Comment by harleyk | Hacker News Reader

Comment by harleyk | Hacker News Reader

This paper is a big deal. I appreciate the laborious efforts the authors underwent to calculate a more accurate infection fatality rate (IFR). The IFR is the number of deaths from a disease divided by the total number of cases. If 10 people die of the disease, and 500 actually have it, then the IFR is [10 / 500], or 2%.

The IFR is different for age-stratified groups. The IFR was calculated before COVID-19 vaccination became widely available.

It's a big deal because it reports a markedly lower COVID-19 IFR than previous studies. The panic may not have been warranted for non-elderly age groups. I read this to also mean that the risks to adolescents did not warrant school closures and for some ages vaccination, even though vaccination risks for this age group were minimal.

"...in many locations excessive hospitalizations may have been driven by irrationally high perceptions of IFR for non-elderly people and they may have caused unnecessary stress and damage to the health care system at large."

It is not a big deal. Ioannidis has been peddling basically the same study for years now. He put his reputation on the line in the early weeks of the pandemic by making some very badly thought out predictions. He believed the pandemic would kill a total of 10k people in the US, and basically ridiculed the people who thought it would be even as serious as the flu. That particular prediction was made when there were a hundred dead; his threshold of 10k was reached three weeks later.

But John wasn't discouraged! He then predicted that there would be at most 40k dead, and that it was imperative for there to be no measures that could hurt the economy. That number was exceeded 10 days after his prediction.

Since then, he has been trying to convince people that he really was right all along, and basically nobody died just like he predicted.

As someone who is not familiar with finding medical literature, I would appreciate if you could link to those earlier claims by the author.

Googling "Ioannadis 10000" gets you plenty of discussion within links on the first page.

They weren't in the medical literature. The 10k prediction was in STAT news (https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-a...) (note: it wasn't a hard prediction, but he seems to treat it as the best estimate based on the data at the time). The 40K prediction was quoted in the Washington Post (https://www.washingtonpost.com/opinions/without-mass-testing...). There's a good summary with additional links, quotes, and commentary here: https://sciencebasedmedicine.org/10000-deaths/

Thank you

> I read this to also mean that the risks to adolescents did not warrant school closures

I'm not saying there's an easy answer, but arguments to keep schools open because adolescents (the students) may be in a lower risk pool tend to gloss over the fact that teachers and administrators are often older adults, who are more at risk.

Every child is attached to a minimum of 1 adult on average. And they are very unlikely to be particularly adept at not acquiring whatever bug is going around let alone something as infectious as covid. So to me that was always the point. One infected kid leads to all kids being infected which leads to all parents getting infected, etc.

The R naught number was very high event for the early strains: https://abcnews.go.com/Health/r0-covid-19-virus-key-metric-o...

Seems wildly unfair to describe as irrational someone seeking treatment when they didn’t need to between March and June 2020.

We didn’t really have good numbers. The numbers we had were extremely alarming. It was entirely rational to seek treatment early given the information at the time.

Ioannidis published very suspect studies with misleading design and motivated reasoning during the pandemic. His credibility is low on this topic.

The paper does seem to address this:

"Several previous evaluations (Ioannidis, 2021a; Levin et al., 2020; O’Driscoll et al., 2021; Brazeau et al., 2020) have already synthesized information on age-stratified estimates of IFR. Most of those used data from early published studies, and these tended to have information from mostly hard hit countries, thus potentially with inflated IFR estimates. Moreover, several analytical and design choices for these reviews and data syntheses can be contested (Ioannidis, 2021b) and many more potentially informative seroprevalence studies have been published since then."

Reply on news.ycombinator.com