Pediatrics data buried if negative, study says

The unpublished data are invariably material that questions the effectiveness of new treatments, leaving parents and scientists alike with rose-coloured views of reality and treatments that are not nearly as effective as others.

"If you want to know if a new therapy works, you're often going to get a biased view by reading the scientific literature," said Terry Klassen, chairman of the department of pediatrics at the University of Alberta in Edmonton. He is lead author of the research, published in the current edition of the Archives of Pediatric & Adolescent Medicine.

Dr. Klassen said there is no evidence that self-interest groups, such as pharmaceutical companies, quash the findings; rather, a series of factors is at work, including competition for space in prestigious medical journals. But the problem has its roots in human nature.

"Everybody wants new therapies to work because that's exciting," Dr. Klassen said. "Finding out they don't work is a lot less exciting. But scientifically, it's just as important." If treatment is to be based on published literature, the literature must include all available data, he said.

Such bias is not new, and it is not restricted to pediatric research.

Ian Tannock, an oncologist at the Princess Margaret Hospital in Toronto, found that one-quarter of all studies on large groups of cancer patients are not published. Again, almost all the unpublished material feature negative results. Most of the drugs and treatments discussed in the unpublished material enter the market anyway.

Dr. Tannock said that this failure to publish violates the tacit agreement between the researchers and patients enrolled in clinical trials.

He told delegates at the recent conference of the American Society for Clinical Oncology that this is an ethical breach and a "form of scientific misconduct."

This view, common among medical researchers, fuels the movement to counter publication bias. A British company will begin publishing The Journal of Negative Results in July.

"We hope to revolutionize scientific publishing," said Jan Velterop, publisher of Biomed Central in London. "We believe that scientific results belong to the world, and they ought to be shared with everyone."

The Journal of Negative Results will be accessible on-line for free, unlike most medical journals that have hefty subscription fees. The journal, will charge $500 fees to researchers, double blows to those whose work has been rejected.

Another initiative, the Cochrane Collaboration, is an international organization that aims to help people make well-informed decisions about health care by preparing systematic reviews of the effects of health-care interventions. Part of its work includes creating a database of all clinical trials, good and bad. To date, it has compiled the results of more than 300,000 trials.

Dr. Klassen said both these initiatives are welcome. But ultimately, what is needed is an international registry of clinical trials and a means of tracking so-called "grey literature" -- including government reports, research from think tanks and theses.

"We need some kind of central repository for information, to ensure we get the best possible data available," he said.

To conduct his research, Dr. Klassen and his team tracked the findings presented at the Society for Pediatric Research for three years. They found that 59 per cent of 447 abstracts presented eventually were published.

But 64 per cent of research with positive results about new therapies was published, compared with 43 per cent of those with negative results.

This story ran on Page A8 of the Globe and Mail, Print Edition