Wednesday, August 5, 2020

Understanding COVID-19 death rates - By Al Maurer

We are daily bombarded with stories about the lethality of COVID-19.  The assumption of every broadcast is that if you don't follow the suggestions of the experts and the executive orders of the governors, you will get sick and die.

Only by extreme sanitizing, staying away from other people, and wearing a mask everywhere — if you must leave home — will you be safe.

Is that really true?

There are so many contradictory opinions floating around in media that it is almost impossible to logically decide.  I decided to look at a map: where were these cases and deaths occurring?  At first glance, it appeared that the numbers were higher in big cities.  New York City, for example, had the highest rates of death—208 per 100,000 population.

It seemed that COVID death rates had more to do with population density than anything else.

To test that hypothesis, I got population density per square mile by state from the U.S. Census Bureau and COVID death rates per 100,000 population from the CDC.  If my hypothesis was correct, the higher the population density, the higher the death rate would be.

Plotting the data gives the following graph:

Death rates are the independent (y) variable, population density the dependent variable.  The relationship is so clear that the regression line is superfluous.  And just in case the eye doesn't make it obvious, the correlation coefficient is 0.992511!  For reference, 1.0 means a perfect 1:1 correlation.  Statisticians begin to posit a causative relationship with a much lower correlation than that.

Also of interest is how close the individual dots appear to the regression (trend) line.  You can always draw a regression line through a series of data points, but the fact that the dots are all so close to the line points to low variability (hence the high correlation).

What does this mean?

Without considering masking, sanitizing, socialist distancing, or any other measures advocated by the experts, population density alone explains death rates.  Everything else is political spin.

It also means that the variations in policy state by state mean little, if anything.  It doesn't matter if the governor is Republican or Democrat.  Texas, relatively loose in its application of rules, and California, relatively strict, both have death rates of 24.  The same is true of Colorado (32) and New Mexico (31).

Why not use "case rates"?

The media love to use case rates.  Epidemiologists use case rates because they focus on the spread of disease.  It's what they do.  But for citizens and their elected leaders, it's the death rates that matter.

Besides, the case rates we are seeing are unreliable data in the statistical sense.  Case rates are really positive tests — and there are different tests measuring different things and with different accuracy rates.  Death is certain.  Death rates, too, are probably somewhat unreliable — speculation a month or so ago went that as high as 40% of "COVID deaths" were caused not by the disease, but rather by other pre-existing conditions.  I assumed that while the data may be exaggerated, it would be exaggerated about the same in every state.

Is this really a causative relationship?

One should be careful: correlation is not necessarily causation.  But it does make sense logically.  If this disease is primarily transmitted between people in close contact in a confined space for 15 minutes or more — as we've been told — then people riding the New York City subway are more likely to catch the virus than a park ranger in the middle of the forest.

So if all the draconian measures taken in the U.S.  have been ineffective, does that mean we should do nothing?  Let's see.

The government of only one country in the world trusted its population enough to let them make up their minds themselves, with little interference: Sweden.  Sweden's infection rates were initially higher but have since dropped off.

Resurgence in the number of cases is to be expected if people stop practicing distancing and get together again in large groups.  Is that a bad thing?  While case rate have gone up, death rates have remained low, at least in the U.S.

And what are those death rates, overall?

The CDC reports for 2018 2,813,503 deaths total, or 863.8 deaths per 100,000 population.  By extrapolation, in the first half of 2020, there would have been a little over 1.4 million deaths in the U.S.  without COVID.  Yet the medical and political establishment in this country is freaking out over an additional 154,471 deaths — a roughly 11% increase if none of the people who died from COVID would otherwise have died.

Yet we know anecdotally that many reported COVID deaths were not caused by the virus.  If doctors were allowed to honestly report cause of death, we would have a better idea of the real magnitude of the COVID impact.  It is likely that we never will.

If stopping the spread of the virus is the goal of government policy, then you may be wearing a mask in public for the rest of your life — and it won't make a difference.  It is hubris of the greatest magnitude to believe that we can eradicate this or any other virus from the face of the earth.

Al Maurer is a political scientist and technologist.  He writes on topics of limited government and individual rights.