09.06.2021 07:00 AM

What Medical Researcher Training Can Learn From the 'Yellow Berets'

A new study shows how the Vietnam War–era initiative shaped a generation of star physician-researchers—and offers lessons that can be applied today.

Around 2015, two researchers who study the biomedical workforce found a treasure trove in the National Archives: a stash of applications from a Vietnam War–era program for young doctors.

The program, established in 1953, brought recent medical school graduates to the National Institutes of Health for two or three years of intensive research training, where they would learn to solve problems that would improve patient care and public health. During the war, applications increased dramatically, as people sought to fulfill their military service obligations through the program. Many of them later referred to themselves as Yellow Berets, a self-deprecating nod to the idea that they had avoided becoming Green Berets.

The highly competitive selection process resulted in an extreme concentration of talent at the NIH. Nine of these researchers later won Nobel Prizes. The development of cholesterol-lowering statins and the human papillomavirus vaccine were just some of the fruits of the research careers launched through the program. Anthony Fauci, the nation’s leading infectious disease expert, is also a graduate.

The program, called the NIH Associate Training Program (or NIH ATP), had been a success, but it was also unusual, because its leaders had been able to cherry-pick the very best med school graduates during a war. That raised some knotty questions: Would these scientists have made great discoveries even if they hadn't participated? And if something about it had made the Yellow Berets great—and could be used to improve modern research training programs—what was it? To say something about the program’s influence, social scientists really needed a control group to compare it to.

These questions went unanswered for generations, until MIT Sloan professor Pierre Azoulay and Census Bureau principal economist Misty Heggeness discovered, with the help of the NIH archivist Barbara Harkins, that stash of 3,075 application cards from decades ago. Harkins and Heggeness digitized them, and then reviewed them thoroughly until they felt confident they had a complete set of cards from 1965 to 1975, representing applicants who joined the program as well as those who had been cut after the first round of interviews. “And now, suddenly, we're in business, because we had found a control group," says Azoulay—the runners-up who were in med school at the same time and had been qualified enough to make it to the first screening, but ultimately did not participate.

Then began a very painstaking effort: Azoulay, Heggeness, MIT PhD candidate Wesley Greenblatt, and a team of helpers Googled each of the applicants who interviewed for the program during that decade. They tracked each person’s achievements, publications, patents, citations, grants, and prestigious appointments (for example, to the National Academy of Sciences), and consolidated this information into a searchable database. In a study published in July in the journal Research Policy, the team came to their conclusion: The Yellow Berets were twice as likely to pursue research careers than the controls; they published more (and more highly cited) papers, and they disproportionately participated in training the next generation of physician-scientists and researchers.

"That's very unique," says Azoulay. "I don't know of any other dataset where you aren't just looking at the effect a few years after the program. Here you can really follow people over the full arc of the career.”

Azoulay, Greenblatt, and Heggeness found that ATP participants were twice as likely as the controls to take a research job as their first post, as opposed to one at a hospital or clinic, which was a more typical choice for medical school graduates. They were also more prolific: On average, Yellow Beret grads published 77 papers and had 5,131 citations; for the controls, it was 37 papers and 1,988 citations. Notably, the majority of them acquired what Azoulay and Heggeness call a “translational style” of research, which involved a lot of back and forth between the clinic and the lab, with the ultimate goal of having a real effect on patient care.

Finally, the graduates were also training the next generation of scholars at a much higher rate than members of the control group: Over their careers, they mentored 141 percent more trainees supported on NIH R01 grants, the bread-and-butter research grants the agency awards to independent investigators. Their trainees went on to be more successful at obtaining those grants themselves: They received 167 percent more of them over their careers than the controls' trainees. Because of this, "you can really say that this program really changed the face of American medicine," Azoulay says.

Experts praised the new study for its clever design and detailed analysis. “The careful analysis of factors that predict persistent academic research careers—as well as the extraordinarily long tracking period following the program—adds to our understanding of how to shape the future biomedical workforce,” writes Alison Hall, associate dean for research workforce development at The George Washington University, in an email to WIRED. Hall is the incoming director of the school’s KL2 program, which provides mentored research training to scholars who already have an MD or PhD, similar to what the NIH provided the Yellow Berets.

Mariano Sto. Domingo, a research scientist at the University of Maryland, Baltimore County, says he was “really impressed” by the amount of data the team gathered on the control group of non-participants. In his own work, he has spent two decades evaluating the effectiveness of the Meyerhoff Scholars Program, which trains students from under-represented backgrounds to pursue scientific research careers. He says it can be challenging to track the students who decline to participate in the program to use as controls.

So what exactly made the ATP so influential on careers? Azoulay and Heggeness speculate that it had several important elements. The first was timing: The participants had just finished medical school and an early, intense exposure to research could profoundly change their career trajectory.

"This is what we call, in today's parlance, a pre-doc," Azoulay says. "But that's not where we invest most of our training dollars. Training dollars for the most part go to graduate school and postdocs. And so one hypothesis that you could test is the timing of the program. We ought to experiment much more with funding pre-doc programs.”

Hall agrees that this is a question worth investigating: “We know that mentored, hands-on research experiences are key to shaping the future researcher. There is still much to know, however, about when that experience should be—medical school? High school? Fellowship?”

Another element could have been the large training cohort—between 200 and 300 during the peak years—which is unusual for research programs, but would help participants become well-integrated into a community and practice science as a team sport.

Finally, the researchers think, the training also emphasized early independence, instead of just working on a mentor's projects. That's important because, rather than getting stuck in existing lines of research, the young scientists were able to pursue novel ideas, while still benefiting from the advice of more senior scientists.

Michael Gottesman participated in the program from 1971 to 1974; he now leads a research group at the NIH studying how cancer cells evolve resistance to chemotherapy, and is the agency’s deputy director for intramural research. He was drawn to the program by what Azoulay and Heggeness today call its “translational” approach, or "figuring out what it was that physicians didn't know that they needed to know to take better care of patients," he recalls.

“All of these people congregated at the NIH, and it was a community,” Gottesman continues. He remembers participating in special coursework to get up to speed on research in biochemistry and molecular biology, which were taught by people who were present or future Nobel laureates. “It was a very heady, exciting time. We were young and we were excited about research opportunities. This was the time of the revolution in molecular biology … it looked like the sky was the limit in terms of what we could do to apply basic science knowledge to human disease,” he says.

Gottesman worked with Martin Gellert, who had at that point just discovered the enzyme DNA ligase, the central enzyme for DNA replication and repair. About one year into the program, Gellert took a sabbatical, which gave Gottesman independence to pursue his own research ideas. He also took over teaching a part of Gellert’s course at the NIH on how DNA is transcribed and translated into proteins. “Not only did I have this incredible opportunity to be an independent researcher, but I suddenly had a fairly major teaching responsibility. It was just a wonderful experience for me,” Gottesman says.

After the program, Gottesman returned to Harvard, where he'd done his undergraduate and medical studies, finished his residency, and started as an assistant professor. But soon, he recalls, he "heard the siren call of NIH" and returned to start his own lab at the National Cancer Institute.

After the draft ended, applications to the ATP declined. The program no longer exists, although a similar one—the Medical Research Scholars Program—supports medical, dental, and veterinary students who do research on the NIH campus. The agency is still working to “capture that lightning in a bottle that was this program,” Gottesman says.

Today, a few universities offer similarly intensive programs. For instance, Hall’s three-year program supports about 20 junior investigators in developing independent research careers. It’s funded via NIH KL2 awards, which are given to new clinicians to do research. "In many ways,” she writes, “the KL2 programs provide similar mentored research training as that in the NIH program, at institutions nationwide."

The Meyerhoff Scholars Program, which is in its 33rd year, also has many of the elements of the ATP, although its emphasis is on biomedical research as a whole, rather than translational or clinical research. The program also includes a relatively large, strongly-bonded 50- to 60-person cohort, and intensive exposure to research at the pre-doctoral level. According to Sto. Domingo, its scholars are about five times more likely to earn a scientific PhD than students who were accepted to the program but declined to attend. It is now being used as a model for similar programs at the University of North Carolina at Chapel Hill and Penn State. Sto. Domingo says new programs at UC Berkeley, UC San Diego, and Howard University are also being established based on its model.

Still, medical research careers have changed since the 1960s and 70s. Today, a major obstacle is the debt burden of med school, which is often in the hundreds of thousands of dollars. Debt can incentivize young doctors to choose lucrative specialties so they can repay their loans. As a result, there is a shortage of investigators who are able to combine clinical expertise with research inquiry, writes Hall. In the US, she writes, each year more than 20,000 people graduate with an MD, but only about 600 earn both medical and research doctorates.

Another challenge, writes Hall, is that it’s getting more difficult to manage a dual career doing research and taking care of patients, because it is harder to obtain research funding to support a laboratory, and there are more opportunities to focus on clinical care.

Because the research ecosystem is always changing, Azoulay envisions the Yellow Berets study as a starting point for further research: rigorous studies that would compare training interventions in terms of timing, cohort size, and other factors. "What we'd like people to take away is not that you should copy what the NIH was doing in the early 1970s,” Azoulay says. Rather, this analysis should inspire new experiments. "We want randomized controlled trials to come to the world of scientific training and scientific funding," he continues. "If we have a bee in our bonnet, it's that one."

Disclosure: Viviane Callier has a contract statistician role that supports some data analysis projects at the National Institute of Allergy and Infectious Diseases, which is part of the National Institutes of Health.