updated 2/22/2006 7:03:31 PM ET 2006-02-23T00:03:31

Dr. Aubrey Blumsohn was stunned: Research results were submitted to a scientific meeting under his name, yet the British bone specialist insists he not only hadn’t written or reviewed the report, he wasn’t sure it was accurate.

  1. Don't miss these Health stories
    1. Splash News
      More women opting for preventive mastectomy - but should they be?

      Rates of women who are opting for preventive mastectomies, such as Angeline Jolie, have increased by an estimated 50 percent in recent years, experts say. But many doctors are puzzled because the operation doesn't carry a 100 percent guarantee, it's major surgery -- and women have other options, from a once-a-day pill to careful monitoring.

    2. Larry Page's damaged vocal cords: Treatment comes with trade-offs
    3. Report questioning salt guidelines riles heart experts
    4. CDC: 2012 was deadliest year for West Nile in US
    5. What stresses moms most? Themselves, survey says

The incident turned into a public feud when Blumsohn charged that the U.S. drug company paying for the study rebuffed his attempts to analyze the data.

It’s the latest in a string of controversies about pharmaceutical industry control of medical research, from hidden antidepressant risks to the undercounting of heart attacks in a critical study of the painkiller Vioxx.

Whoever pays for medical research — not necessarily the scientists who do the work — controls what doctors, and the public, learn about its outcome. Scientific journals, including one that published some of the reports Blumsohn now questions, are grappling anew with how to ensure that they print complete results.

“This is a major problem, for both researchers and scientific journals,” said Dr. Joseph Lorenzo of the American Society for Bone and Mineral Research, which has convened a task force to consider whether extra steps are needed to protect against “hidden biases” in industry-funded research.

It’s an important question, considering that the pharmaceutical industry provides about 70 percent of the financing for studies of medications in the United States.

Questions about that research started making headlines in 2004, when a Food and Drug Administration reanalysis of industry antidepressant studies concluded those drugs sometimes increase the risk of suicide in children.

Risks downplayed
Then Merck & Co. pulled its arthritis drug Vioxx off the market, after research found long-term use doubled the risk of heart attacks. Critics say the risk was downplayed until then — and last December, the New England Journal of Medicine revealed that it had published a 2000 Merck study that failed to disclose some heart attacks, making the drug appear less risky than later determined.

Reeling from bad publicity, the industry pledged to do better at revealing results of clinical trials. Editors of leading medical journals attempted to force them to do so, by declaring they would no longer publish results of any studies that hadn’t been listed in a public registry. And Sen. Charles Grassley, R-Iowa, is pushing for legislation to mandate full disclosure.

But Blumsohn, in a Capitol Hill visit arranged by the watchdog Government Accountability Project, is raising questions of more subtle influence — which Procter & Gamble, the company that funded his work, disputes.

P&G hired his lab at Sheffield University to analyze samples from thousands of women who used its osteoporosis drug Actonel. The goal: to determine what rate of bone renewal gives the most protection against fractures.

Missing data
E-mails that Blumsohn provided to The Associated Press suggest P&G denied him access to the patient data until months after results had been submitted in his name to the bone society. Blumsohn finally got a brief look in July 2003, only to conclude that about 40 percent of the data was missing, skewing the results, he said.

“We’ve allowed the basic rules of science to be flouted without a murmur from anyone,” contends Blumsohn, who is meeting this week with Grassley’s staff and drug regulators in Washington.

P&G spokesman Tom Millikin said Blumsohn “was provided full and unfiltered access to all of the data that was relevant to the work he performed.”

That appears to be in line with standards outlined by the industry’s Pharmaceutical Research and Manufacturers of America.

“This issue is about a relationship fraught with misunderstanding, and we sincerely regret that,” Millikin added, noting that Blumsohn willingly discussed the research at two medical meetings.

Blumsohn counters that he presented only data he could confirm.

'Veneer of credibility'
“Access to data means you’ve got the numbers. They redefined 'access to data’ meaning a company statistician would give you some tables,” he said. “These companies are using scientists, university scientists, to give their research a veneer of university respectability and credibility.”

The researcher’s case made headlines in Britain, and was raised in Parliament last December after the university suspended him. Blumsohn said it was for speaking to the press; reports at the time quoted the university as saying it had encouraged him to raise his concerns using proper channels.

Many leading scientific journals require researchers to affirm that they analyzed all the raw data, not averages or compilations from someone else. Yet a recent survey, published in the New England Journal of Medicine, of 122 universities’ standards for drug-company research found 17 percent reported disputes over control of or access to data. Also, they reported widespread disagreement over whether the companies that pay for research should be allowed to help write the results for publication, or insert their own interpretation of those results.

The American Association of Medical Colleges last month published principles for industry-funded research that affirm the importance of access to full data.

© 2012 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.


Discussion comments


Most active discussions

  1. votes comments
  2. votes comments
  3. votes comments
  4. votes comments