Thread by @ProfDFrancis

What is unusual about Guinea-Bissau, nestled between Senegal and Guinea?

(Hat tip to Joe Schmoe MD, @joeschm80238897)

Well, you know how you and I and everyone's brother-in-law are all talking crap on Twitter about how effective or ineffective face coverings are, yeh?

Well, the WHO is actually doing a trial in Guinea-Bissau.

Guess how many patients they need?

Answer given here

apps.who.int/trialsearch/Tr…

A clinical trial in Europe or USA, for a thing like this, would likely cost somewhere between $1,000 and $5,000 per patient (just a rough guess), because we have endless forms to fill and non-research hangers-on that need to fiddle with paperwork etc.

So $60 to $300 million.

W.H.O .managed to avoid the multi-year delay in arranging the funding etc by going to Guinea-Bissau, where I absolutely guarantee it won't cost $1000 to $5000 per patient.

"Large simple trials" will less and less be done in developed countries, because the development of a country also develops a snowballing bureaucracy that extinguishes motivation to any but the most intensive trials (so that the admin cost is not the dominant element)

But enough moaning.

66,000 patients. That is the kind of size we need to see effects with confidence when each person delivers just a "yes/no".

One binary digit of information (strictly speaking less than one bit, because there is a large majority of "no").

It is for that reason that Cardiovascular trials that have "yes/no" for Major Adverse Cardiac Events as the endpoint need to be ginormous.

And in contrast, trials that have continuous variables (such as exercise time increase in drug trials) can be highly effective when small.