Drug testing

Do we have the right guinea pigs?

As America's makeup changes, doctors worry that drugs are being tested on the wrong mix of people

By ARTHUR ALLEN

There’s a WWII-era story that Sam Oh uses to explain the problems with most medical studies. During the war, U.S. military pilots lost control of their planes and crashed at alarming rates, even in training flights. Eventually, engineers discovered why: The cockpits had been designed in 1926 using average pilot height, weight, arm length and seven other measurements. But pilots were never average in all 10 dimensions—and only rarely in more than three. By designing the cockpit for the average pilot, the military had designed it to fit exactly nobody. After the introduction of adjustable foot pedals, flight suits, seats and head rest, the crashes pretty much stopped.

Oh, an epidemiologist at the University of California, San Francisco, sees in that story a painfully clear analogue to modern medicine. When new drugs are tested, they’re tested on an average population, and physicians are told about an average set of benefits and side effects. The patients who take them, though, won’t be average at all: They’ll be particular individuals. “Diversity is critical to good science,” he says.

The United States is an increasingly diverse country, to the extent that by 2050, whites will fall below 50 percent of the population. But you wouldn’t know that by looking at the patients in the clinical trials that are used to test medicines and new devices. Minorities still make up a disproportionately low percentage of participants in trials for cancer, asthma and other drugs, according to research by Oh and his colleagues.

In medicine, factoring in diversity is more than a matter of political correctness: It can be life or death. The anti-seizure drug carbamazepine causes Stevens-Johnson syndrome, a horrible and sometimes deadly skin reaction, in many people of Asian ancestry. Puerto Rican children often don’t respond well to albuterol, the emergency drug used to improve breathing for people having asthma attacks. Neurological side effects to the anti-HIV drug efavirenz vary based on genetic ancestry. Plavix, a top anti-platelet drug, often can’t be processed by Asians — although it is still on the formulary as the drug of choice in cities like San Francisco, where Asians make up 37 percent of the city’s population. In each of these drugs, racial differences in response were discovered long after they were approved and put on the market. They might have been found earlier if the clinical studies that led to Food and Drug Administration approval had been more racially inclusive.

“You need to have diversity to study as much of the disease spectrum as possible — differences in exposures, in risk, in genetics, in social and cultural differences,” Oh says.

As long ago as 1993, Congress passed a law requiring the National Institutes of Health to include more women and minorities in research studies. But while women are now overrepresented, enrollment of minorities has lagged. A recent NIH report put the total at 28 percent in 2015, but some research indicates the agency’s statistical methods skew the figures higher than they actually are. The underrepresentation will only be exacerbated as America becomes more diverse, unless scientists can overcome the cultural habits that make it hard for them to reach many groups in the country.

NIH says it is getting better at inclusion, and some new research projects have diversity built into the architecture. The National Institutes of Health’s Precision Medicine Initiative, a massive research study to enroll at least 1 million Americans in a database containing DNA, life histories and hundreds of other measures, aims to have 40 percent of the volunteers be people of color, to reflect America’s current makeup — a percentage that will presumably climb as the population shifts in the future.

But they’ll have their work cut out for them. Oh and the leader of his laboratory, pulmonologist Esteban Burchard, who run a two-decade study of genetic and environmental links to asthma among Latino and African-American kids, say their own findings raise questions about whether the NIH has reached the 28 percent of non-white trial participants that the agency claims. While inclusion may be growing, less than 2 percent of the more than 10,000 cancer trials funded by NIH over the past two decades focused on minority groups. Less than 5 percent of NIH-funded respiratory research reported inclusion of significant racial and ethnic minorities, according to Burchard’s findings. “We have been tracking and trying to be inclusive for decades, and we have not moved the needle,” concludes former NIH Deputy Director Kathy Hudson.

Historically, there are a number of reasons representative numbers of minorities don’t appear in research studies. For one, they often simply weren’t asked: Patients of color are less likely to be seen by the high-end academic specialists who are part of research networks. People who are too busy working and too poor to take time off tend not to participate in research; the same is true for people who don’t understand the project or can’t speak English well. People tend to trust researchers more when they come from similar racial or ethnic backgrounds, but blacks and Hispanics represented just 4 percent and 7 percent of doctorate degrees awarded in biomedical sciences in 2013 — and less than 2 percent of NIH-funded research leaders are black.

And African-Americans, especially, may also have avoided research, thanks in large part to a history of mistrust dating to the 1932-1972 Tuskegee study, during which 600 black men initially were studied, including 399 Alabama sharecroppers who were observed, untreated, and who were allowed to slowly die of syphilis. The study was conducted without the benefit of patients' informed consent.

THE EMPHASIS ON diversity reflects an interesting turnabout in America’s attitudes toward medical research. For the first two-thirds of the 20th century, scientists predominantly enrolled prisoners, the mentally ill and racial minorities in medical research. Participation in research was for the poor, while the drugs that came out of it were for the rich. It was only in the 1970s that scientists recognized the need to warn participants of the risks of medical research and condemned the excesses of trials at places like Tuskegee and the Willowbrook State School in Staten Island, where institutionalized children with mental disabilities were intentionally infected with hepatitis.

Since 1981, U.S. law has required research studies to be approved by Institutional Review Boards, committees of experts who examine testing protocols to be sure they’re honest and ethical. Among other things, scientists are required to clarify to research subjects that an experiment might provide a worse outcome than normal care. The turnabout in attitudes toward human subject research began less than a decade later, largely as a result of activism by suffering AIDS patients, who wanted the FDA to speed approval of drugs that might save their lives — and raised their hands to be the test subjects themselves. “Patients were saying, ‘Don’t protect us to death, we’re happy to be guinea pigs,’” said medical historian Steven Epstein of Northwestern University.

At the time, a lot of drug research involved white men aged 40 to 60 — people who looked like the researchers who studied them. But AIDS activism also increased among women, who had been left out of major experiments for cardiovascular disease — often because they were based at Veterans Administration hospitals where there were few women, notes Thomas Inui of Indiana University. Soon this awareness spread among minorities as well, he said: “They realized that [anti-hypertensive] beta blockers weren’t as effective at controlling high blood pressure in African Americans.”

Nowadays, well-off, well-educated cancer patients are the ones most likely to seek treatment in clinical trials. Rather than seeing themselves as vulnerable subjects being experimented on, they see clinical research as the place to get the best care and the most promising treatments (although clinical trials are still experiments, and they still usually fail).

So the problem has shifted, from research exploiting the underclass to research over-emphasizing white and upper middle-class America. Fixing that problem is more than a matter of imposing new ethical guidelines: It takes money and effort.

At the NIH, to ensure that its Precision Medicine Initiative hits its goal of 40 percent people of color, the agency is using a range of PR and recruiting tactics. Last year, it renamed the program “All of Us,” a name that sounds more like a comforting community program than a forbidding research enterprise. And for recruiting, the NIH is using a “trusted intermediary” strategy, says Janet Lambert, acting director of engagement for the project. Reaching beyond the major U.S. research centers, NIH is contacting churches, patient groups and advocacy organizations like the NAACP and the National Council of La Raza, and has budgeted $15 million over three years to help these groups enroll their members.

In African-American communities, even people who don’t know the specifics of Tuskegee inherited the mistrust it engendered, says Robert Winn, a leading cancer researcher at the University of Illinois-Chicago. “For me, the very first goal is to engage, be open and honest about their misgivings but explain how we’ve moved on,” says Winn, who has spoken to community leaders at storefronts and churches as a leader of the Precision Medicine Initiative in Illinois. “We can also say that if you aren’t involved in the [research], you aren’t visible. If those groups are invisible — minorities, rural people — they are increasing disparities, because we will have less understanding of these communities.”

More broadly, Burchard and Oh argue that it chiefly comes down to money, and that NIH should provide better incentives for researchers to enroll minorities in studies overall. “NIH funds me,” says Burchard. “I’m hitting them with a stick, but I’d like to have a carrot of budget increases so our studies could use more diverse populations.”

The cost of broadening a study population can be considerable. To produce valid answers about, say, a given drug, scientists need to enroll statistically relevant numbers of subjects, to prove that an effect isn’t there just by chance. Enrollment costs vary, but a 2015 PhRMA study estimates the drug industry spends $42,000 per patient in Phase III trials — the studies used to determine whether a drug is safe and effective enough to go on the market.

The numbers of patients, and the costs, grow each time researchers search for valid answers about a drug’s effect in a subpopulation — a different race or ethnic group or economic level, or among people who have a secondary illness. While such secondary illnesses — or “comorbidities” — are part of health care in the real world, investigators tend to avoid dealing with them in research trials, to simplify the variables under consideration.

This sometimes impacts minorities in particular. For example, a 2004 trial comparing tamoxifen and raloxifene for breast cancer initially had 12 percent minority enrollment, but only about 6 percent of those included in the final result were minorities. This was because a higher percentage of minority women had comorbidities such as diabetes or hypertension that excluded them from the trial, notes Doris Browne, a medical oncologist and former NIH investigator.

For the All of Us program, the NIH has a decent budget — it’s currently funded through a $200 million fiscal year resolution. Diversity won’t be a luxury as long as Congress continues to approve funds for the program, which would get an additional $300 million under the 2017 continuing resolution. The resolution is up for another vote in March, and its prospects are uncertain. “There is concern about implementation under the current administration,” Burchard says. “But this is a bipartisan issue — whether you supported Orangutan or Bernie [Sanders], we all develop diseases and would like to find ways around them.”

IN THE QUEST to make research more relevant for more patients, not every scientist and doctor thinks diversity in clinical trials is the most important area to focus on. David S. Jones, a doctor and historian of science at Harvard, argues that racial differences in response to drugs or treatment are relatively uncommon, and rarely clear enough to guide treatment decisions. Not even genetics tests are definitive in most cases, Jones says. While researchers have known for decades that patients with congestive heart failure respond to vastly different dosages of the blood thinner warfarin, for example, genetics testing alone hasn’t provided guidance helpful enough to change clinical practice, which still requires physicians to adjust each patient’s dosage based on how they respond to the drug.

The FDA in 2005 approved a much-ballyhooed heart medication, BiDiL, for African-Americans alone, based on the fact that it worked for black patients in a clinical trial. But while there is some evidence of different heart disease processes in African-American populations, BiDiL has sold poorly, and many critics believe its approval was a political rather than scientific process.

“I just don’t think there are that many significant differences between populations,” says Jones, who points out that there are political risks to such research as well. “You need to balance the marginal utility of finding them against the very real risks of reaffirming that races are so different.” Social concepts of race — which have to do with skin color — have little to do with biological differences, which are highly variable among and between what we think of as “races.” And in any case, the greatest biological differences in the world exist within the populations of Africa.

“If the NIH really wanted to look for genetic differences, they’d want to compare, say, South Africans with Nigerians and Ethiopians,” he said. Genetic differences shouldn’t be ignored, but “fascination with high-tech tools might divert attention from approaches that could be more significant.”

He also points out that other factors — such as lifestyle and family income — may be far more important to a patient’s treatment. “The social scientist in me tells me that one reason the normal dose of albuterol doesn’t work in an African-American kid in the inner city is because his family hasn’t been able to afford the medication consistently,” Jones said. “Physicians would much rather talk about genetic variations than those kinds of social questions.” Noncompliance affects 50 percent of patients and is probably higher among families that lack economic resources, he notes. “A $2 billion moonshot for non-adherence would provide much better return on investment — but no one wants to do it.”

If the Precision Medicine Initiative succeeds, it could answer some of these questions. It is aimed at taking all health factors into account — genes, to be sure, but also environmental elements, and the combination of genes and environment known as epigenetics — the science of how factors outside of DNA, including exposures to toxins or stress hormones or foods — can affect how genes are active in the body. This may help explain why identical twins share many traits, but not all of them, for example.

“What I really like about All of Us, in particular, is the emphasis on different types of diversity,” says Winn. “It’s broad in scope in the best way in terms of getting all classes, races, ethnicity, regions, cultures and all levels of education and technology.”

“You’ll get no argument from me that your ZIP code will determine your outcome more than your genetic code. But your genetic code is not insignificant,” adds Winn. “Let’s stop having the “or” conversation and make it an “and” conversation. Health, he says, is about "access to clean water, less stress AND access to travel, hospitals and the miracles that science has to offer.”