One of the facile errors respecting pediatric Covid vaccine statistics?

Citing "absolute risk reduction."

ARR has vastly different implications depending on disease; whether *eventually all people will get it unless vaccinated* (endemic virus) or not (cancer)=key.

Join me!

Not everyone sees this.

Allow me to explain!

Let's imagine that you have a medical intervention that reduces the rate of a disease from 1.5% to ZERO.

Let's imagine 2,200 test subjects.

16 cases in control.
0 deaths in vaccine group.

ARR=1.5%

Is this a big or small number?

The answer is that it depends on the size of your risk pool and lethality of disease.

Imagine a relatively unusual but deadly form of cancer.

A trial with 2,200 subjects would be a pretty big deal.

If you found an ARR of 1.5% for DEATH you might be disappointed.

Why?

Well let's say the mortality went from 20% (control) to 18.5% (injection of interest).

ARR=1.5%

Your treatment went from 220 deaths to 204 deaths.

16 saves.

That could be spun as "meh" in an "unusual form of lethal cancer."

If a grand total of 2,000 US residents die of this rare cancer yearly, you just saved 38 people per year.

Nice but not "changing the course of history."

How about a virus that *everyone will eventually get*?

In the case of *infectious disease*, the final risk-pool is = the TOTAL POPULATION of the demographic in question.

So not 20,000/yr who develop a rare fatal cancer.

But the whole demographic you're assessing.

Say, 75 million US kids.

What does ARR of 1.5% to do in that case?

Well, for SARS-CoV-2, apparently the following, according to Peds Pfizer data:

Vaccine: 0 cases
placebo: 16 cases.

But let's assume that, similar to adult trials, a large enough study would find the *mortality* to be 97% reduced.

Can estimate what this means for kids?

Yes!

Here's the intellectual insight that a lot of commenters who have not thought this through all the way keep forgetting.

We already know the OUTBREAK FATALITY RATE of Covid-19 for kids so far!

It's 282 Covid-19 deaths (US kids 0-17 so far).

(Out of 74 million at risk pool).

The FINAL EXPECTED FATALITY RATE of that population hinges simply on what percent of kids have been exposed so far.

That we *do not know*.

But watch how easily we can estimate the parameters!

Let's assume 1/2 of all kids in the US were infected with Covid already. That is way high but I do it to make a point.

That means eventually another 282 US kids will die, if no vaccine.

So if vaccines decrease cases 100% (and hopefully the deaths by 97% as in older groups)...

Then, assuming 50% of US kids already got Covid (way too high an estimate), you could save 274 kids lives with the Pfizer vaccine.

The risk? vaccine side effects: eg fatigue (apparently life makes you tired; the placebo had notable fatigue rates too, as we saw in adult trials).

The key is to recognize that a trial of an INFECTIOUS DISEASE that EVERYONE WILL GET can have a low ARR and still have an astronomical societal impact.

But wait! There's no way that 50% of kids already got COVID-19, you say.

It's lower, you say?

Ok, let run another scenario:

Let's say only 20% of all US kids got SARS-CoV-2 last year.

In that case, we can expect eventually 1,410 kids would die of the disease (ie. final pandemic fatality number for US kids) *unless* we vaccinate.

How does 1,410 compare to other diseases we vaccinate against?

Hard to know. But it's a bunch.

Example: seasonal flu kills around 125-150 US kids 0-17 each year (CDC Wonder), and we vaccinate against that with a yearly vaccine that is literally made via best guesswork.

So, as always, it is denominator, denominator, denominator.

For a pandemic (or endemic) virus, the denominator of interest is the FINAL NUMBER OF PEOPLE WHO WILL BE INFECTED.

That's literally everyone, since herd immunity is not going to be reached.

In sum, absolute risk reduction is a great measurement when you have a limited pool of risk, and you're worried about things that we worry about with regard to things like cancer: side effects vs time left alive but often suffering mightily, cost, etc.

Absolute risk reduction is also useful infectious diseases...

But the numbers have *vastly different implications* because the final risk pool is ORDERS OF MAGNITUDE larger than non-contagious entities with low incidence and prevalence.

So...the Pfizer trial in the kids today?

Means we just might have found out that we can now save THOUSANDS of kids lives in the US alone.

Now, I still think we should prioritize higher risk people if we can (i.e. send doses overseas if we can).

But these data suggest that the FINAL PANDEMIC FATALITY RATE for kids remains unknown and is in our hands!

Vaccines might literally eliminate all further pediatric deaths, depending on rollout + whether decreases in cases correlates to decreases in deaths.

Today was a good day.