A rethinking of the U.S. clinical trials system is needed to truly understand the risks and benefits of current and future devices and drugs, Duke Clinical Research Institute Director Robert Califf, MD, asserted at the FDA Science Forum April 28.

"I think our system of evaluating medical technology is built on a model that we now know is wrong," Califf said at the Washington, D.C. event. Califf, who is also associate vice chancellor for clinical research at Duke Medical Center, further characterized the problem. Doctors, he said, are "trained to believe that if they know a little biology, they somehow understand the balance of risk and benefit of using therapeutics."

Califf previously has suggested that extensive reliance on surrogate biomarkers as endpoints at the expense of statistically robust clinical measures is a misdirected approach.

At an April 22, 2004 FDA Science Board meeting on the then-new "Critical Path Initiative," he warned: "Things that we think ought to work based on surrogates, not only don't work - they kill people."

FDA's Critical Path encourages the use of biomarkers to speed clinical trials (¹ Also see "FDA Updates CLIAC On Diagnostic Guidance Pipeline, Co-Development Trends" - Medtech Insight, 28 February, 2005., p. 29).

Califf favors a greater reliance on "simple" clinical trials of large populations to produce long-term data geared toward endpoints such as major illness or death.

The rationale for straightforward trials, which often have broader eligibility criteria, is to provide a better real-world picture of therapeutic outcomes and lend greater statistical weight to modest treatment differences.

During the FDA Science Forum, he cited the Women's Health Initiative's 2002 published risk/benefit study of hormone replacement therapy that involved more than 16,000 women as a good starting point. "That was a transition from...jury-rigged studies to support marketing positions to a true depiction of benefit and risk," he opined.

The Duke physician invoked devices in his argument that an overhaul of the clinical trials infrastructure is needed. He referred to a dialysis device, for instance, that was 510(k)-cleared. However, a large, simple study conducted to attract a government grant suggested higher mortality in patients who used the device versus those who did not.

Califf says those larger postmarket studies are both technologically and financially feasible. "We are now in a state where the limitation is our culture, not the technology," he argued.

Current computing capabilities are sufficient to handle the large amount of data from very big trials, he contended. Furthermore, such data management "is entirely possible with the current amount of money we are spending in the evaluation of drugs and devices, for most of what we do to define the benefits and risks - not only acutely, but chronically," he suggested.

FDA, including CDRH, has been closely examining methods of postmarket product surveillance, particularly after Merck's market withdrawal of Vioxx (² Also see "Drug Safety Oversight Initiatives Spur CDRH Postmarket Soul-Searching" - Medtech Insight, 28 March, 2005., p. 3). However, comments by some FDAers suggest that postmarket trials of the type suggested by Califf in many cases might not be feasible.

The authority to mandate a firm to perform such trials is limited. The agency can use Sec. 522 of the FDA Modernization Act of 1997 to compel companies to execute a study.

Still, initiation of a Sec. 522 study requires a strong "burden of evidence...almost like a search warrant," CDRH Office of Device Evaluation staffer Neal Muni said at the science forum. Consequently, Sec. 522 "is something that is not often used." More effective use of this tool has been advocated recently by Office of Surveillance & Biometrics Director Susan Gardner (³ Also see "Regulatory News In Brief" - Medtech Insight, 25 April, 2005., p. 22).

For now, CDRH is focused on improving the completion rate of condition-of-approval studies. Other barriers to postmarket trials also exist, however.

"Clinical trials may not be feasible, either ethically or from a resource perspective, in the postmarket period," CDRH OSB Division of Postmarket Survaillance Director Tom Gross, MD/MPH, said. "There are lots of issues attendant with the overall use, and clinical trials may not be the vehicle to address those issues."

Larger Clinical Studies Required To Assess Device Risk/Benefit – Duke’s Califf

More from Archive

More from Medtech Insight