Hacker News Reader

Comment by DennisP | Hacker News Reader

It'd be nice if they mentioned how many cases there were, compared to number of shots.

They link to the CDC slide deck. Slide 14: https://www.fda.gov/media/150054/download

488 total reports from Pfizer-BioNTech, 301 reports from Moderna.

Slide 18 shows proportions. 12,169,692 doses administered. Expected number (base rate) of Myocarditis/pericarditis cases in persons under 25, 10 to 103. Cases observed: 277. So it roughly quadrupled the base rate. Which brings the adverse reaction rate up to... 0.000276%

The debate between presenting absolute vs. relative rates is a really interesting one in epidemiology, for the record, with roots spanning everything from "absolute risk regression models are actually kind of tricky" to the early arguments about smoking and lung cancer, casual criteria, and R.A. Fisher being on the wrong side of history.

This only requires that a person with myocarditis be 3-4x as likely to notice it after a COVID vaccine than without a COVID vaccine. That seems consistent with a real effect but also consistent with no vaccine effect at all.

The whole article could use a lot more context, which has been typical of COVID reporting. For one thing: is heart inflammation a rare side-effect for vaccines and/or immune responses generally? If not, to what degree (if any) does this exceed rates we see for other vaccines?

That should be second-paragraph-of-article sorts of information.

Exactly this.

I'm curious how much we're seeing effects that are present but just haven't been measured previously, because of the size of the newly-vaccinated population.

The second dose is obviously very stressful on the body - people are often down for a couple days dealing with the reaction - so it seems unsurprising that there's /some/ kind of measurable secondary effect. Stress is known to correlate with heart attack.

It's a postmarketing surveillance system [1]. Not a controlled trial, so the true exposure vs. reported cases isn't known for these data from VAERS.

> "There were 283 observed cases of heart inflammation after the second vaccine dose in those aged 16 to 24 in the VAERS data. That compares with expectations of 10-to-102 cases for that age range based on U.S. population background incidence rates, the CDC said."

They use Observed vs. Expected cases in order to calculate e.g. a reporting odds ratio [2], which can loosely guide you in finding events which occur disproportionately.

1. https://en.wikipedia.org/wiki/Postmarketing_surveillance

2. https://allaboutpharmacovigilance.org/43-reporting-odds-rati...

Agreed. Context matters here.

I'd hate for this to get blown out of proportion like the blood clot issue did, causing massive confidence declines in places all over, especially those struggling to roll out vaccines and keep their COVID cases under control.

Obviously, if the numbers are alarming, we must react accordingly.

I would like to see it uncovered. I did not sign up for heart inflammation and any potential future problems that may come out from it.

They are alarming based on the measure of Observed vs. Expected cases.

Observed (O) cases in that age group: 283

Expected (E) cases in that age group: 10-102

So at least double the top of the confidence interval, for expected.

Right, but you've got to look at the absolute rate, which in this case is 283 out of 12 million, which is roughly 24 people per million.

You can look at the risk of what it's preventing (covid has already killed ~1700 people per million in the US), or compare to other activities and risk levels.

Relative risk comparisons between different activities are often not useful. For example, the risk of shark-death at the beach might look alarming when compared to my risk of shark-death-at-home, but actually it's vanishingly unlikely.

The trouble is that the people apparently at risk from this are not the people at risk from Covid-19. The risk to people between the ages of 12 and 24 from Covid is really, really small. Especially towards the lower end of that age range, most countries seem to be seeing low single-digit per million deaths: https://www.thelancet.com/journals/lanchi/article/PIIS2352-4...

The risk of death from Covid-19 is low in that age range.

But what about the risk of heart inflammation from Covid-19 in that age group, possibly with no other Covid symptoms?

There's also the relatively high chance that you'll pass the infection on to another person, who may not be nearly as healthy/resilient as you.

That’s the right question to ask

but that's out of 13m doses, so ~200 out of 13m. I'm not sure how that compares to the expected harm from catching covid but I imagine the vaccine still ends up massively favourable

This depends a lot on the probability of contracting Covid at all. Roughly speaking, if there's a 1% chance to contract the disease, the vaccine has to be at least 100 times safer. As the chance to catch Covid decreases over time, the risks associated with vaccines become more of a problem.

> As the chance to catch Covid decreases over time, the risks associated with vaccines become more of a problem.

This is a big assumption. It could also be that it becomes part of the virus background infections we all deal with except that it is high risk if you get it the first time.

When looking that the risk of an non vaccinated individual catching Covid today keep in mind that a lot of people have been vaccinated and the incidence numbers de-facto refer to incidences of the non vaccinated part of the population which would justify a smaller denominator than used (whole population).

Without vaccines the risk of catching covid is of around 70% at least.

Is that for an individual not taking the vaccine but benefiting from herd immunity of others, or a person in a while society without vaccines?

For society in general.

That said, there is no guarantee that herd immunity will ever be reached at least in the US. If herd immunity isn't reached then it's basically guaranteed you'll get it eventually as the disease becomes a "background disease".

Per the article. The age group of concern "accounted for less than 9% of doses administered."

Also, underreporting is a factor - not everyone knows about VAERS, not everyone goes to the Dr. for chest pain etc.

I agree it's favorable, but it's important to understand the effects of any medicine on subgroups of a treatment population.

Have to really consider if the 200 are but a fraction of the actual though. I mean how many are going to be reported?

I'd wager it's approximately the same fraction of myocarditis cases that are reported in the unvaccinated group. There may even be a bias toward over-reporting myocarditis in the wake of a vaccine, as people are primed to look for side effects.

They do, in this exact article. 283 cases were noted in the VAERS database, there are about 130 million fully vaccinated people in the US.

I assume they don't track healthy people in the database of adverse vaccination reactions.

So they do. I'm pretty sure they added that after I read it.

Reply on news.ycombinator.com