The new feature of the antidepressant drugs of the 1990s was that they had milder side-effects than their predecessors. Combined with aggressive marketing, this meant that annual sales in Sweden increased from just under EUR 18 million to over EUR 100 million in the space of just a few years.

Most countries have an established system for self-regulation of pharmaceuticals advertising. Sweden has been held up internationally as a good example. The Swedish pharmaceutical industry's trade organisation appoints an information examiner to monitor the marketing to ensure it doesn't violate established ethical guidelines. There is also a board appointed by the industry that deals with complaints. Misleading advertising usually results in a fine.

"The system of self-regulation was established in 1969 in Sweden, but this is the first academic study of how it really works", says Shai Mulinari, who is a researcher in both social sciences and biomedicine.

Together with a colleague, he has gone through all the advertising for antidepressants published in the professional journal for doctors, The Swedish Medical Journal, between 1994 and 2003. In the study, they found that 34 per cent of all adverts had been judged as misleading by the industry's self-regulation system.

"The figure should really have been even higher because we discovered that a large amount of misleading advertising had passed through the review process unnoticed. For example, many advertisements that were not picked up contained exaggerated claims about the effects of the drugs", said Shai Mulinari.

The fact that one third of all advertising failed the review process can, on the one hand, be regarded as a sign that the industry's self-regulation functions fairly well. On the other hand, in Shai Mulinari's view, the system can be perceived as toothless:

"The consequences of being convicted were marginal. In total, only 0.009 per cent of sales revenue went to fines for unethical marketing", he said.

Today, the level of the fines has been raised, but Shai Mulinari still believes there is reason to look more closely at how well the industry's self-regulation of advertising really works.

For example, according to the Lund University study it took an average of 15 weeks from the publication of the unethical advertising to the announcement of a verdict. During that time, the advertisement could in theory have been published in the journal 15 times. Another problem identified by the researchers was that only 12 per cent of the reports were initiated by doctors and only 8 per cent by the Medical Products Agency. In other cases, it was business competitors who reported one another or the reports were initiated by the industry's information examiner.

"It is important that doctors and the Medical Products Agency report impropriety, otherwise all responsibility is placed on the industry", said Shai Mulinari, who thinks the review process for the pharmaceutical industry's marketing is discussed far too little in Sweden compared with many other countries.