Carl Hennigan

We're just going to go, let's do it.

And countries are doing it.

And what we'll do is end up doing non-randomised comparisons, which will be full of all sorts of problems.

But the environmental exposures will be done.

Lots of that work will look back retrospectively and look at things like temperature and all of that and latitude.

And they are difficult to interpret, but I suspect we'll have so much data in this outbreak that we will be able to infer some interesting sort of effects from the environmental exposures.

Look, it's really important that point because it could be somebody out there is diverging from the sort of herd, trying out some treatment or trying some strategy and comes up with something that could be groundbreaking.

And that's how research works.

The unexpected can just jump out from somewhere.

You can't force people into a box and say, do this.

And I suspect there'll be people looking at the data or looking and they come up with something that we didn't expect.

But I think that's how research works.

And we should never forget that.

The question is the waste problem comes when you think, oh, if you'd, all them hydroxychloroquine trials, if you'd have done them with a minimum sample size and a bit more robust and followed up these patients in this way, by now we'd probably have an answer.

Can I say we're having a bit of a rabble coming in onto the podcast, which is very welcome.

And this is the fun factor of this.

But I'm going to try and stay on a serious point.

For instance, I'm going to write this up.

I've often thought there should be a minimum sample size in randomized controlled trials.

And that actually 100 people is a good starter.