Hacker News Reader

Comment by officialchicken | Hacker News Reader

Unpublished, non-peer reviewed, and sample size of 8... is there a well known name for opposite of FUD?

This seems like an unreasonable assessment. The preprint is available (https://www.biorxiv.org/content/10.1101/2021.01.25.427948v1), submitted for peer review, and the article is pretty clear from the start that this is a preliminary finding.

Where was it submitted? I couldn't see that anywhere.

I'm not sure; the Moderna release mentions submission but doesn't specify.

Well, in your analysis, what's the minimum sample size and why?

"Peer review" is not a suicide pact. In the middle of a pandemic, where days matter, you publish information as you get it, and act with that understanding.

I disagree. In a pandemic it's important to get solid and clear information to the population. If you start saying things that don't go through the proper channels to be verified as reasonably true, then you create massive confusion when you potentially have to go back on that information. Scientists can and do make mistakes. A lay person might not understand what "preprint" or "not peer reviewed" mean yet this article is on the BBC which the average lay person does trust as a source for reasonably true information.

I think that if "the other side" sticks to the rules and doesn't publish things like :https://www.theguardian.com/society/2021/jan/20/covid-vaccin... using small sample sizes and various assertions about how the immune system works then you are right!

On the other hand, if you are in a situation where no holds bared screaming to get attention is the norm then it may be good to share information quickly and transparently.

Last week the media in the uk ran headlines saying that covid was spreading like wildfire, despite the evidence from the number of cases showing the opposite (halving every 2 weeks)

The day later they ran headlines sayign that covid wasn't spreading like wildfire.

The source of the first scaremongering inaccurate dangerous headline was a study which said:

  During the period 6 January to 15 January, SARS-CoV-2 virus was circulating with a higher prevalence than between 25 November to 3 December with 158 in 10,000 infected. There was no strong evidence for either growth or decay in prevalence averaged across the period 6 January to 15 January.

Which led to headlines like

"Covid-19 cases have increased more quickly since lockdown started in England, study finds"

(Lockdown started on Jan 5th - when about 60,000 cases were being identified each day. The headline was Jan 21st, when about 30,000 cases were identified)

https://inews.co.uk/news/covid-19-cases-england-increase-sin...

The 7 day cases identified from Jan 6th to 15th dropped from 55,885/day to 40,242/day

That either means the number of unidentified cases balooned, or the REACT study was too small to identify changes over the course of a week (it's not designed to). In the former case we'd expect fewer tests were being done, but tests throughout January have remained averaging about 550,000 a day.

The media will print whatever they can, so it's important to give headlines which make it clear the data is inconclusive. "appears" is a weasel-word that is well used in this case, and far better than misreporting a scientific paper

> is there a well known name for opposite of FUD?

False hope

hype

Hubris

This result is the absence of bad news. No more, no less.

Regardless of what this early result is, the only thing most of us can do is try to keep social distancing, use masks, and get vaccinated to protect ourselves as best as possible.

If the vaccine doesn't actually protect us against the newer strains, we should still take it because we need protection from the more widely strain regardless. The only difference is whether we'll need to take a second vaccine in another year.

The more guessing and conjecture I hear about the effectiveness of current vaccines on mutations, the angrier I get. Are there any clinical trials going on right now to determine how well various strains respond to the current vaccines? Until Pfizer or Moderna puts out a press release saying their products cover the top 4 major strains, I'm not believing anything.

There are two ways to do clinical trials on vaccine effectiveness. One is to vaccinate someone and then deliberately infect them; this gives you good data quickly, but due to the deadliness of COVID-19, this is generally regarded to not be an option. Another is to vaccinate a very large number of people and wait to see if they get sick with the virus from natural exposure. This is ongoing right now on a very large scale; millions of people have been vaccinated, and sooner or later people will come in contact with these new variants. However, it takes time to collect enough data with this approach to provide useful results, and in this case it can be confounded by the lack of a control group (you don't know how many exposure events happened that did not lead to infections).

Regardless, the current vaccination campaign is highly effective against the original - still very widespread and deadly - virus, so it's worth getting it, and as people get vaccinated with it, it will eventually be apparent whether it works against various variants as well. If it doesn't, it's likely that a booster vaccination can be developed for these variants (and indeed, Moderna has begun work, as an insurance policy, on a booster for the south african variant).

Actually the UK is performing a challenge trial with young adults. Sign up here to get infected: https://ukcovidchallenge.com/covid-19-volunteer-trials/

Sample size of 8 is not really a huge issue for this kind of analysis.

The hypothesis is that the Moderna vaccine produces antibodies against the new variant.

8/8 people sampled produced antibodies.

Can we confidently say that 100% of people will produce antibodies? No, not with a sample size of 8.

Can we say confidently that 50% of people will produce antibodies? That the vaccine gives you a coin toss's odds? If it's just a coin toss, this result would be the same probability as 8 "heads" in a row, or 1 in 256.

So, the result is hardly FUD or meaningless. They've also submitted to journals for the world to inspect the results.

Reply on news.ycombinator.com