Remember when doctors used to sell cigarettes? Those hilarious 1950s ads with father figures in white coats recommending Chesterfields for your throat? You can’t get away with that now – Carol Vorderman got the sack from Tomorrow's World for trading on her flimsy scientific authority to shill for Benecol, and even the mildest medical claims have to be couched in ever more elaborate disclaimers to avoid opportunist litigation. There is, however, one place where no such constraints seem to apply – the world of medical communications agencies.

It works like this: when a drug company has a new product, it needs to go through a huge range of clinical trials to get approval. Even afterwards the company will fund trials to compare the product's performance against similar drugs in the market or to see how factors such as age, sex and secondary illness affect it. If this data turns out remotely favourable, they will want to make sure that it gets publicised. This is where the agencies come in. From the raw data their writers will put together something that looks, smells and tastes like a scientific paper. It will even have been vetted every bit as carefully as a real paper, but instead of accuracy the agency’s contact in the drug company will be looking for the rigorous incorporation of the drug’s ‘key marketing messages’.

Of course, the paper still has to go through peer review, but this is where the medical writers really earn their pay. You still can’t claim that the results say more than they do, so the trick is to use scientifically meaningless phrases such as ‘approaching significance’[1] and to make huge logical leaps. Imagine that you’ve been paid to write that drug X is of more benefit to the elderly than drug Y, but all you have to work with is a study of Mexican teenagers. Well then, you mention that age could have been a factor in the results, as previous studies have indicated that… and so on. Competitor drug Z isn’t mentioned in your study at all, but you know its big selling point is that it gets metabolised quicker than X so you get a quick dig in about ‘limiting the study to drugs with long-term efficacy’. Sometimes you don’t even need research data – get enough papers together and you can write a ‘review article’ that’s been subtly spun in your client’s favour. Once you’re done, send it to one of the professors you’ve got on your books, and as long as they approve you can slap their name across the top like a badge of honour. Then bundle it off to your journal looking for all the world like a piece of research instead of a piece of advertising copy.

Journals such as the Lancet and the British Medical Journal are entirely wise to the process and will bounce these papers without blinking, so they tend to appear in lesser journals eager for material. The internet, however, has made them as accessible as their higher-powered counterparts, and because papers from medical communications agencies tend to cite each other extensively, search engines based around ‘relevance’ or ‘impact’[2] will give them priority. Busy doctors with limited resources are being misled by a forum designed to earn their trust. They may be aware of the problem but if they can’t tell whether what they’re reading is science or advertising, that awareness can do little more than increase their general uncertainty. The publication of scientific research is widely perceived as a bastion of factual integrity. In issues ranging from climate change to the dangers of MMR the most impassioned rhetoric can in theory be beaten hands down by bringing peer-reviewed papers to the debating table. Medical communications offers the corporate world a way to capitalise on this integrity, tarnishing it (as ever) in the process.

References
[1] Significance in clinical trials refers to a stringent set of statistical criteria. The phrase ‘approaching significance’ in a clinical trial paper has no more meaning than the phrase ‘slightly unique’.

[2] 'Impact’ is also in this case a technical term used in the scientific community to assess the importance of both papers and journals. It essentially measures the number of times papers from a particular journal have been cited elsewhere.