What “Overreaction” Means

Every death is a tragedy, at least on some level. And everybody dies.

Here, for this analysis, we will recognize upfront that tragic deaths have occured due to COVID 19. Nobody sane is happy about that, and I and everyone of good will has nothing but sympathy for those suffering such loses.

Having acknowledged that, we will set aside our perfectly healthy and human emotional reactions to take a look at what should reasonably be done.

Death is a petty, fickle thing. The most innocent activities can get you killed. For example, every year:

• about 3,500 Americans a year drown in their backyard pools.
• about 2,700 Americans die in motorcycle accidents.
• about 36,000 Americans die of falls.

And on and on. About 4o,000 Americans die every year in automobile accidents, not counting pedestrians who get run over; commercial fishing and raising cattle are two other comparatively dangerous occupations, with disproportionately high accidental death rates.

We all know this, or at least used to. Used to be, we accepted reasonable risk and the inevitable tragedies as part of the price of freedom. What a ‘reasonable’ level of risk is varies, but, generally speaking, if we felt (it’s mostly feelings – we’re people, after all, not computers) the risk was either unavoidable or at least partly something we can control, we were generally good with it. Thus, backyard pools can be fenced, children can be watched; motorcyclists can be very careful and wear helmets; we throw a non-slip rubber mat in the shower.

Something else we used to understand: bad stuff happens. The neighbor kid will climb pool fences, and sometimes drown; the careful, law-abiding motorcyclist will die when somebody runs a red light or makes a crazy lane change; and grandma is going to die of something, and falling and whacking her head is a possibility no matter what we do.

In moments of clarity and humility, we accept that all we can do is take prudent precautions, that bad stuff will happen no matter what steps we take. We recognized we’d end up living lives of paranoia and slavery if we can’t accept a good level of risk.

Now? Check this statement out:

“The truth is, there is no such thing as an accident. We know what to do to save lives, but as a nation, we have not consistently prioritized safety at work, at home and on the road.”

Accidental Deaths Hit Highest Number in Recorded U.S. History the National Safety Council, 2018.

“…no such thing as an accident.” I don’t know whether to prescribe some basic philosophy, psychological counseling, or serotonin reuptake inhibitors to treat such an insane misapprehension of reality. On a logical basis, that claim is sheer fantasy. As propaganda, however, it’s genius. Under this rubric, LITERALLY EVERY BAD THING THAT HAPPENS IS SOMEBODY’S FAULT. “We” know how to prevent every bad thing. If every bad thing has not been prevented, some other “we” – certainly not us! – has failed! Implied: “we” – first “we” again – need MORE POWER in order to properly “prioritize safety”.

The National Safety Council is a hoary institution, dating back to the Wilson administration, a time when our betters were pretty public about their conviction that they could solve all problems if we little people would just let them, so it would be tempting to think this view represents some sort of outdated extreme. Alas, this is exactly the view expressed by Governor Cuomo and others to justify anti-COVID 19 measures that would have been considered crazy a generation (or, perhaps, one administration) ago.

And measures not remotely justified by the information we had available at each point of this fiasco.

How did we get here? A big part was played by models that predicted millions of dead and a collapse of health care systems under the strain of tens of millions of critically sick. Those claims were made not just on extremely unlikely worse-case scenario assumptions, but in the face of known contradictory evidence. Yet, our well-schooled, by which I mean functionally innumerate and alogical, population continues to fall for it. Let’s take a look at models at a high level.

The first thing a pro, even so meager a one as myself, wants to know when constructing a model is: what does the data look like? Is it clean and consistent? Is it reported in an orderly, reliable manner? Is it clearly defined? Because your model – your predictions – will never be any more accurate or reliable than the data they are built upon. This is a truism: garbage in, garbage out.

So, something clear to anyone who cared to look from Day 1 of this mess: the data is garbage. Teasing anything meaningful out of this inconsistent, incoherent pile of junk was always going to be a challenge, and would always be highly uncertain.

Simple questions a pro would ask:

• What constitutes a ‘death’?
• What constitutes a ‘case’?
• Are these defined and reported consistently from place to place and over time?

And slightly – and only slightly – more sophisticated:

• Anybody have an interest in the numbers going one way or the other?

Meaning: is it in anyone’s interest to over or under report any of this?

Let’s take it one by one:

Death: Completely inconsistent or undefined from place to place and over time. In some places, such as Britain and New York, a ‘death’ for COVID 19 purposes means: person was determined to have COVID 19 at the time of death. ‘Determined’ means that the person who fills out the death certificate puts COVID 19 anywhere on it. It does not mean: the deceased tested positive. It could mean: the deceased had symptoms consistent with COVID 19, for example, a fever and a cough. Sure, that’s also consistent with the flu or a common cold, but whatever. ‘Death’ as used in the reports, at least from some of the places reporting, literally means that an asymptomatic person who tests positive and then gets run over by a lorry on his way home counts as a COVID 19 death.

That may seem far fetched, but this next example certainly isn’t: yesterday on Twitter, people were mourning the death of a man well known in certain geeky circles. His death was classified, and people were referring to it, as a COVID 19 death – that’s what the official report said.

He was 79, almost 80, and had been in a nursing home for the last few months after having suffered a massive stroke. Now, I gather few healthy people spend much time in nursing homes. Unfortunately, I have, because my father and two older sisters spent time in nursing homes prior to their deaths, and I have had a number of other occasions over the years to visit a variety of such facilities.

The majority of people in these places were put there to die. If they are very lucky, some family member will get them released into hospice care so that they don’t have to die alone in a place stinking of urine. But few are there to get better; at least, if they are there to recover, they are not generally there for long.

I’d be happy to be contradicted by medical pros or somebody with experience working is nursing homes, but here’s what I suspect: if you’re pushing 80 and have a massive stroke and don’t bounce back within 2 months, the only way you’re leaving that nursing home is feet first. Barring a miracle, of course.

No rational person would count that death as due to COVID 19. He died in the aftermath of a massive stroke. The stroke killed him. The most COVID 19 might have done was speed up his death some. But the flu or a cold would have done as much, most likely.

Now put this anecdote together with some other bits of information that have been generally available from the beginning: that people who were old, or with preexisting medical conditions, or both, were disproportionately killed by COVID 19. That COVID 19 broke (a small portion of) the Italian medical system when it ripped through nursing homes in Lombardy. That the average age of death was 81 years old.

The ability of even the most vigorous 80 year old to fight off and recover from illness is seriously impaired in almost every case. Now, start with what in almost every other case is a very minor illness – very few people who aren’t old or already sick so much as develop serious symptoms – and you’re in for a lot of deaths, provided you count every old or sick person who catches COVID 19 and dies from whatever causes as a death FROM COVID 19.

If I were building this model, as a first-pass data clean-up, I’d exclude ALL deaths of people over, say, 65 and ALL deaths of people with preexisting conditions. THEN, if I could figure out a reasonable way to do it, add back in some small fraction of those deaths on the assumption that some, at least, would have survived, say, a year and a day if they had not caught COVID 19.

That would be the professional, responsible thing to do. The number of death, after such a merely prudent clean-up, would be much smaller that currently reported. Using the raw data would be irresponsible, as it clearly misrepresents reality.

Let’s take a look at that. Back in March, when we were losing our minds:

That likely explains why although older Americans represented 31% of the cases, they accounted for 45% of hospitalizations, 53% of ICU admissions, and 80% of deaths, the CDC reported.

STAT March 19, 2020

Ninety-five percent of New York City’s almost 200 deaths from the new coronavirus had underlying health conditions, though almost half were under the age of 75, according to data published by the city’s health department on Tuesday.

Bloomberg, March 25, 2020

So, we’ve known for weeks now that only 5% of the deaths associated with COVID 19 were otherwise healthy people, and that 80% were people over 65. But rather than take prudent precautions around the sick and elderly, we shut down the economy.

If I were building this model, I’d use that 5% – that’s 1,200 deaths at the moment – as a baseline number, then add back to it some percentage – half? I’d need to do some research, but that seems generous – of the other deaths, under the assumption that those were people who, despite being abandoned to die in a nursing home, might still make it a year and a day if they hadn’t caught the virus. I’m being only slightly flippant here – to have sane numbers, you’d need some approach that filters out people who died of other causes while infected (or suspected of being infected) with COVID 19, to better reflect the real risks. If you’re dying anyway, COVID 19 isn’t really a risk, is it?

So, maybe I start with 13,200 deaths in my model. Seems high.

Next, let’s look at the Chinese. I don’t know how they define a COVID 19 death, and wouldn’t trust them to tell the truth anyway, but: if the Chinese, for their own selfish reasons, decided to count only deaths where COVID 19 seriously contributed to an untimely demise, excluding from the count people who were already seriously sick before they caught the virus, their reported numbers, which seem so low, might actually be accurate. At least, more accurate than the ridiculous numbers coming out of the West.

And – here’s the main point – if, and it’s certainly a big ‘if,’ the Chinese were in fact reporting in this manner, they would be reporting COVID 19 deaths in a way that’s much more like what a reasonable, normal person would consider a COVID 19 death.

Now consider the ‘consistent over time’ issue. New York changed to the ‘count every death where COVID 19 shows up anywhere on the death cert’ method a few days ago. They got what must have been a gratify uptick in deaths – gratifying, because why else would you do such a ridiculous thing? In this current panic, was there any reason to think deaths in New York were being underreported? Really?

Imagine if you fanned a panic and things didn’t turn out too bad, so that, come election time, your opponents can point out that your panic mongering was directly responsible for millions of lost jobs and falling incomes and destroyed businesses. Better do something about that. Today, New York ‘found’ some more deaths:

New York City today has reported 3,778 additional deaths that have occurred since March 11 and have been classified as “probable,” defined as follows: “decedent […] had no known positive laboratory test for SARS-CoV-2 (COVID-19) but the death certificate lists as a cause of death “COVID-19” or an equivalent” [source]. We will add these to the New York State total as soon as it is determined whether the historical distribution can be obtained

Worldometers

Wow, nick of time. First, you change the requirement so that every death with COVID 19 anywhere on the death certificate is classified as a COVID 19 death, then you mine the old death certificates to find any where they listed COVID 19 but didn’t list it as a significant cause, so you can then throw them into the count. That this happened just as the totals for New York and the US were turning downward is one of those amazing coincidences that just keep popping up.

The US in general changed its method of counting COVID 19 deaths a week or so ago, to be more ‘generous.’ US counts are therefore inconsistent over time. Even this ignores the enormous amount of discretion local officials – doctors and coroners – have in what they put on a death certificate, especially in cases where cause of death is complicated, as it always is when preexisting conditions are involved. Even under the same rules, what counts as a COVID 19 death in Peoria may not be what counts as one in Santa Barbara, nor one on Tuesday versus one on Saturday, even in the same location. Lots of uncertainty here.

Finally, add in the political, professional, and social pressures. We’ve just burned a couple of trillion dollars in economic activity, gotten millions laid off and many millions more are seeing income reductions – you want to be the guy who calls ‘Ooopsie!’ on all this? Or do you double down? Do you want to be the one doctor who refuses to play games, or you just going to follow orders and put what they tell you to put on the death cert? New York and the recent redefinition of COVID 19 deaths shows us exactly how that’s going to work.

Similar problems exist everywhere. France, for example, had a spike in deaths when they ‘discovered’ a whole bunch of people who died in nursing homes had not been counted. France, by the way, is experiencing a ton of social unrest at the moment; Macron would like it to go away. Maybe, just maybe, he might like this whole thing to drag on a bit?

As of this moment, 125K deaths from COVID 19 have been reported worldwide. There is no reason to believe this number is remotely accurate by any reasonable definition of ‘death caused by COVID 19’ and every reason to think it wildly overstates the number of such deaths. This problem has been obvious from Day 1.

Next, we have similar issues with Cases, which I will not belabor again here. The real issues with cases is that people still insist on thinking cases = number of infected people. With rare exceptions, case counts will NOT INCLUDE asymptomatic people, for the simple reason that, so far, almost all test are of people suspected of having the virus. Somebody with no symptoms is most likely not getting tested. In the few general semi-random tests done so far, around 50% of infections are asymptomatic. Further, 96% of cases show mild symptoms. Again, many people with mild symptoms are likely not getting tested.

Again, all I’ve done here is ask the basic questions any competent model builder would ask about the data, and discovered that cases must significantly understate the number of infections in the world. Yet people who should know better keep insisting cases and people infected are the same. For example:

Worldometers

Must harp on this one more time:

Case Fatality Rate (CFR) = number of people identified as COVID 19 deaths on their death certificates / number of cases.

Fatality Rate = number of people who died from COVID 19 (however defined) / number of people infected with COVID 19.

Unless we test every human being on the planet, we will never know the true number of people infected. It was always be an educated guess. All we can reasonably know is that it must be more, and probably a lot more, than the number of cases.

Now let’s put these two things together:

Since the number of deaths has been demonstrably overstated, AND the number of people infected is demonstrably understated by the number of cases, THEN: the CFR OVERSTATES how deadly COVID 19 is. Based on my semi-educated proposed adjustments, outlined here but explained in more detail in earlier posts, I’d guess that factor is at least 4. If your CFR is 2.5%, then, the logic here shows the real fatality rate, the real risk someone who catches COVID 19 will die of it, can’t be more than 2.5/4, or 0.625% EVEN IF we make the completely insane assumption that ALL deaths WITH COVID 19 are in fact deaths FROM COVID 19. If my guesses are in the ballpark, cut those numbers in half again.

These number ignores the age/health component to risk. If you’re healthy and not old, your risk is effectively 0 – too small to mean anything in practice. Effectively, you’d be better off getting some exercise and and maybe cutting some calories than doing anything at all to avoid COVID 19. Even if you’re old, if you are otherwise healthy, your risks are tiny.

It also needs to be said that this sudden concern for the sick and elderly is ironic, by which I mean, rank hypocrisy. We routinely warehouses the sick and elderly under the care of minimum wage workers so that they are no bother to us while we wait for them to die. Again, ever been to a nursing home? For every resident who gets daily visits from loved ones, there are probably 4 who rarely, if ever, get a visitor. Ask somebody who works in one of these places. We abandon them to die, and now we’re all concerned about them dying? Give me a break. (Note: we have my 82 year old MIL living with us. She’d still be in a nursing home if we hadn’t sprung her. There are cases where this is not possible, where more care is needed than a normal family can provide. But there are plenty of cases where it would be possible, if anybody cared to do it. And nothing prevents frequent visits or at least phone calls.)

So, what would simple prudence dictate as proper actions to take in order to protect the vulnerable from COVID 19?

• If you are sick or have been exposed to COVID 19, stay away from the sick and elderly (duh).
• Similarly, if you are sick or elderly, you might want to consider avoiding crowds for a while. I say ‘consider’ because lack of human contact can damage people too, and loneliness is a big problem for many elderly people. If gramps want to accept the risk and go to the church potluck to be with his friends, who are we to say no?
• Wash your hands.

I leave the proper steps to best care of hospitalized and institutionalized people to the pros. The rest of us should get back to work.