I’ve been very interested by this study, which has got a lot of coverage today. It’s a meta-analysis of studies on anti-depressants and apparently shows that they don’t work. Or rather, to be more scientific about it, they don’t perform clinically significantly better than placebo, except in the most severely depressed patients. And even there, this appeared to be because the placebo effect is very weak in these patients, rather than because the anti-depressants are more effective.

My undergraduate degree was in pharmacology, and I remember seeing papers showing that anti-depressants were not that much better than placebo. So why has this new paper been able to say that the difference is even slighter than that? Because they used Freedom of Information legislation to get unpublished clinical trial data.

As you may or may not know, publication bias means that negative or neutral results are less likely to be published, so published papers offer a misleading picture of how well drugs work. (Papers from PLoS and the BMJ on this, plus a more readable account from Ben Goldacre) And this is particularly likely to happen when drug companies have paid for the research. It seems like anti-depressants have pretty much no effect but because this bias means they looked like they did, millions (billions?) of pounds of NHS money has been spent on buying them.

Now this seems to me an excellent How Science Works issue. It’s thought-provoking and shows that there are limitations to science, but that we can understand and study them. It has concrete effects that we can understand and connects with ethical, social and economic issues. And it leaves open a discussion about what we can do about this. But is it a good basis for a lesson, or is it too esoteric for year 9s and 10s to grasp? Teachers, feel free to give me your thoughts on this.

I admit I am a bit obsessed with thinking about what should go into the teacher packs at the moment. I have a lot of ideas that I get excited about but then have to discard most of them. I have to be realistic that we can’t do everything and things have to be simple. But then to come up with really good ideas you have to sift through a lot of rubbish usually. Or you do in my brain anyway.

In the next few days I’ll start working with the teaching materials development team in earnest and things will start firming up. This will be a relief as hopefully I’ll stop breaking off in the middle of conversations with people to excitedly write notes to myself. Then five mins later break off again to frown at the notes and cross them out…

In the meantime, any bright ideas for activities, or pleas for things for us to cover, gratefully received. Top request so far in terms of topic is for something on reliability of data (and how it’s not the same as accuracy). Top request for skills development is activities which teach students how to debate and discuss issues, perhaps by modelling and leading them through a debate, “It’s all very well the curriculum telling us that you’ve got to have all these debates with the kids, but they don’t actually know how to do it.”