In labs all around the world, big data have researchers "acting like kids in a candy store," the NIH's Charles Rotimi said at an FDA meeting on enhancing demographic subgroup data for drugs, biologics and devices.
But whether that store is offering sufficient variety is a matter of opinion as researchers, regulators and patient advocates try to work out how to broaden the offerings. Many of the people speaking at Monday's meeting wanted much more inclusion in clinical trial data based on sex, age and ethnicity. But others tried to push the conversation beyond simple, self-reported labels to the science behind why one person may respond to a treatment while another one doesn't. The result was more questions than answers.
In opening the meeting, FDA Commissioner Robert Califf said the challenges today are figuring out how to interpret which differences matter and what they mean in specific diseases. Other questions are what sample sizes are necessary to extrapolate data across a subset and how modeling can be used when it's confounded by complex demographic characteristics. He pointed to the term "Hispanic" as an example. For the most part, it is an American term that has little meaning elsewhere, and it encompasses a broad range of ancestry, he said.
Further developing that thought, Rotimi maintained that ancestry is more important than self-identified race or ethnicity in medicine. He reminded the group that in the real world, doctors treat individuals – not an "average" of a subgroup. When it comes to disease, it is the message encoded in a patient's DNA that matters, he said.
However, Rotimi pointed out that close to 95 percent of genomic studies are focused on European ancestry. That's backward, he said. Being more inclusive of broader ancestry isn't a matter of political correctness. It's a scientific imperative, he added.
SOCIAL JUSTICE OR SCIENCE?
John Whyte, director of professional affairs and stakeholder engagement at the FDA's drug center, suggested that questions such as weight or renal function may be more important in some instances than sex, age, ethnicity or even geography. "Is it a social justice issue or a scientific discussion?" he asked. A scientific hypothesis should drive trial inclusion criteria, he said.
Several speakers during the public hearing insisted it's about social justice. Okey Enyia, a staff worker on Capitol Hill, insisted, "Social justice is a tenet of public health." Deploring the paucity of black men in clinical trials and noting that the FDA has an office of women's health, he asked if the agency had plans to set up an office devoted to health issues faced by black males. "Black male lives matter," he said. (The FDA does have an office of minority health.)
Adolph Falcon, of the National Alliance for Hispanic Health, called for the FDA to require drug and device makers to do a better job of representing the U.S. population in terms of sex, age and ethnicity in clinical trials. A genetic database must better reflect the population or else health disparities will continue, he said.
Other speakers echoed Falcon's call for a diversity mandate, insisting that the FDA shouldn't approve drugs and devices that aren't adequately evaluated in minorities and women. Another suggestion was to require labeling indicating that a product wasn't sufficiently tested to determine its risk-benefit profile in specific subgroups.
VARIABILITY THAT COUNTS
Such comments led to questions of whether adequate diversity in a trial should represent the overall U.S. population or the prevalence of a disease in certain subgroups. Given that subset data may be more important in some diseases or conditions, the FDA's Robert Temple asked when the FDA should require a certain proportion of sex, age or ethnicity inclusion.
At the same time, he wondered how the FDA could deny approval to a drug or device that demonstrated a substantial improvement in care but didn't include the right number of participants from various subsets. And if a life-saving drug or device were approved, would it be ethically feasible to do a randomized, controlled postmarket study of its efficacy and safety in a minority population? he asked.
How do doctors know the treatment would be life-saving for a particular group if it wasn't tested in that group? countered Rita Redberg, a professor and director of women's cardiovascular services at the University of California-San Francisco. She referenced a clinical trial of an implantable cardioverter defibrillator that enrolled only men. When she asked the sponsor why no women were included, she was told that women don't have heart failure.
Although women make up half the population, they account for only 20 percent to 25 percent of most cardiovascular trials, Redberg pointed out. The exception is hypertension trials that enroll up to 80 percent women.
Jocelyn Ulrich, the assistant vice president for science and regulatory advocacy at the Pharmaceutical Research and Manufacturers of America, said requiring clinical trial populations to represent the overall U.S. population could extend phase III trials 20 months or longer and it wouldn't necessarily provide a sample size large enough to power a subgroup analysis.
Also, mandating inclusion based on self-identification would be a poor proxy for factors that truly matter in health outcomes, including other medications or conditions a patient may have, genetics and geography.
Clinical pharmacology and biology should be used to establish the plausibility of a variable response rather than implementing a mandate, Ulrich said.
Allison Kalloo, founder of #iParticipate, countered Ulrich, saying that inclusion based on ethnicity is a "matter of social justice" and that "more is always better."
"Changing our thinking is hard," the FDA's Barbara Buch said. As she closed the meeting, she asked a few more questions: Is the FDA asking the right questions? Is it using the right labels or does it need to redefine them? To move forward in improving the variability of the data, the FDA has to listen to others, Buch said.