Carl Hennigan

I have serious concerns about a lot of the research that is being published at the moment, how much we can trust it, how much we can verify the results.

And people only have to look at what's happening with the antibody tests in diagnostics, where we're getting over 100 tests emerging from different manufacturers, and much of them are not working when you verify them in your own population.

This is a serious, serious problem.

that needs to be assorted.

It's almost like you need a license to be able to do research.

And then you have to act according to good clinical practice to be able to do the right things at the right time.

Unfortunately, many parts of the world are not conforming to these standards.

And for me, this is a huge disappointment, but the important thing is it means we are still in a position

Halfway through this pandemic, nobody has a clue really about any treatment that could work.

And we've said this last week is a systematic approach with good clinical practice and the right research done in the right way could help us.

What's happening at the moment is adding to the disaster.

Can I come in on this point?

Because it's incredibly important.

We've realised there's a massive oversight in the architecture of evidence-based medicine.

Let's put aside diagnostic and prediction rules.

What about all these models that are coming out?

There are over 60 now, and there are huge structural problems with them.

They never refer to each other.

and say in the context of this other model so there's no systematic review approach they don't update on a daily basis they don't derive the data and then validate it and you'd expect them to be living structures because what they want to do is say we've derived this model and now we're going to test it for the next four or five days with data as it emerges to see how stable it is and if it's still changing then you go we can't use it but they publish it and go here it is hear the truth and spend all of their time defending it

So out of this, we are going to produce a new structure for how mathematical models fit into the architecture of EBM.