Carl Heneghan

so description so there are there are large scales you can look at it on a practice right now in uh open prescribing data which is run by a chap called ben goldacre at our department who that you can look and benchmark different practices but as individuals we get back fed back our data in audits so that's not widely available but each practice and each individual because as you sign a prescription it is it is annotated to my recorded to my name

The last audit I did was actually against antibiotic prescribing because we're not high users of opiates in out-of-hours care, but antibiotics is important.

And it was useful to benchmark you against individuals, but also looking at the cost as well.

Well, interestingly, earlier this month, the Nelson trial, which was a trial of 13,195 men and 2,594 women, looked at lung cancer screening and gave the 10-year follow-up data.

Really important outcomes and important to look at what it means.

Now, one of the things is, and I had great difficulty with this paper, was getting my head around the exact results.

Yeah, and I think this is really interesting.

And what it reports in the paper is it reports the cumulative rate ratio for death from lung cancer at 10 years was 0.76.

95% confidence interval, 0.61 to 0.94.

Yeah.

Well, interestingly, the first thing you know is that's lung cancer-specific death.

So you have to go in the back of the paper to find out about all-cause mortality.

Correct.

And there's a huge debate at the moment about what should we do with lung cancer screening.

We talked about this earlier last year with NHS England rolling out case finding and trucks in high risk areas.

But there are two important papers for me that really influenced my practice and my thinking.

The first is going back to 2003.

There's a paper by Gergir Gorenza on using natural frequencies.

Well, instead of using percentages and trying to use these relative measures or statistics, what you're trying to say is in a natural occurring cohort of a patient, like 100 or 1,000 patients, what are you likely to see?

Now, the second paper that's really interesting is a paper by Steve Wollison and Lisa Swartz that says, these papers are dominated by relative risks.