Some known unknowns

Because there is still so much to still be learned about the spread and lethality of the novel coronavirus, it is difficult to answer the question of whether the shutdowns are necessary and, if so, how long they will remain that way.

We not only don’t know certain things that we would need to know, but what we think we know might be mistaken, a point emphasized in a recent essay in the UK Spectator by Dr. John

Lee, a retired pathologist and former consultant to the British

National Health Services (NHS).

Lee’s central argument is that the raw data we tend to pay the most attention to, on the number infected and deaths from infection, is often presented in ways that can mislead.

The first data point — the numbers infected from country to country — tells us only how many people demonstrated sufficient symptoms to receive testing and then tested positive. As Lee points out, that number may only be a small minority of those infected because many (most?) of the infected show few or no symptoms. There is, in other words, no way at this juncture to acquire an accurate estimate of how many people have been infected worldwide, even though such knowledge is crucial to our assessments.

The second data point — the number of people infected by the virus who have died in each country — fails to make the crucial distinction between deaths with the virus and deaths actually caused by it. Along these lines, even Neil Ferguson, primary author of the Imperial College study of the virus that sent shock waves throughout the world, has estimated that up to two-thirds of those who have been infected and died would have died anyway because of severe underlying conditions — they died “with” but not necessarily because of the virus.

What this means is that it is possible that both the numerator and the denominators currently being used to assess the most crucial number, fatality rates, are being interpreted in ways that dramatically increase that rate, to truly alarming levels, often above

5 percent.

Put differently, if the numerator (the number infected) is 20,000 rather than an officially reported 1,000, and the denominator (the number dying because of rather than simply testing positive for the virus) is 20 instead of 50 you get a fatality rate of 0.1 percent instead of 5 percent, a rate, in Dr. Lee’s words, “similar to seasonal flu.”

But the broader point is that the data we are collecting and relying upon to track the spread of the virus and its effects, and thereby make sound public health decisions, is incomplete and susceptible to costly misinterpretation. As Lee notes, “there is no international standard method for attributing or recording causes of death. Also, normally, most respiratory deaths never have a specific infective cause recorded, whereas at the moment we can expect all positive

covid-19 results associated with a death to be recorded.”

So lots of deaths that are primarily caused by other illnesses are now statistically being attributed to the virus if the patient had tested positive for it.

Such considerations are important because the interpretation of statistics is going to be fundamental to how we make decisions going forward. Unless we plan to hunker down for years and essentially return to the dark ages, which is what it might take to reach the point when no one is any longer at risk, we are going to have to open things back up at some point, probably sooner rather than later. And that decision will have to be made and then defended on the basis of statistical metrics reflecting reasonable cost-benefit assumptions. The claim that there is no necessary trade-off between saving the economy and saving lives is true, but not indefinitely so.

If we assume, as seems reasonable, that we should move to end the shutdowns only after the “curve” of the virus in terms of infections and fatalities has clearly begun to bend downward, how far downward does it have to actually bend before it is “safe” and how will “safe” be defined?

Although extending the shutdowns for another month or even another week will save more lives, elected officials will at some point have to accept a certain additional increment of lost lives in order to stave off economic catastrophe, while knowing that the demagogues among us will reflexively accuse them of having “blood on their hands” because they prioritized money over lives (as if the two are ever that easily distinguished).

The calculus will be as morbid as it will be necessary: Is it worth it to shut down things for another two weeks at cost “X” if it means 200,000 fatalities nationwide compared to 180,000?

If we somehow make it through the pandemic with far fewer lives lost than some of the scarier models predicted, skeptics of the shutdowns will claim that they were never really necessary in the first place, that we risked destroying the economy and much else out of panic.

But the same outcome will also allow those who emphasized the dangers of the virus from the beginning to claim that it was the shutdowns that averted disaster.