COVID19: Frightening Numbers, Frantic Confusion and The Truth About Statistics

COVID19: Frightening Numbers, Frantic Confusion and The Truth About Statistics

I went to the Dollar Store last night to get a bag of penne. There wasn’t any. None of the canned beans I like so much either. I wandered around the store and noticed empty shelves. You never see empty shelves at the Dollar Store, but I did last night.

While I was checking out the empty shelves I walked past a man and woman having a conversation in equal parts Spanish and English, a common enough occurrence here in Tucson that I was able to get the substance of what they were talking about.

The man was saying that he had been to the VA hospital earlier that day and that he had learned there had been four deaths related to COVID-19 in our county. The woman gasped and said something about where it might all end.

At the same time, I was thinking, “four flu related deaths in a county of million people…what a relief”, and wondered how many zeros are in that fraction of a percentage. When I got home, I did the math found out.

Three zeros.

Expressed as a percentage four in a million is 0.0004%.

I don’t even know how to say that. A tenth of a percent is easy (.1%) because we see that number and use it so much. So is a hundredth of a percent (.01%). But four out of a million is so vanishingly small we don’t have the words to express it.

Yet the woman at the Dollar Store was alarmed.

The high fatality rate in Italy resulting from COVID-19 infection is an example of what can happen when social distancing measures are not taken soon enough. But something left unsaid is the unusually large number of elderly in Italy. Like other flu variants COVID-19 is more dangerous to older people.

So is the high rate of Italian fatalities the result of failure to act quickly, the average age of the population, or a combination of both?

Or maybe something else completely?

Time will tell.

This is why the initial reports from China pegging the COVID-19 fatality rate at 3.4% are so suspect.

There are a number of variables we know about, but we have no idea what we don’t know. China is a communist country after all, and official statistics are often found to be unreliable. Since the early days of the epidemic we’ve learned a lot about containing it, so initial statistics may not generalize to the present.

There is more certainty about flu fatalities in the recent past…

According to the Centers for Disease Control (CDC), during the 2018–2018 flu season about 45 million people caught the flu in the United States and 61,000 died. That is a fatality rate of .1%, or one tenth of one percent.

The Washington Post reports that as of Match 16, 2020 the national death toll in the United States form COVID-19 stood at 85, out of 4450 confirmed cases for a death rate of 1.9%. That’s far higher than the death rate for seasonal flu, but will likely go down. About 80% of people who catch seasonal flu never go to the doctor because symptoms are so mild.

An unknown number of people have recovered from COVID-19 and never came in contact with the medical industry. We aren’t really how serious the disease is.

Again, the numbers seem solid, but uncertainly about the total number of cases creates uncertainty.

That brings us to the issues of contagion. How contiguous is COVID-19?

Contagion is measured with a statistical value called R0. If you’ve taken a statistics class you might be reminded of correlation values measured with the R Squared calculation. R0 is a related concept. It’s the ratio of people carrying the disease to the number of people they infect.

Right now, the R0 estimate is R2.28.

That means that one person will infect slightly more than two others.

But wait a minute…

That number is mentioned in a study of an outbreak on a cruise ship where people are packed together, the same as jails and assisted living facilities — settings known to be highly contiguous.

Will COVID-19 spread as quickly in more common settings?

Probably not. But we don’t know for sure yet.

The way we are reacting to COVID-19 illustrates something about how we tend to inaccurately assess risk. When we are using statistics to assess risk it’s easy for our emotions to influence our conclusions and subsequent behaviors.

For example, according to Skydive Perris, parachutes fail about once every thousand times they are used, or 0.1%. According to Planned Parenthood condoms fail fifteen percent (15.0%) of the time they are used.

There would be Congressional hearings if aspirin failed that often.

Yet people are far more likely to put their faith in a condom than a parachute.

Of course, you won’t die if your condom fails. But you won’t die if your parachute fails either, because parachutists are required to carry a reserve, or back up parachute. Condom failures have their own back up, too, in the form of “day after” abortifacients.

Just the same, most people would rather risk a condom failure than a parachute failure.

The interesting questions is “Why?”

Read the rest at