John Ioannidis

Practically what you see when you have an outbreak, you see the worst cases, those that have the severe symptoms, the worst outcomes.

However, this is just the tip of the iceberg.

Usually there is far more

that is going on in terms of infectious load in the community.

And the question is, how much more is the part that you're missing?

So the early estimates, if you take the base fatality ratio or infection fatality ratio for SARS-CoV-2, the quote by the WHO was 3.4% based on early data that looked at the denominator, how many people died, and the denominator, how many cases we have documented

based on testing.

However, testing had been extremely limited almost all over the world and it was not based on representative random samples of the population.

In one setting where we have representative sample of the entire population at the moment or close to that is Iceland that has a very well designed study covering the Icelandic population.

They test volunteer samples that represent the general population until the time that we're talking

based on what they have tested, they see about 1% infection rate in the community, which is about 3,500 Icelanders being infected, and only one death has been recorded, which is an infinitesimally lower rate compared to the 3-4% that went into the original calculations.

Another situation is the cruise ship, the Diamond Princess, where everybody was tested, more or less, with a few exceptions, and the

The number of deaths were 7 out of 700 people found to have been infected, which is 1%.

But this is a very elderly cohort.

If you adjust for the age structure and gender structure in that population to project it against a typical general population of the U.S.

or the U.K., the infection fatality rate is much, much lower than 1%.

With sample size of 10,000, you can get a very reliable picture with relatively limited uncertainty about what is the ballpark that you need to know to make an informed decision.

So it can clearly tell you whether we are at 0.1% of the population or 1% or 5% of the population being infected or higher.

And this is really what we need to know urgently and be able to track in terms of big changes downstream.

If the percentage is very low, then the best strategy would be to just try very, very forcefully to track every single case and the contact and make sure that you isolate people who are infected.