For you amateur (or professional) data scientists

Some of you - okay, lots of you - have emailed me to point out that the chart showing United States influenza deaths in the last previous Stack is NOT adjusted for the growing population.

In other words, all else being equal, we would expect influenza deaths to rise as the population grows, so maybe the fact they haven’t shows the vaccines work after all? At least a little?

Before I answer, a side note: I love Team Reality! You keep me honest.

Now, let me point out a few other things the chart is not adjusted for:

The aging population (which should tend to increase deaths)

Falling levels of smoking and lung cancer (decrease)

General improvements in healthcare, including mechanical ventilation (decrease)

Tamiflu (decrease, maybe, the evidence is weak)

Different strains of flu (depends on the strain)

No doubt you can come up with others.

Do you see where I’m going with this?

All else is not equal. Trying to measure the net effects of all those changes is difficult even for someone with no incoming ideological bias. Using population-level data can be tricky, because many different changes can occur simultaneously. Measuring the effect of any single variable with precision can be difficult.



The key words there are with precision.

In the 1980-81 flu season Americans received 12 million flu shots. In the 2017-18 season, they received 155 million.

In the 1980-81 season, the United States had 227 million people and about 20,000 flu deaths. In the 2017-18 season, the United States had 327 million people and 52,000-81,000 flu deaths (the Centers for Disease Control initially estimated 81,000, then years later - after Covid hit - offered the new lower estimate. Which had the effect of making Covid look even worse. Which definitely wasn’t why they made the change. But I digress.)

Here’s the point.

If the flu vaccine reduced the risk of death in any meaningful way, increasing its use 12-fold - use that is explicitly directed at the elderly, who are most vulnerable - would have SOME visible impact on overall mortality from influenza deaths. It has none.

Okay, 2017-18 was just one year - what about the long-term trend?

At best, we can say population has increased slowly, while flu vaccinations increased very quickly - and flu deaths have jumped around without any discernible pattern.

More CDC data, this time for the 30 years ending in 2007:


This dataset actually looks even WORSE for the shots. Please note the numbers run distinctly higher from top to bottom. That is, the trend is toward more influenza deaths year after year - AS FLU SHOTS INCREASE.

To be clear: this evidence doesn’t prove flu shots cause flu deaths. But it sure doesn’t prove they do any good, either.

If your friendly public health bureaucrats wanted to prove influenza shots reduced influenza deaths, they could do so lickety-split.

They would just have to run a large randomized controlled trial (ideally two or three trials, over two or three flu seasons) that enrolled, say, 25,000 elderly people at high risk of death from flu and split them into placebo and flu shot groups.

This trial would be perfectly ethical, because at the moment we have clinical equipoise - a fancy way of saying that no one really knows if the shot does any good. It would cost a tiny fraction of what we now spend on flu shots, not to mention the associated flu shot advertising campaigns.

And it would give us a definitive answer. Within a few months after the end of flu season we would know - not guess, know - whether flu vaccines actually reduced flu deaths. We could even look at ALL-CAUSE mortality, and find out whether the vaccine might be causing other problems that we are not catching (or possibly even have some hidden benefits, who knows? That’s why you run the trial!).

So why haven’t your friendly public health bureaucrats run this trial. They have pushed these shots increasingly hard for the last 40 years, and especially the last 10. Don’t they WANT to know if they work?

I’ll give you a hint.

Of course not.

As long as they don’t know the vaccines are useless, they can continue to insist they are useful - and that the only problem is that not enough people have received them.

The risks of finding the truth far outweigh the rewards. Not for the people getting the shots. For the bureaucrats pushing them and the companies selling them.

And we know who really matters.