Peer review analysts lay into industrial sponsors


19 September 2005

The pharmaceutical industry has come in for severe criticism over its role in clinical research sponsorship.

Unrelated research groups have highlighted numerous concerns over research sponsored by drug companies. One group concludes that industry-sponsored systematic reviews of drugs should not be trusted.

Anders Jørgensen and Peter Gøtzsche of the Nordic Cochrane Centre in Denmark investigated whether systematic reviews carried out by the Cochrane centre shared methodology and conclusions with industry-sponsored meta-analyses of the same drug.

They concluded that, compared with industry-sponsored reviews, Cochrane reviews were more likely to have stated their search methods, had comprehensive search strategies, used more sources to identify studies, and made an effort to avoid bias in the selection of studies. 

'All reviews supported by industry recommended the experimental drug without reservations versus none of the Cochrane reviews,' concluded Jørgensen and Gøtzsche. 'Systematic reviews of drugs should not be sponsored by industry. And if they are, they should not be trusted.'

Federico Soldani and colleagues, from Harvard Medical School, have arrived at similar conclusions with preliminary results from the Clinical Knowledge and Research Methodology audit, a project analysing recent research reports in leading psychiatric journals. 

Soldani's team, which assessed 75 studies of drugs for treating bipolar disorder, report a positive 'funding-outcome' association. The team offers a number of explanations, including careful selection by companies of which studies to fund and which to publish, and a lack of appropriate comparator or dosage in industrially-sponsored trials.

'Even in highly respected journals, most reports on bipolar disorder are of limited methodological quality,' concludes Soldani, who fears that funding source might influence psychopharmacology research results, conclusions and dissemination.

Meanwhile researchers at Ohio State University and the University of Massachusetts have examined issues relating to the failure to publish clinical data. William Gardner at Ohio and Charles Lidz at Massachusetts contacted over 300 authors and discovered that about half of them had participated in trials that weren't published. Sponsors were more likely to prevent publication of a trial when no differences between their product and a competitor's were found than when differences were found, say the researchers.

'The present state of affairs could not exist without the collaboration, or acquiescence, of academic researchers,' said Gardner and Lidz. 'This should be changed.'

Data from these and other groups were presented at a meeting on peer review and biomedical publication organised by the Journal of the American Medical Association and the British Medical Journal Publishing Group. 

Richard Ley, spokesman for the Association of the British Pharmaceutical Industry, told Chemistry World that it is not in the industry's interests to skew the results of trials or other research.

'Companies themselves want accurate results from their research,' said Ley. 'They are in a long-term business, with a new medicine typically taking 10 to 12 years to develop at a cost of around £350 million. With so much at stake you simply cannot cut corners - quite apart from the ethical dimension. The industry is not interested in flawed, short-cut science.'

Ley added that it was not surprising that industry-sponsored trials often produce positive results, since drugs reaching this stage have passed many earlier hurdles and show pharmaceutical promise.

It is the responsibility of academics to ensure that they did not sign restrictive agreements with sponsors, said Ley. The global pharmaceutical industry is launching a register of clinical trial data, where results from all trials will be published on the internet regardless of outcome, he said. Simon Hadlington