Looks like a straight forward article. Instead of a conclusion you are left with a discussion, which sadly is quite weak, but this seems to be the point. (As I understand it after skimming through it)

Basically you have a data set and some statistical evaluation. The gabs are filled with some medical and statistical "Mumbo Jambo" I didn't read. Seems professional, the graphs are pretty clear cut, which is nice.

The discussion basically states that the data is flawed for the purpose of making clear connections and urges the necessity to provide better data in the future. They aren't even able to make their derived call data public, due to privacy laws in this case.

They are absolute right and have a valid point here. No reason not to publish this and it conveys some important messages.

We should gather better data and find a way to share it, without touching on the privacy of individuals. A very difficult topic I personally wouldn't touch with a stick.

> A very difficult topic I personally wouldn't touch with a stick

Agree completely. This topic has got so ideological that if (when) it comes up in conversation I tend to demur and politely change the subject.

Perhaps at some point there will be a genuine attempt to gather more data on just who genuinely benefits from vaccination (obese/elderly/other comorbidities?), and who doesn't (young/healthy).

One thing we can be sure of is that no such study will ever be funded by Big Pharma.

Of all the parts of the paper to gloss over, the medical and statistical "Mumbo Jambo" is the most important one to evaluate!

Would you like to discuss why they use Spearman instead of Kendall and if this negative binomial distrubition asumption is valid in this case?

It won't have any effect on the final result, which basically states that the data set is insufficient.

> but the Nature imprint seems impressive

It's Scientific Reports from the same publisher, but not Nature, the journal itself. Scientific Reports has a very high acceptance rate, i.e. they publish pretty much anything that passes peer review.

Traditional paper-bound journals are intrinsically limited re: number of articles that can be published. OTOH online publications aren't so restricted, so why not publish all that are worthy?

I think the real question is the quality of peer review. A "high rate of acceptance" isn't bad if accepting an article happens only after good quality peer review. But a high rate of acceptance without adequate peer review would be a big source of concern.

Not sure which of these publishing considerations applies to this article. I'll have to go over it again in detail when I'm more rested, anyway on quick overview I didn't spot any glaring problems in it.

Don't know in general but there is another study with the same statement for young people with increased heart problems. Cannot find it, but it is just a few weeks old and I think the country of study was Denmark. Not sure though.

I am leaning to towards not reliable, as they claim the trend started in Jan 2021, which is a little early for when Israel was vaccinating that population: https://www.theguardian.com/world/2021/jan/31/israel-covid-v...

It seems like the authors probably went into this with a conclusion in mind…

I’m not very knowledgeable about this, but I was wondering if MIT (professor?) co-authorship precludes papers from coming to the wrong conclusions? Or is it we’ll know that when a co-author is from MIT that the paper is as rigorous and objective as possible?

I am just comparing the potential veracity of a peer reviewed article in Nature of two MIT authors and one emergency physician from Ben Gurion University with a one line dismissal and a newspaper link. Which one is "we'll [sic] known" to be more plausible?

This is not published in Nature. Scientific reports is much less prestigious and selective.

I said possibly less reliable because of some interesting data mismatches, not dismissing it. considering then other comments addressing the flaws in their stats, it seems quite wrong. I have a Masters degree from UBC and work as a data scientist, not that actually matters…

Can you point to the specific error in the paper? I have a PhD from MIT, but I am not sure why "UBC" is relevant to the paper in question.

I cannot, it is not my area of expertise. For the same reason "MIT" is relevant.

Ok I understand

Please follow HN guidelines when responding, you could have simply stated, "Looking at figure 1b, it seems like there was time delay between the first dose and the increase in rates".

This is a wrong venue to condescendingly teach other people how to write comments. The goal of this forum is to contribute information or insights. If you you think I violated HN guidelines, please downvote and report to dang.

Respectfully, it is exactly the right venue? I sorta thought that was the point here. And sure, done.

on the page, there is a link to his google scholar page He is a serious scientist: https://scholar.google.com/citations?user=RBSyaTIAAAAJ&hl=en

I guess many anti-vax people confirmation-biased this article and are tweeting about it, and if I can guess, you're pro-vax and your confirmation bias is skeptical about it (if it was an article that touted how good the vaccines are, you'de be less skeptical about it, and less critical pro-vax minds would be on Twitter saying "See, I was right!").

I don't know ryandamm, so I can't judge whether he is pro- or anti-vax, but you should be fact checking every article before sharing it.

Doing basic checks like "what else did the authors publish" seem fairly sensible.

Actually it isn't really since the scientific community regularly has the problem of ostracising those that offer much progress. It is a decent method in practice but ultimately fallacious. The corpus of published works is far too massive so it is a compromise. A bad one, because the pressure to publish has a negative effect on scientific rigour.