About 40 percent of Americans belong to a racial or ethnic minority, but the people who participate in clinical trials tend to be more homogeneous. Clinical trials are the studies that test whether drugs work, and inform doctors' decisions about how to treat their patients. When subjects in those studies don't look like the patients who could end up taking the treatments, that can be problematic. In short: Clinical trials are too white.

"We've known for years that certain drugs don't work on parts of our population," says Sam Oh, an epidemiologist at the University of California, San Francisco Center for Genes, Environment and Health.

He was one of a group of 14 researchers who found that diversity in biomedical research often does not reflect the American population. The study was published Tuesday in the journal PLoS Medicine.

"Only 2 percent of cancer studies and less than 5 percent of pulmonary studies have studied enough minorities to provide useful information," Oh says.

Because of genetic differences, some people's bodies process drugs in very different ways. Racial and ethnic categories can serve as a proxy for those differences.

The blood thinner clopidogrel, or Plavix, doesn't work in the 75 percent of Pacific Islanders whose bodies don't produce the enzyme required to activate the drug. For them, taking the medication is like taking a placebo.

People with epilepsy who are of Asian descent are supposed to get genetic testing before being prescribed the seizure medication carbamazepine, because the drug can damage the skin and internal organs of patients with a certain gene variant.

And, says Oh, "African-Americans and Puerto Ricans don't respond as well to some of the most common asthma controller medications, and that's really a tragedy since these two groups are the most affected by asthma in the United States."

Coming across disparities like the asthma example, Oh says, "you begin to wonder, 'Well, why is this the case?' And part of that reason might be because our biomedical studies in the past have not recruited as heavily in those populations."

There are a lot of reasons why minorities are underrepresented in biomedical research, from limited access to specialty care centers where patients are often recruited to trials, to fears of exploitation in medical research. That's based on scandals like the Tuskegee Institute syphilis study, where researchers crossed serious ethical lines with minority subjects.

"There have been some bad experiences — some very bad experiences — which have appropriately led people in minority communities to have less trust in the research environment than they otherwise might have," says Dr. Michael Lauer, who oversees grant applications at the National Institutes of Health.

But, he says, things are looking up. "For example, between 2010 and 2014, the proportion of participants in [Phase III] clinical trials who are black has increased from 10 percent to 23 percent," Lauer says.

The participant population in all NIH clinical trials combined more or less matches the racial and ethnic makeup of the U.S. But when zooming in on trials of certain diseases, the demographics tend not to mirror reality as well.

The importance of increasing diversity among study subjects has been on the NIH radar for almost 30 years, when the Office of Minority Health was set up.

The review boards that decide which studies will get funding, however, remain 90 percent white, says Oh, based on information the group accessed with a Freedom of Information Act request. The community of researchers applying for grants also skews white. That's one of the things Oh and his colleagues say needs to change, because people are more likely to sign up for a clinical trial if the recruiter looks like them, or at least speaks their language.

"So it's really important when you want to do science in diverse communities, you have a scientific team and a scientific staff that is also diverse," Oh says.

After the NIH found that black applicants were 10 percentage points less likely than white applicants to win a research grant, a position was created last year to focus on increasing diversity in the field.

Oh and his colleagues point out that "the award rate for R01 or equivalent grants has been consistently lower among non-white applicants than white applicants" for about 30 years.

The tight NIH budget of late hasn't helped diversify study subjects. The shallow pot of grant money — which got a big boost from Congress on Tuesday — has encouraged researchers competing for funding to look for all sorts of ways to slim their proposed budgets.

"One of the ways, unfortunately, to reduce the budget is to take out money from recruiting diverse populations," says Oh. Recruiting minority participants can cost more because of the advertising, community outreach and translations that may be required to attract a broader subject group.

The researchers called for diversification among the committees that review grant applications, and evaluating minority recruitment as one criterion in determining the scientific merit of studies.

Copyright 2016 NPR. To see more, visit http://www.npr.org/.