Hacker News Reader

Comment by MrAlex94 | Hacker News Reader

Seems at least 60% is good enough:

> Simulation experiments revealed that to prevent an epidemic (reduce the peak by >99%), the vaccine efficacy has to be at least 60% when vaccination coverage is 100% (reproduction number=2.5–3.5). This vaccine efficacy threshold rises to 70% when coverage drops to 75% and up to 80% when coverage drops to 60% when reproduction number is 2.5, rising to 80% when coverage drops to 75% when the reproduction number is 3.5.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7361120/

Also, it prevents hospitalisations completely which is crucial.

I asked elsewhere but seems this thread is the busiest:

Also, does anyone know - supposedly the Oxford vaccine had swab tests for the virus every week to catch cases even if trial participants weren’t showing symptoms, whereas for the Pfizer/BioNTech vaccine they were only swabbing if someone showed symptoms. Would that not have a large effect on efficacy numbers?

View on HN

From the linked paper:

> Using Microsoft Excel, version 16, with the Crystal Ball add-in, the team developed a computational transmission, clinical, and economics outcomes model

I did not expect the simulation model to be written in Excel. Never heard of "Crystal Ball" before. From Oracle's website:

> Use Monte Carlo simulation to automatically calculate and record the results of thousands of different what–if cases

And that’s part of the reproducibility crisis. R/Python code published on GitHub is a much better way.

There‘s whole books on numerical simulation with Excel. Which is absolutely incomprehensible - if you are math-wise enough to do simulations, you should be computer-wise enough to use a real programming language.

I upvoted this. It's quite a lot more important than your elegant one-liner reveals. I think there's a gaping hole in many people's skills that need to be addressed. We're educating people in how certain models work, with lots of equations and diagrams, but we don't give most people access to the power of computation that would greatly help their studies.

I’ve done this type of numerical work in code and it’s much more difficult than it first appears. It’s not like web development this stuff is much more like real computer science.

Tools like Excel and Matlab exist for a reason.

Matlab, yes. Matlab code can be versioned, and opened in VIM for independent review. It can have unit and integration tests. It can also often be run in alternative implementations with few changes, such as Octave.

Excel? You'd be lucky if Excel N-2 can properly handle the file, much less an "alternative implementation" such as LibreOffice Calc.

You can have manual tests (automated through VBA if you wanted) and reviews. The review tools are lesser known features built into Excel. I know from my experience accounting firms will take your spreadsheets and code too and audit it. I’ve worked with bankers who would effectively run a suite of tests before using their updated spreadsheets. These concepts aren’t unique to programming.

Curious as to how your claim follows?

It is well known in the Excel modelling world and a lot of people use Excel for modelling.

I’m not sure coding these things up is always a great idea. I’ve done various types of programming over the years but in Excel you get things like dependency tracing, error checking, UIs for every components, Excel does have advantages for a lot of work. The equivalent code these domain experts produce is usually pretty awful.

> some participants routinely swabbed themselves for SARS-CoV-2 testing, even if they weren’t showing symptoms. Differences in infection rates between people who received the placebo and those who got the Oxford vaccine suggest the vaccine blocks transmission, says Ewer. (The Pfizer and Moderna trials tested only people who showed symptoms.)

https://www.nature.com/articles/d41586-020-03326-w

the 62% efficacy rate was for symptomatic cases, which makes those numbers comparable to the Biontech/Pfizer one. I think it was around 50% effective in stopping asymptomatic cases, which is also pretty good. But that main number comes from symptomatic cases

Yes Pfizer/Moderna didn't do weekly swab tests and only tested the ones with symptoms unlike Astrazeneca.

The other big difference is the sample size is much bigger with Astrazeneca one and is across continents as well.

One thing I have been wondering: Given that R has been near to 1 in many countries, wouldn't we need much lower immunity rates and/or vaccine efficiencies if we maintain the minimal impact measures such as public masks?

My understanding would be that 10% immunity for instance would push R down by 10%. Once we get it below 1.0, we effectively won.

Is this for a vaccine that confers sterilizing immunity (ie, prevents infection and therefore onward transmission, as well as preventing severe disease)?

I don't believe we yet have evidence that any covid vaccine provides this. I've only read that any vaccine prevents disease or hospitalization.

There are some indications in animal studies for Moderna and Pfizer, and some preliminary data given by Moderna at the FDA meeting about reduction of transmission, but it's nothing definitive.

60% is not good enough to be worth taking it for me, personally. I am sure it is good enough for society, but I want a vaccine that allows me to go back to normal life, and if I take the 60% it becomes a coin-flip, plus I will be at the end of the line for the effective vaccines.

The fun thing is to read the paper and see Bruce Lee in the author list.

In Germany only 2/3rds of the population want to get vaccinated. So they'd better be vaccinated with a vaccine with very high efficiency.

70% efficacy in that case wouldn’t prevent an epidemic, but since it prevents hospitalisations and severe effects of the disease it’s still worth it.

It’s interesting how the vaccination speed affects epidemic. What’s the effect of vaccinating everyone within a month vs within half a year.

I wouldn't call 2/3rds as "only", this is quite high. In Poland we have only ~40% that want to be vaccinated (but maybe this a result of Germany having much older population).

I'm a bit reluctant to vaccinating until I see some more tests on wider population and we could test for long term complications (e.g. reduced fertility which couldn't be tested earlier) and I'm a person that vaccinated my children (and myself) for all kinds of diseases.

So I assume there are more people like me (aside from the usual "anti-vaccination" crowd).

The 2/3rds already include people who in principle want to get vaccinated but would like to wait a bit more, like you. Only 1/3rd wants to get vaccinated as quickly as possible.

Reply on news.ycombinator.com