While sat downstairs working on my laptop tonight, I was kinda catching up on TV in the background. I had some Sky+ programmes to watch – The Bill, Casualty, Supernatural – nothing particularly high-brow, but good background watching. So I ran out of programmes to watch and one that I was saving to watch with Nathan made me stop and think, so on it went.
The programme was a BBC4 documentary called “Can Dogs Smell Cancer?”. It originally aired on Saturday 11 March 2006, the day after my birthday, but I’d kept it until now.
Wow. Now, from my work with SARDA and with Arthur I know both how intelligent dogs can be as well as how sensitive their noses are. Indeed, the study shown in the documentary goes on to say that dogs showed abilities far beyond any man made equipment. However, when you hear of some of the statistics from that programme – the ability to detect smell at on part per trillion; in one study a 90% or more accuracy rate and the ability to detect cancerous cells in a five-year-old sample…these are truly incredible.
I was, however, disgusted with the impression that the Cancer Research UK spokesman gave – his comments and his demeanour in the documentary indicated that he thought this was nothing but a big joke – a comment echoed by many viewers on the “Have your say” page on the BBC. Cancer Reasearch UK, as well as the producer, responded to comments made by the viewers. The Cancer Research UK response was particulary interesting, although I have to admit that I think I’m going to withold donations from them until they start funding more of this kind of reasearch and stop poo-pooing it.
A fantastic documentary, highly recommended if you can get a copy.
One thing to keep in mind with any study of animal sensitivity is to keep the tests properly blinded. If CRUK believe (as I do) that the tests were unblinded, a 90% success rate with the handler knowing if it (or isn’t) cancer being checked translates to anything between 0% and 90% of the dog actually picking it up from the sample.
Thus the “90% success rate” doesn’t, in itself, say anything. I haven’t seen the programme in question, but none of the previous reports on cancer-sniffig-dog-studies I’ve seen imply they’re using a double-blind setup and the results are therefore essentailly worthless and CRUK are doing the right thing by not spending money chasing down probably-blind alleys. I could, possibly, see an argument for a properly-blinded test. Shouldn’t take more than a few weeks of researcher time…
I mean, when you train your SARDA dogs, you don’t let the handler know where the sought-for person is, do you?
The two studies I mentioned were both blinded – one double blind. The one year study (Cambridge) had a success rate of 80% until the dog decided to have an off-day. Problem is that they need more money to work out *why* it was having an off-day (could even be training issues).
The second study was done by the Pine Street Clinic and their results are on their site. Theirs was a 5-year study.
And no, the handlers don’t know where the person is. 🙂
That’s much better than any of the ones I’ve seen, where they use people diagnosed with and people not diagnosed with cancer as sniff targets. At least one of them had the treating doctor in the romo and at least one had the dog handler knowing who it was. With proper double-blinding in place, I’m happier for you to condemn CRUK. 🙂
Hmm, sounds really interesting, wish I’d seen it. I read the commments on the website and the CRUK person talks about developing electronic sniffers. I really don’t see the point in this, isn’t it far cheaper to just train dogs than put research into something that might not even work. I mean, for all we know it may be a combination of things that lead dogs to make a ‘diagnosis’, not just smell.
*sigh*
Oh, by the way, have been seriously considering coming up (down?) to S. Wales now that you’ve mentioned it, get in touch! 😀
Sean