09 janvier 2007

Researchers see bias in private-funded studies

Polly Curtis, health correspondent
The Guardian

Research into the health benefits of drinks including fizzy pop, juices and milk may be severely biased in favour of food industry funders, American doctors say today. A survey of research on the nutritional value of drinks found that studies funded entirely by food and drink companies were approximately eight times more likely to produce results favourable to their funders, compared with studies which had no industry funding.

The findings threaten to revive the row which started in the pharmaceutical industry about how independent scientists can be when they receive funding from a commercial source. The authors of the review of research conclude: "Industry funding of nutrition-related scientific articles may bias conclusions in favour of sponsors' products, with potentially significant implications for public health."

The researchers, based in Boston, reviewed 206 articles which examined the nutritional effects of drinks and found a "significant" bias towards positive results in the research which was funded by the food industry.

"Whereas bias in pharmaceutical research could have an adverse effect on the health of the millions of individuals who take medications, bias in nutritional research could have an adverse effect on the health of everyone," they said.

David Ludwig, lead researcher and specialist in childhood nutrition at the Children's Hospital in Boston, said: "We found that the industry-funded studies were up to eight times more likely to be positive than the publicly-funded ones. In the aggregate we now have evidence of a bias and that bias could have a substantial impact on human health."

The paper, which was published in the peer-reviewed public access journal PLOS Medicine, suggested that companies were more likely to sponsor research which would show their product in a good light; investigators might consider a question which would "sell" to the funders when starting a study; authors might be selective in what they include to meet their funders' agendas or may over- or under-represent findings accordingly.

But other scientists said that the high incidence of positive results in industry-funded research was because they were more likely to fund research that was useful to their aims.

Susan Jebb, a senior scientist at the Human Nutrition Research department at Cambridge University, which receives taxpayer funding via the Medical Research Council, said that industry funding was "essential" as long as there was a strong understanding that it did not influence the outcomes of the research.

"There isn't enough public funding around so we welcome industry funding to help. The public should be able to have faith in the scientists to do research properly and fairly," she added.

Rappelons une fois de plus que le mélange des genre n'est jamais bon en sciences. L'origine des fonds qui financent une étude scientifique fait partie des critères d'évaluation de la qualité de ces études. Le conflit d'intérêt est une des sources importantes de biais positif et de biais de publication.

Aucun commentaire: