That’s not surprising, given the increased hospital demand. (It’s actually surprising it’s taken this long, as I’ve mentioned in previous reports.) Death counts are still not very high.

Why’s this happening? Not sure. There still hasn’t been time for Gov. Walz’s mask mandate to have an impact on most of these figures.

I remain inclined to give some credit to the apparently spontaneous increase in Minnesotan social distancing (in terms of miles traveled per day) that started 4th of July weekend. That data lags by weeks, so we won’t know for sure until September, but that decline in Minnesotan travel does seem to line up nicely with covid’s slowing growth. I’ll keep tabs.

How long will this plateau last? Not sure. As I’ve said before, covid outbreaks are part opportunity and part luck. It seems unlikely that this delicate stability can last for long, but it’s difficult to predict whether the curve will bend upward or downward next.

IN CLOSING, SOME OTHER NUMBERS:

Some of these numbers have changed slightly from previous weeks. This is due to new data, mostly old tests that didn’t get reported in a timely manner:

Average week-over-week growth in est. cases, June 25th-July 1st: 32%

Average week-over-week growth in estimated cases, July 2-8th: 32%

Average week-over-week growth in estimated cases, July 9-15th: 36%

Average week-over-week growth in estimated cases, July 16th-22nd: 18%

Average week-over-week growth in estimated cases, July 23rd-29th: -4%

Average week-over-week growth in estimated cases, July 30th-Aug 6th: 2%

Date I expect Minnesota’s second peak to surpass its first peak: [not currently trending in that direction].

Please note that predictions I have made about this pandemic have had some hits and some very notable misses.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few months ago.

Reminder: I break these out between long-term care (LTC) and non-LTC residents where possible. Furthermore, I focus on non-LTC because most of the people reading this are not LTC residents, and most of my advice is not applicable to them.

For about the last month, Minnesota’s covid epidemic has been trending in the direction of a “second peak”. However, this week, that trend stalled out. By any measure, cases this week climbed much less than they did in recent weeks.

This could be just a short blip, but it could also be the start of a second decline. Either way, it’s very welcome news, because it means Minnesota will remain in a covid “valley” for at least a little while longer. Continue enjoying your summer safely.

DETAIL:

As you can see, Minnesota’s case counts plateaued this week. That could be just a “pause” like the one we had in late May, but it could plausibly be the summit of our “second peak”… which would be great news, because this “second peak” was barely anything. This data lags by (in this case) six days.

“LTC” is short for “long-term care” (basically, nursing homes). These are non-LTC cases.

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but accurate enough for us to trust the trendline.

Good news continues in nursing homes. My estimates show the same thing the state data shows: that new cases have nearly vanished from long-term care facilities. (This estimate assumes test positivity rates between LTCFs and the general population are more or less the same.)

Here is the raw, official state data I use to build these estimates:

Test positivity rates (higher rates = more undetected cases = higher case estimates) have also plateaued, which is good news. For a while there, it looked like they were just going to keep climbing at 1% per week, but, instead, it has frozen up for the past… almost two weeks! Here, let’s zoom in on this:

Test positivity rates for this week aren’t final yet, but preliminary data shows they’re stable, maybe going down a little. We’re not yet streets ahead, but neither are we streets behind.

Back to bad news: hospitalizations are going up. This data is current as of yesterday at 4 PM:

…but this is to be expected. Cases went up for weeks. Hospitalizations go up a week or two after cases go up, and deaths go up a week or two after hospitalizations go up. If cases remain stable or start to decline, then I would expect hospitalizations to do the same in a week or two… and, happily, we have plenty of hospital bed capacity at this point, because our second wave is nowhere near the Red Zone.

Wait, did I say that deaths go up a week or two after hospitalizations go up? Am I sure about that?

Last week, I thought the jump in non-LTC deaths was the leading edge of the expected increase in death toll after a month of increasing cases. So far, that hasn’t panned out. Indeed, as you can see, the non-LTC death toll jumped down this week just as aggressively as it jumped up last week.

Can we thank Gov. Walz’s new mask mandate for this? Absolutely not. I think a mask mandate is probably good, but the mandate was imposed on July 25th. We don’t even have full data from July 25th yet. Even if it were immediately adopted perfectly by everyone, the mask mandate would take a few weeks to have any visible effect on our numbers. (Besides, remember, we still don’t know how effective cloth face coverings really are, so I’m not getting my hopes up.)

A slightly more promising explanation for this week’s covid stall-out is an increase in social distancing statewide. We’re still waiting for complete data, and it may not pan out… but there was an apparently spontaneous increase in Minnesotan social distancing (in terms of miles traveled per day) starting 4th of July weekend. The timing is just about right to explain a slowdown today. My reservation is that it’s not a very big increase, nor do I see it in other data, but I’ll keep tabs.

So covid growth stalled this week. Why? Not clear. Hopefully it keeps doing that, though.

IN CLOSING, SOME OTHER NUMBERS:

Average week-over-week growth in est. cases, June 25th-July 1st: 32%

Average week-over-week growth in estimated cases, July 2-8th: 31%

Average week-over-week growth in estimated cases, July 9-15th: 32%

Average week-over-week growth in estimated cases, July 16th-22nd: 14%

Date I expect Minnesota’s second peak to surpass its first peak: [not currently trending in that direction].

Please note that predictions I have made about this pandemic have had some hits and some very notable misses.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Takes A Breather a few weeks ago.

Reminder: I break these out between long-term care (LTC) and non-LTC residents where possible. Furthermore, I focus on non-LTC because most of the people reading this are not LTC residents, and most of my advice is not applicable to them.

UPDATE 13 August 2020: This post originally posted the week-over-week growth rates incorrectly. Instead of 32%, 31%, etc., the post said case growth was 132%, 131%, etc.. This incorrectly implied that cases were more than doubling every week! Cases in the last week of June were indeed 132% the size of the cases the week before, on average… but that means they only actually grew by 32%. Sorry. Dumb error.

In Minnesota, covid is climbing steadily back toward a new peak. Unless something about Minnesotan behavior changes within the next week or two, we are likely to surpass our April/May first peak in about a month.

However, for now, we aren’t there yet. Continue enjoying your summer safely. Squeeze in an outdoor cookout with 10 or fewer of your more responsible friends, go eat on the patio of that restaurant you love, continue not going to bars, and mentally prepare for the reality that life will likely become more restricted again as fall arrives.

A mask mandate is probably coming from Gov. Walz, and that’s probably good.

DETAIL:

As for the past several weeks, case counts continue to trend upward, but remain well below April/May peaks. This data lags by (in this case) six days.

“LTC” is short for “long-term care” (basically, nursing homes). These are non-LTC cases. As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but accurate enough for us to trust the trendline.

Here’s some good news, though: my estimates show that the epidemic has gone to a very low ebb in long-term care homes. It was killing grannies left and right just a few months ago, but now we’re seeing very few cases, even with expanded testing:

Here is the raw data I use to build these estimates:

This positivity chart is also okay news. While our test positivity rates are getting higher (indicating that we are missing more cases), it appears that we are still detecting most cases of covid. I worried we would quickly lose control of disease surveillance when the second wave came. We may still, but we are holding on for now.

Back to bad news: hospitalizations are going up. This data is current as of yesterday at 4 PM:

It seems the laws of epidemic arithmetic are still holding, despite my hopes: more cases in the past few weeks has led to more hospitalizations today. More hospitalizations today means more deaths in a few weeks. Perhaps the deaths will be fewer, because more young people are getting it, but deaths are unlikely to stay in decline.

For now, though, deaths are indeed down, and we can continue to celebrate that… even as we see what looks like the beginnings of a trend in the other direction:

IN CLOSING, SOME OTHER NUMBERS:

Average week-over-week growth in estimated cases, July 2-8th: 31%

Average week-over-week growth in estimated cases, July 9-15th: 32%

Very rough estimate of pre-symptomatic/early-symptomatic Minnesotans outside LTCs, as of July 15th: 8,000

Where that number was on May 15th, at first peak: 23,500

Very roughly where I expect that number to be as of today, once we have all this week’s data in: 10,500.

Very roughly where I expect that number to be in a week: 14,000

Very roughly where I expect that number to be in two weeks: 18,000

Very roughly where I expect that number to be in three weeks: 23,500

Date I expect Minnesota’s second peak to surpass its first peak: August 19th.

Please note that predictions I have made about this pandemic have had some hits and some very notable misses. The pandemic may surprise us by slowing down again, just like it did in May, and I hope it does.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Isn’t Killing Minnesotans Like It “Should” a few weeks ago.

Reminder: I break these out between long-term care (LTC) and non-LTC residents where possible. Furthermore, I focus on non-LTC because most of the people reading this are not LTC residents, and most of my advice is not applicable to them.

UPDATE 13 August 2020: This post originally posted the week-over-week growth rates incorrectly. Instead of 32%, 31%, etc., the post said case growth was 132%, 131%, etc.. This incorrectly implied that cases were more than doubling every week! Cases in the last week of June were indeed 132% the size of the cases the week before, on average… but that means they only actually grew by 32%. Sorry. Dumb error.

In Minnesota, covid is still in a valley, not a peak, but there are growing signs of a second wave. Fortunately, the new wave is coming much more slowly than the last one, and it may still dissipate before developing into much of anything. Minnesota is by no means in the dire straits that some Southern states are currently in, nor is it likely to find itself in the even-more-dire straits that some of the Northeastern states experienced in April and May.

Continue enjoying your summer safely, with an emphasis on outdoor activities and small gatherings — but begin mentally preparing for a possible shift back toward a more restrictive outlook.

A mask mandate is probably coming from Gov. Walz, and that’s probably good.

DETAIL:

As last week, case counts continue to trend upward, but remain far below April/May peaks. This data lags by (in this case) six days.

As I state every week, my daily estimate of “actual new cases” is derived by taking the current 7-day average positivity rate, dividing it by 2% to yield a multiplication factor (minimum 1.0), and multiplying the officially reported non-LTC cases by that factor. This is crude enough that, when positivity is significantly above 2%, the precise numbers may be way off… but accurate enough for us to trust the trendline.

The trendline is currently heading up after several weeks of steep decline, and that is bumming me out. On the graph, it doesn’t look like much, because the April/May peak was so much higher… but my estimate indicates that Minnesota’s case count bottomed out three weeks ago, on June 21st, and that the number of cases jumped 50% between June 24th and July 1st. That’s slower than the extremely fast spread we saw in late April… but, if we keep on this pace, we’ll get back to May/April peaks in fairly short order. Hopefully we won’t.

Here is the raw data I use to build this estimate:

I take great comfort from the fact that test positivity rates are still holding under 4%. The ideal is 2%, but 5% is the WHO’s official guideline. That means, even though covid is spreading faster again in Minnesota, it’s likely we’re still catching the majority of our covid cases. States like Arizona, with positivity rate over 25%, are likely catching fewer than 1-in-20 of their actual cases — which makes their high reported case loads all the more alarming.

Unfortunately, this is probably the last week we’ll be able to say that. Preliminary data for July 7th-July 12th is not included in these charts, because it’s preliminary… but I can already see that test positivity is going to be over 4% on at least two days next week. That means there will be a bigger gap between “officially reported” new cases and actual new cases.

Good news, though: hospitalizations are still flat, as they have been for nearly a month. Unfortunately, there are indications in the Southern states currently experiencing an outbreak that more cases is likely to lead to more hospitalizations, so we may see this number tick up in the next week or two — but we haven’t yet, and I’ll take that win.

This data is current as of yesterday at 4 PM:

Deaths, happily, continue to trend downward. This still appears to be simply because we’ve had fewer cases. Now that cases are going back up, we should watch out for deaths ticking up in 2-3 weeks… but, hopefully, the uptick will be just as slow as the growth in cases.

All data is either directly from here or derived from data from here: Minnesota Department of Health: Situation Update for Covid-19 . I went into a little more detail on some of these data in my post Covid Isn’t Killing Minnesotans Like It “Should” a few weeks ago.

Reminder: I break these out between long-term care (LTC) and non-LTC residents where possible. Furthermore, I focus on non-LTC because most of the people reading this are not LTC residents, and most of my advice is not applicable to them.