A dose of our own medicine

A report this week struck right at the heart of what we do. Yes. it was another analysis of the world's medical journalism.

There was a conflict of interest but we did not hold back. We reported the story.

Like most such analyses, it concluded that most medical journalism in the mainstream media is flawed. That's no surprise. But the study, from Alberta, Canada, did make some interesting points.

It suggests four kinds of basic information that should be included:

• the number of patients;
• the methods used;
• the doseage;
• conflicts of interest.

Tania Bubela, who conducted the research, seemed to want to make the case that reporting is biased against alternative therapy. So I think that was why dose was important in her mind - as it is less standard in, say, herbal medicine than in mainstream.

Methods is an interesting challenge. Yes we always try and report the methods used. Frankly its tedious for the writer and the reader. The problem is that the "gold standard" is the randomised control trial. This means that neither the doctor nor the patient knows whether they are receiving the real pill or a fake - known as a placebo. As we saw in the case of acupuncture, creating a placebo can be problematic. It's okay with pills but with other therapies, how do you fake them?

There are other statistical methods, mainly involving the analysis of populations. These will tend to tell you about the impact of diet and lifestyle. They are rarely conclusive on their own but taken together, over the years, they can be pretty convincing. It was that kind of analysis that proved that smoking causes lung cancer - it took a long time.

And that's where numbers come into it. Yes, we would always report the number of people in a trial. It's very important. If percentage differences between a treatment working and not working are close, you want to see large numbers - thousands. If they are overwhelming smaller numbers, say hundreds, will do. The trouble is we still live in a world where some people have done statistics (and hence standard deviation etc) at school and others have not. Increasingly the younger generation, certainly in Britain, will have at least a smattering of statistics.

Finally there's conficts of interest. I touched on this in a recent posting. Frankly it's not always easy to know. But yes there are some stories which are disseminated actively by commercial interests. We are more cautious of these than others, perhaps. And we see sourcing as important. If we report a therapy, the report will generally come from a journal or a conference, possibly via a university. We will identify that source in detail. That's not usual in the mainstream media.