The issue of evidence on the Breastlight breast monitoring product for women has got me thinking about how much evidence one should have before an intervention is declared to work. In an ideal world, there would be continuous monitoring of all interventions so that very late or very rare complications would show up. In the real world that doesn’t often happen and we have to make do with initial studies and then voluntary reporting of problems thereafter.
So the quality of the evidence at the start would seem important. Here’s another Boots product: advanced electronic cold sore treatment which says it ”reduces the duration of the attack and speeds up the healing time”. It’s a good concept, and certaintly cold sores are common and often unpleasant. So does it work?
Unable to find anything about this product on PubMed, I contacted Boots who sent me this paper: A randomised double-blind study comparing the effect of 1072-nm light against placebo for the treatment of herpes labialis
The full paper isn’t available for free, but here is the basic flowchart: 32 volunteers were recruited. 18 got a placebo treatment, 14 got the active device. 4 were lost to follow up, and 1 was excluded for other reasons. That left 15 in the placebo group, who had a mean healing time of the sore of 9.4 days, and 12 in the active group, who had a mean healing time of 6.33 days. The concluding paragraph of the paper notes that further studies would be of interest, presumably because this is a very small trial which relied on people self reporting when their cold sores were better.
Is it enough evidence on which to market a product? There is nothing wrong with this trial, but it is small and the findings have not, as far as I know, been replicated. Boots say this is the only trial evidence there is – they also declined to tell me how many of the devices they’ve sold. It would be possible to do a large trial of this device. Shouldn’t this be done before it’s marketed?
Margaret McCartney
Clive Cookson
Andrew Jack