Too cozy for comfort

LOTS OF WARNING signs have appeared recently suggesting that clinical trials to determine the effectiveness of new medicines may not always be conducted on the up and up.

Cozy relationships between academic and government researchers who perform drug studies and the pharmaceutical companies that underwrite their costs encourage a bias toward favorable outcomes.


Such bias could show up in the way a test is designed; it could affect the way test results are released; it might also be responsible for squelching unfavorable results altogether - as in the recent example of test results that were deep-sixed suggesting the antidepressant Paxil encouraged suicidal thoughts among teens.

Pressure is building within the medical community and among consumer advocates to create a more transparent process, in which potential conflicts of interest are disclosed and the results of all clinical trials conducted in the United States would be available in a master registry.


Tough measures are long overdue, and should be quickly embraced by both the Bush administration and the pharmaceutical industry. The voluntary guidelines currently in place are clearly inadequate to the task of preventing the science from being tainted by the huge financial stakes involved.

As a practical matter, there's no way to remove the taint entirely. The collaboration among government agencies, academic laboratories and pharmaceutical companies in the development and testing of new medicines has gone on for three decades. In 2002, a total of $5.6 billion was spent on clinical trials involving human subjects, 70 percent of which came from the drug industry.

It makes sense that the industry should bear most of the cost for product testing. But in recent years, as drugs have played an ever larger role in health care and the industry has scrambled to expand and update its products, the line between researcher, regulator and investor has blurred considerably. Government and academic researchers are often paid subsidies by drug companies, and may be inclined to set up a test that yields the most favorable results.

Editors of medical journals, who publish reports on clinical trials often used in pharmaceutical marketing, are attracted by what's new and interesting rather than showcasing dud drugs - and are limited to publishing what's made available to them, in any case. On some occasions, they have published research reports that failed to disclose a financial link between the researcher and the success of the drug he's touting.

The American Medical Association wants to put into place new safeguards that attempt to screen out bias from clinical trials, and at a minimum announce any potential conflict of interest by those involved.

Even more important, though, is the AMA's call for a mandatory registry of trial results. That would ensure the information is available even if the findings don't suit the interests of the drug company footing the bill.