Hacker News Reader

Comment by jsnell | Hacker News Reader

The title is funny, since Ioannidis is actually the one proposing measures ("do nothing") based on no evidence and exaggeration. Basically every number in the paper is tainted by him cherry-picking the most optimistic number possible, and then trying to twist it further into something even better.

Like his argument for the R0 being near 1.3 is just that it's "probably" the case.

At some point we need to be able to call the data we have good enough for making decisions.

View on HN

R0 is at the initial onset, without any immunity and any preventive measures. If the virus had R0 of 1.3 it would be easily containable. The reality is that it has an R0 of 2-6 and an effective R > 1 in most except Asian countries.

A R0 of 1.3 with an incubation time of 1 week would not lead to the doubling of cases in 4 days. The whole thing is a crime against math.

Weirdly enough, just yesterday I took all available sequence data currently on GISAID, used it to estimate R0, and got ... 1.3.

Intersting. Can you post what you did maybe on a Jupyter notebook?

The imperial college had it estimated 2.1 - 3.3 before "stay at home" measures. Obligatory influenza is 1.5.

If that were true then how can you possibly explain how rapidly the pandemic is spreading?

Long time to spread.

It's curiously spreading at the exact same rate that tests are being administered. Curious indeed..

Maybe you got the terminology mixed up like I had? `* 1.3` is just about the growth rate per day, which isn't R0.

Um. He doesn't advocate do nothing. He specifically says the opposite. Read the paper. WTF?

Sure he recommends researching vaccines faster, giving people flu vaccines since the flu is more serious than Covid, admitting fewer Covid patients into hospital care, and doing more research so that we're better prepared next time when something actually serious happens. Or if this is serious, doing more research so that the data is the most accurate possible.

I think calling that doing nothing is pretty generous. Since you disagree, what measure that he's proposing did you have in mind?

he seems to advocate individual hygiene and avoiding the public when sick, which also btw seems to be the response that countries like Japan, Taiwan and Singapore have taken, where complete lockdowns or closures have largely been avoided. Together with tracing they seem to have handled the situation just fine.

Taiwan is doing more than public service announcements. Singapore too. Japan closed all schools in Feb. Tracing won’t work without first getting the number manageable.

> together with tracing

This is key — these countries were able to avoid lockdown by testing and tracing early, before the case load became unmanageable. In the US, we’ve missed that opportunity

But US still needs to develop that capacity as quickly as it can, because once quarantine brings the virus under approximate control, testing and contact-tracing are what can eliminate it.

Completely agreed!

South Korea, with one-sixth the population of the US, peaked at 10,000 cases. We haven't passed that per capita case load - and even if we do, lower population density should make it easier to get the virus under control in the US (except in NYC).

We're still a long way from Korea's testing capability. Mass use of surgical masks in public probably also helps.

On Friday, March 20, The Atlantic said over 100,000 people in the US have been tested.[1] More recently, Mike Pence said 250,000 people have been tested.

We're a few days away from Korea's testing capability, if we haven't already matched it.

I think southeast Asia in general handles pandemics better. The people know how to respond and do so more quickly than Americans. Wearing masks, not going on spring break, etc.

1: https://www.theatlantic.com/health/archive/2020/03/how-many-...

300 tests per million people is much, much better than we were a week before that.. But South Korea is at 6,000.

https://ourworldindata.org/grapher/covid19-tests-per-million...

Those are old numbers. Roche alone is sending out 400k test kits per week (over 1000 per million people).[1]

According to [2], more than 290,000 Americans have been tested for the coronavirus (close to 1000 per million people) and in Washington and New York, over 3000 people per million have been tested.

What is South Korea's testing capability (tests per week)? They've had several weeks to get a lead in absolute number of tests performed, but if they've only done 6000 per million people in all that time, we've probably matched them in testing capability.

1: https://diagnostics.roche.com/us/en/news-listing/2020/roche-...

2: https://www.msn.com/en-us/news/us/one-map-shows-how-many-cor...

He seems to recommend the protection of the high-risk slices of populations and campaigns to increase awareness regarding the importance of hygiene. Also figuring out the real incubation period since, he claims, the original patient that was found contagious was already symptomatic, but researchers did not ask.

Reply on news.ycombinator.com