Carl Heneghan

In America, under FDA, none of these are approved.

So that's the first thing to know.

And I think that's hugely important to understand that with smartphone apps, we have a significant problem because the thing is, as an individual, you may use one of these and be falsely reassured that everything's all right when you should have presented.

And in clinical practice, it's never a one-off assessment.

It's a moving feast.

If I'm worried about somebody, I will go, look, I'm uncertain.

I may refer you on to the two-week wait because I understand it's an indeterminate result, but my safety netting says I'm going to refer you.

These are going into NHS innovation programmes.

We have to have a serious debate and understanding of the limitations of these processes.

And the most important aspect about these is they really do have, for serious diseases, require to have a sensitivity.

100% they can have reduced specificity because when they're negative they have to rule out that you had skin cancer so then a person can be reassured it's okay to say they may be positive and they have false positives they'll create a lot of work but actually that's okay because at least they'll be safe so there has to be some performance standards that say sensitivity for something like skin cancer has to be a hundred percent then we'll let it through

because there are other factors at play but if you're if you're saying like it's either or either you're going to go and see your gp or you're going to use this app which is going to give you the answer so so you're saying yes integrate it into clinical prediction rules and then you have to do the right type of research you have to derive it in a clinical set and then you have to validate it in a second set so you're starting to have a higher bar for the type of evidence you would want to approve these devices none of them have gone through that type of performance standards

And I would, yes, say great.

But also remembering you can't have sensitivities of 95, 96 percent when you're going to say, oh, four in 100 people who've got a melanoma, we're just going to let go back out on the street.

You have to have processes like you're saying.

It's part of a clinical pathway, which at some point will involve clinicians or a referral pathway.

Can I?

I want to come in.

Just to say, I noticed that Des Spence, Bad Medicine, is back in the BMJ.

A GP from up in Scotland who wrote many years ago, and he's back, and he's one of the most interesting topical people to read.