When the murder of George Floyd in 2020 ignited calls for racial equity across the U.S., the field of medicine confronted its own thorny questions about race. James Diao, then a medical student at Harvard Medical School, was among the many people who zeroed in on one particular issue: If race is a social construct, why was it a factor in clinical tools used to determine a patient’s risk of disease?
“These big questions were really fundamentally not just scientific problems, but also human and moral problems about how values are incorporated into these seemingly dispassionate tools that we use,” said Diao, whose papers analyzing the impacts of race and its removal from clinical calculators have since played a role in policy decisions that touch millions of patients. “I became really, really obsessed with this idea of what kind of assumptions we’ve baked into these calculators.”
As the health care system continues to grapple with the role of race in many other clinical tools, Diao’s work on the subject has modeled how to balance a quantitative approach to the issue with the perspectives of patients, advocates, and policymakers. Now a resident at Brigham and Women’s Hospital and one of STAT’s 2024 Wunderkinds, Diao’s career has been shaped by his willingness to listen.
“The more I’ve learned, the more I’ve realized there’s more to learn,” he said.
The search for a better way to calculate kidney function
When the pandemic hit, Diao had already been working with Arjun Manrai, an assistant professor of biomedical informatics at Harvard Medical School, since he was a college sophomore studying statistics and biochemistry. Manrai, who recruited Diao once he began his medical training at Harvard, investigates how clinical algorithms work — and fail — when they’re applied in different populations. As the two were cooped up at home, they started delving into the different ways that hospitals at the time were adjusting their calculators for kidney function, known as eGFR, to remove race.
They Zoomed. A lot. “I think there was a period where we Zoomed with each other every single day,” said Manrai. Diao looped his partner Gloria Wu, a public health researcher, into the work. (This August, Manrai officiated their wedding, where the pair showed off their ballroom dancing skills.)
Diao didn’t just stay in his bubble. On Twitter, as fierce debates unfolded over the harm of both keeping and expelling race from the kidney calculator, “I was a fly on the wall,” said Diao. “I’m grateful I didn’t say anything then. I had so much to learn.”
To catch up, he read. “Fatal Invention,” Dorothy Roberts’ book about the fallacy of race as a biological category, was at the top of his Covid reading list. He talked with the medical students advocating for changes to the race-based kidney calculator in their hospitals, and joined meetings of the grassroots Institute for Healing and Justice in Medicine with other medical students who had started questioning the clinical equations they were being taught.
As he continued his research, Diao had to carefully balance empirical and moral arguments for considering race in clinical decisions. “He had a really difficult job,” said Rohan Khazanchi, then a medical student studying health services and health equity. Diao was young, and a relative outsider, working with teams of physicians and researchers that have been studying — and using — clinical algorithms for years.
“You have to keep it focused on the data, so that different people can use that data to inform the policy decisions without feeling like you have your thumb on the scale,” said Diao. “But then on the other hand, if you try to lean too far into that, people think you don’t care about the underlying issues.”
At the end 2020, Diao was the first author on a study in the Journal of the American Medical Association that quantified the impact on Black patients if medicine were to remove race outright from the existing eGFR calculator, without changing it in any other way. Removing race would increase earlier access to kidney care, including transplants, he wrote with Wu, Manrai, and others. But it would also prevent some patients’ access to chemotherapy or medications that have doses based on kidney function — a finding that some interpreted as advocating against removing race from the calculator.
“It was really stressful to have to try to defend ourselves,” said Diao. But as the National Kidney Foundation and American Society of Nephrology convened a task force to reassess the role of race in eGFR, Diao was able to present his case, testifying about the research along with other medical students who had advocated for the removal of race from their hospitals’ equations.
“He cited some of the numbers,” recalled Khazanchi, “but he also talked about the challenges of talking to an individual patient and saying, ‘By the way, I’m using your race to determine the stage of your kidney disease.’”
The next year, Diao and others published a perspective on the search for a better kidney function equation in the New England Journal of Medicine that went beyond the simple removal of race from the existing eGFR calculator.
They aimed to move beyond the “false dichotomy” pitting those approaches against each other, said Manrai, who co-authored the paper. “There’s many other race-free equations, and many other ways of changing this equation to remove the stratification by race that are more accurate and that have different strengths and weaknesses.” Months later, the task force issued its recommendation supporting two of the race-free approaches highlighted in that paper, citing it directly. In practice, that decision has led thousands of Black patients to have their wait time on kidney transplant lists adjusted.
Bringing patient perspectives to clinical algorithms
Race meant one thing where Diao grew up, in the suburbs of diverse Houston, where his parents worked as oil company engineers. But when he would visit his dad’s hometown in China and showed his school yearbook around, he remembered, many people assumed every dark-skinned student was of African descent.
Seeking to expand his global perspective on race, in 2022 he went to the other Cambridge across the pond, to study health policy and expand his global perspective on race. “He really listens to a lot of people who have different perspectives, and actively seeks them out,” said Manrai. “He’s not just coming at it from one community, one angle. He really is listening to lots of folks.”
Meanwhile, Diao was continuing his work on another race-based clinical algorithm. The American Thoracic Society, spurred on by 2020’s racial reckoning, had begun to reconsider its approach to lung function testing, which had long assumed Black patients had lower baseline lung volumes than white patients.
Diao’s approach was similar to his work on the kidney function calculator, said Khazanchi, who worked on the project: “For all patients, what are the positive and negative implications of shifting from a race-based pulmonary function testing algorithm to a non-race based algorithm?”
By the time Diao published yet another first-authored paper in the New England Journal of Medicine this year, the ATS had a verdict: It would move away from race-based lung function testing. But the research provided critical context as health systems figured out how to execute that recommendation.
Shifting away from race-based tools would likely make Black patients more eligible for worker’s compensation. With more accurate lung function estimates, they could get access to treatments and therapeutics. But better diagnosis could lead to more invasive, potentially risky interventions, or even take certain surgeries off the table. After the study was published, the Veterans Affairs department launched an investigation of the change’s impact on disability payments, anticipating a smaller effect than the study predicted.
On the strength of his work, Diao recently became the 22nd student in the history of Harvard Medical School to graduate summa cum laude. With Khazanchi, he’s diving right into residency in Boston, where he’s been drawn to cardiology.
“Something clicked” about the specialty, said Diao. He loves the high stakes of helping a patient with a heart attack. But cardiology is also an especially data-driven specialty — a place where he could see his computer science and statistics chops making an impact. “The field is moving really, really quickly and using all this data that it’s collecting for the benefit of the patient,” said Diao. “This is something I could really contribute to and be part of.”
For now, he’s focused on helping patients avoid cardiovascular disease. In one of his latest publications in JAMA, Diao projected the impacts on statin eligibility with the likely adoption of another new race-free tool, this one used to predict the likelihood of strokes and heart attacks. That new calculator, PREVENT, was built by the American Heart Association as a wholesale revamp of the previous tool, aiming to make it more accurate by incorporating patients’ BMI and kidney function.
It also, in a first for a U.S. tool of its scale, attempts to refine its predictions by incorporating a patient’s social determinants of health, using ZIP code to estimate factors like income, education, and housing status. Diao played around with the new tool, trying different ZIP codes around his hospital. “Boston has some of the largest disparities in life expectancy between its neighborhoods of any city in the U.S.,” he said. Plugging in different ZIP codes in the area, “you end up with these really dramatic differences in the prediction.”
As the AHA contemplates whether and how to adopt PREVENT into its clinical guidelines, the question is whether those ZIP code-based predictions are more accurate — “and if so, how do patients feel about that?” said Diao. If patients are uncomfortable with their race being used to determine their risk of disease, how would they feel about their estimated income factoring in?
In one study he’s working on now, Diao and his colleagues are “just asking people,” he said: “Asking people what they feel comfortable with being used in their care, and when they feel comfortable with it being used in their care.”
He’s still listening.
