Statistics are hard! (Whine)

We hear numbers all the time. And in particular, as it relates to COVID, there's a lot of noise among all the numbers you hear.

And oddly, interpretation of them became a partisan political issue. One person says X and the other says Y and they both support the assertion using statistics.

The media try and "pick a side" and try and report on what they see, but they wind up just reporting on either X or Y and never really question what they are reporting.

It's that way here. In Europe. And in most of the countries where you have an educated population and a healthy GDP. I guess that's how this goes?

Anyway, I heard an Australian broadcaster yesterday who had the audacity to not just report on what he saw on a series of graphs but to ask "why?" ... and that question led to some deeper insights in what the number means.

Look, when it comes to COVID you have some key metrics: number of cases, number of hospitalizations, and number of deaths.

And people argue about whether the death was COVID-related or if there was some underlying factor. Or if the person died of "something" but never got tested for COVID. So is the number accurate?

It's easy to get confused.

But here's the thing: if someone tested positive for COVID and died, then they died *of* COVID. Yes, there may have been an underlying condition. And that condition may have contributed to the persons death. But would they have died from that condition at this time if they didn't have COVID?

This is an infectious disease that does amplify and exacerbate some conditions. And it is worthwhile tracking these conditions so we understand the virus. But to dismiss the deaths as not Covid, when someone tests positive is trying to promote an agenda.

I heard a state medical examiner explaining that it evens out anyway, which makes some sense. Some people may never have been tested. Some may have died of other causes. But the fact is they died.

You can see how hospitalizations might be similar.

Now to infection rate. There's a problem here. You get tested and... hey you're positive! Now quarantine yourself and maybe we do contact tracing, and maybe not. But you are just a data point - unless you get sick.

If we took the actual testing seriously, invested in more rapid testing, took the time to understand that a positive or negative test on its own isn't particularly helpful - the jury is still out on your ability to transmit and what positivity actually means - then maybe we could do something meaningful with the info.

But as it stands? Not so much.

So in the end, we get a stand-alone number of positive test results. It would be helpful to actually know the rate of transmission. To understand this number in the context of number of tests administered, or against the population on the whole.

And then, perhaps the most meaningful thing we can use is the number of deaths stacked against the positivity rate.

That statistic may be useful.

I have been looking at the (reported) tests administered. That's about 80 million. With a population of 330 million, that's roughly 1/4 of the country's population that's been tested. Some people have been tested more than once, so if we assume something like 40 million real people; and that takes us to closer to 1 in 8 people being tested.

Probably not enough to really understand the spread, and do something meaningful, but it's a start.

We also know the number of positive tests. It sits currently at about 8 million. Or close to 10% positivity rate.

But here we get stuck again with the numbers because we don't know some things about how frequently people get tested, or how that relates to the population, or where they got tested (say an urban or remote, so you don't know if there's an outbreak), or the transmission rate.

It's an indicator. But on its own it's not the most meaningful number.

As an indicator, it would suggest only 2.5% of the population has tested positive. Perhaps less, given the number of people who get retested.

But as it's based on the testing rate, that's probably not accurate or fair. The rate will go up if there's more testing. That's how math works.

Now the death rate against the positivity rate is meaningful. If we take the ~270k (.27 million to keep the units the same) Covid related deaths against the 8 million positive tests, we are at about 3% of the population who have died as a result of the virus.

But here's something interesting. Back in the July peak, the deaths vs positivity rate was about 2%.

So something is happening. What, though?

And you may realize that this virus has mutated since it was first discovered, so that makes the statistics less meaningful in a way.

It just keeps us guessing.

This is where we really need an organization like the CDC to step up and do their job, reporting accurately and consistently and tracking the spread. That's what they are chartered to do.

Unfortunately, though, the dumbass in the Oval Office decided to neuter them and further politicize everything about the spread.

He's now saying "we've controlled it and eradicated it" which is dishonest and misleading.

Even the stats can't support that.

So I suggest you question what you hear. What does it *mean*? Is it useful
On its own?

Don't get bogged down in the stats you hear or see. Think logically about why this piece of information is being shared.