What we are learning

Some of the pieces in the virus puzzle are starting to fall into place, allowing us to draw some broad, if still tentative, conclusions.

First, there seems to be a growing belief that the virus was in this country infecting large numbers of people weeks before the first official case was diagnosed.

Second, that the number of people who have been infected across the country (the crucial denominator necessary to calculate fatality rates) is likely higher, perhaps much higher than the official estimates indicate—the much publicized Stanford study of Santa Clara County estimated that from 50 to 85 times (!!) more people were infected than assumed, while a recent study of blood donors in Stockholm found the presence of virus antibodies in more than 10 percent of those tested (10 percent of America’s population would add up to more than 30 million infected).

Third, the profile of those hardest hit by the virus seems remarkably consistent throughout this country and Europe—the elderly and/or those with serious underlying health conditions. Everyone is potentially at risk, but the young and healthy are vastly less so than the aged and unhealthy.

Fourth, if the elderly and/or sick are by far the demographic most at risk, then the geographical epicenter is clearly the metropolitan New York area. As of a week ago roughly half of the nation’s fatalities had occurred in just NYC’s five boroughs and outlying suburbs.

Those are useful things to know, particularly as we get about figuring out a way to open things back up in as safe a manner as possible, but they still leave us with some huge known unknowns.

First of these is the one that matters most—the actual fatality rate, defined as the percentage of those who contract the virus and then die because of it. Along these lines, the Stanford study might feature certain methodological flaws that limit the applicability of its conclusions to other parts of the country. Although other reports also suggest that we are seriously undercounting the number of infected, they tend to contain much lower actually infected to officially infected ratios than the Stanford team found.

Getting a better sense of these numbers is crucial because if the number of Americans infected is around 1 million and the deaths attributed to the virus (however loosely) are 50,000, then the fatality rate is what we currently see, a scary 5 percent (one in 20 who get it can be expected to die). But if 20 million Americans are actually infected rather than a million, that fatality rate drops to a much less terrifying 0.25 percent. If 30 million are infected, it drops closer to the 0.1 percent fatality rate estimated for influenza.

But the relevant point is that, until we have far more and more carefully conducted studies testing for virus antibodies, we will have no way of truly knowing, even within broad parameters, how many more people are infected than the official numbers indicate. For the sake of argument, if the real number is “merely” double the official number (rather than 20 or 30 times), then the fatality rate still lands at a level (2.5 percent) likely to cause overwhelmed health care systems and catastrophe.

All of this matters for how we, in a cost-benefit sense, assess the necessity of the shutdowns and whether or not they should continue. Only by acquiring firmer numbers on the number actually infected can we get a more sure understanding of how easily the virus spreads and how lethal it is, and therefore what approach to it makes the most sense. Presumably, that approach would be different if the virus spread easily but wasn’t particularly lethal as opposed to if it was less contagious but highly lethal.

Then there is the mystery of New York, the outlier in so many respects and the place which most clearly contradicts both the assumption of vastly more infected within the population and the lower fatality rates that flow from such an assumption.

There were always reasons to expect NYC to be harder hit than most other places—population density, status as entry point for foreign travelers, reliance upon crowded subways and other mass transit, an older population—but virtually nothing in what we know thus far about the virus explains the extreme disparities in numbers between it and other at least somewhat comparable urban centers like Chicago, Boston, Philadelphia, San Francisco and Detroit.

As Robert VerBruggen and others looking at the statistics point out, if the Stanford estimates of unofficial infections were applied to New York City, the number of infected would be larger than the city’s population, an obvious statistical impossibility. Even the substantially lower ratios of actually infected versus officially infected suggested in other studies still mean, in New York’s extreme case, that a majority of its citizens have already been infected, statistically possible but certainly improbable.

Thus, many of the answers to our most pressing questions will likely be found when we figure out why the state of New York, with a population of less than 20 million, has nearly six times the number of official virus fatalities as California, Texas, and Florida combined (population 90 million).