Hacker News Reader

Comment by DanBC | Hacker News Reader

Also not so good: this is a corporate press release. The data will come later.

The results came from the independent data safety and monitoring board that is analysing the trial results on PFE/BNTX's behalf. PFE/BNTX themselves don't currently have access to more data. But there is no reason to suspect that these results are not accurate.

At least this one has a decent looking 'n' value. The BBC also address this in part

> There are still questions - how long does immunity last, does the vaccine work as well in high-risk elderly people, does it stop you spreading the virus or just from developing symptoms?

I'd rate the news as "vibing" on a scale of clickbait to party for the world

The current n=94. The real n=43.5k data apparently will come later. Article doesn’t specify when, but “hopes” to have enough of the larger n analyzed to submit to regulators by 3rd week of November.

No, 43.5k is the number of people in the trial and 94 is the number of confirmed COVID cases within the entire population.

The trial will be complete from an effectiveness perspective once there are ~150-200 cases.

Safety is a different story and may take more time to wait for side effects.

Thankfully safety is looking pretty good according to the press release. However I look forward to reading expert analyses of the actual data to confirm.

> However, the data presented is not the final analysis as it is based on only the first 94 volunteers to develop Covid so the precise effectiveness of the vaccine may change when the full results are analysed.

Thanks! Misread the 94, but the n=43.5k is also misleading since they have not analyzed all of the results and duration of the study vs exposure risk is still tbd. If the 94 participants that got COVID also abided by all the other precautions, such as masks, etc, then it stands to reason that the actual infection rate in the population approximating pre-covid behaviors may also be higher.

To be clear: 94 out of 43500 people in the trial got covid, and if the vaccine is rated at > 90% effectiveness, that means at least 85 of those 94 people were in the placebo group, and less than 10 people were in the vaccinated group.

Does it say how many people got the real vaccine?

FWIW, N=94 seems ridiculously small to me. I realize there's ethical concerns in increasing exposure etc but in my experience, that's in the ballpark of sample sizes where unacknowledged heterogeneity can be manipulated to distort findings.

I'm not suggesting Pfizer is up to something nefarious, just that the biomedical literature is littered with studies with N < 100 for the outcomes of interest and they don't replicate. The N > 40k in some ways is irrelevant if none of those participants have the processes in play. It's more like recruiting N >40k to get N=94 to study.

I'm also not suggesting with 90% effectiveness that there's probably no effect, just that my guess is a true value of 60% isn't outside the realm of possibility once you consider all the factors that could be in play besides the random sampling variation that's assumed by their tests.

Seems fine to me. 85 people from the placebo group caught covid. If the vaccine was completely ineffective you'd expect roughly the same number in the vaccinated group. The actual number was 9.

What you're saying is the 9 is pretty close to 34 (60% effectiveness). No, it isn't. You'd need to quadruple the number of vaccinated cases to get there. Even if you double them, you're still at 80% so I'd take that as the realistic worst case.

Edit: Also, this is BioNTech's vaccine, not Pfizer's. They partnered with Pfizer as they have the facilities and scale for a massive rollout, but it was developed by BioNTech. I wish they were getting more credit in the US & UK coverage.

Judging from other threads these trials have been independently reviewed, so it appears it's not "just" a press release.

> these trials have been independently reviewed

That's how clinical trials are done.

Yes, but in other cases there have been "independent" reviews that weren't really independent, e.g. test makers who did their own test validation but got academics to do the writeup ... who worked at the university the manufacturer was spun out of. There's a lot of bad behaviour going on right now in the scientific world that would normally not be tolerated, but standards have dropped lot in the name of speed.

That seems wrong. The data probably came first, but in less popular publications.

Reply on news.ycombinator.com