Carl Heneghan

But on a balance of probabilities, for me personally, talc doesn't do much for me now.

And looking at the evidence, there's still some uncertainties and it's slightly leading to more likely to cause cancer than not, if particular, if it's got micro deposits of asbestos.

And if this stuff does have that, that should be banned around the world.

Okay.

The first thing is you may actually alter the intervention, which makes it not realistic in clinical practice.

Well, for instance, certain non-drug interventions, if you try and blind the participant or the healthcare practitioner to actually the intervention, then you'll have a problem.

And I'll give you a good example, self-monitoring of oral anticoagulation.

You know, you need to know what the INR result is.

You need to know what the test result is.

And you need to feed that back to the patient.

So that becomes a problem if you say we can't feed this back to the patient because you're changing what you do in the actual real world.

Second is recruitment to trials, because often if you go to patients and I've seen this happen, you go, we've got these two drugs, but we're not going to tell you what's going on and which one's the best one.

And patients go, well, I'd really like to know what I'm on and what I'm taking.

And so that actually has an issue with recruitment to the trials and feasibility.

So I think there to me are the first two.

And then third is sometimes it's just not possible to blind people to what's going on within the intervention.

And I'll give you examples.

The drugs might have specific side effects.

And once you get them side effects, you know what's happening.

And if you give a placebo that's inert and doesn't have the side effects, people can guess anyhow what's going on.