Pharma-sponsored trial result reporting biased: report

A survey of Australian medical research specialists has found that bias and manipulation in reporting of clinical trial results are rife when the trials are sponsored by pharmaceutical companies.

An editorial in the June 6 online edition of the Medical Journal of Australia warns that the increasingly close relationship between academic research and the pharmaceutical industry is putting patients at risk by allowing unsafe drugs to reach the market.

Prof Peter Gotzsche, director of the Nordic Cochrane Centre in Copenhagen, in Denmark, said the research agenda now predominantly served the interests of companies, not patients.

"Even when the results for the active and control therapies are no different, industry-sponsored trials come to a positive conclusion in favour of the sponsor's drug five times more often than do not-for-profit-sponsored trials," he wrote.

Gotzsche was reviewing the results of a major survey of Australian medical specialists, conducted in 2002 and 2003 by a team led by David Henry, Professor of Clinical Pharmacology at the University of Newcastle.

Among the members of the study team was Dr Graham McDonald, external licensing coordinator for Merck, Sharp and Dohme (Australia), whose parent company, Merck & Co, is facing the threat of a major class action for damages, after it withdrew its anti-inflammatory drug Vioxx (rofecoxib) from the market amid evidence that it may have caused thousands of premature deaths from heart attack and stroke, because it increased the risk of blood clotting in some patients.

As evidence of the serious consequences that could arise from bias in the analysis of trial results, Gotzsche said a meta-analysis (an extensive review of the research literature on a specific subject) supported by Merck had concluded that there was no increased risk of arterial thrombosis with rofecoxib.

Harm down-played

But another, independent meta-analysis, conducted without industry sponsorship, had concluded there was an increased risk of thrombosis associated with rofecoxib -- moreover, this risk had been apparent in research papers available to the authors of the industry-sponsored meta-analysis four years before Merck pulled its drug off the market.

"Such down-playing of harms in published papers has often required the collaboration, or acquiescence, of academic clinical researchers," Gotzsche wrote.

He noted that the industry's share of biomedical research activity had risen from 32 per cent to 62 per cent in the US during the past 20 years.

Gotzsche said surveys had shown that company-sponsored clinical trials were being manipulated in a number of ways. In some cases, results were not published at all, while in others, the control treatment was commonly disadvantaged by design, analysis or interpretation.

The Henry survey had reported important breaches in research integrity in industry-sponsored research, and for several reasons, the prevalence of the problems it reported probably represented only the tip of the iceberg. For starters, only 823 (39 per cent) of the 2010 specialists who had been contacted to participate had completed the questionnaire.

While only 9 per cent of respondents had reported one or more episodes of potentially serious research misconduct, the authors noted that this was equivalent to 21 per cent of those who had an active research relationship with industry.

The authors' criteria for assessing misconduct or bias had not considered protocol changes to be serious research misdemeanours. But on this reading of the results, at least one primary outcome was changed, introduced or omitted while research was actually in progress in 51 out of 82 trials -- or 62 per cent of cases.

"We think this is a serious problem, as, with a median of 27 outcomes per trial, one would expect one outcome to become significant by chance, even if the compared treatments were identical," Gotzsche said.

He said the MJA's own survey of research triallists denied the existence of unreported outcomes, despite clear evidence to the contrary. "We did not reveal to them until later that we had access to their trial protocols through the scientific ethics committees."

Gotzsche said research misconduct, and bias through intervention in research, could be markedly reduced if all trials were required to be registered at their inception, as a condition of them being allowed to proceed, and all trial protocols and data were freely accessible. Measures to this end were already underway.

But Gotzsche said it would be even better if drug-testing in patients was a public enterprise -- whether financed by industry or not -- with blinding during data analysis and manuscript writing.

"It is clear that governments could save money and treat patients better by investing much more in trials and academic trial centres than by relying on industry's own trials and conclusions," he wrote.

"Who would buy a washing machine that is five or 10 times more expensive than other machines just because its manufacturer has compared it with other machines and claims that it is the best? Unfortunately, such absurdities are often seen in healthcare, and are allowed to happen even in the absence of any direct head-to-head comparisons."