STAT Wunderkind James Diao on racial and clinical algorithms

When the 2020 murder of George Floyd sparked calls for racial equality across the United States, medicine faced its own thorny questions about race. James Diao, then a medical student at Harvard Medical School, was one of many people who focused on one particular issue: If race is a social construct, why is it included in the clinical tools used to determine a patient’s disease risk?

“These important questions really weren’t just about scientific problems, but also about the human and moral problems of incorporating values ​​into the seemingly impartial tools we use,” said Diao, whose papers examining the impact of race and its removal from clinical calculators have played a role in political decisions that affect millions of patients. “I’m really obsessed with the assumptions we put into these calculators.”

As the healthcare system continues to struggle the role of race in many other clinical toolsIn Diao’s work on this topic, she presented a model for balancing a quantitative approach to the problem with the perspectives of patients, advocates and policymakers. Now inhabitant at Brigham and Women’s Hospital and one of Wonder Kids 2024 according to STATDiao’s career was shaped by his willingness to listen.

“The more I learn, the more I realize there is still a lot to learn,” he said.

Finding a better way to calculate kidney function

When the pandemic hit, Diao was already working with Arjun Manrai, an assistant professor of biomedical informatics at Harvard Medical School, because he was a sophomore studying statistics and biochemistry. Manrai, who recruited Diao when he began medical training at Harvard, studies how clinical algorithms work – and how they fail – when applied to different populations. While they were both at home, they began to wonder what hospitals were like back then by adapting their calculators to kidney functionknown as eGFR, to remove race.

They enlarged it. Very. “I think there was a period where we were on Zoom every day,” Manrai said. Diao brought in his partner Gloria Wu, a public health researcher. (In August this year, Manrai officiated a wedding during which the couple showed off their ballroom dancing skills.)

Diao didn’t just stay in his bubble. As fierce debates raged on Twitter about the harms of maintenance and the exclusion of race from the kidney calculator, “I was a fly on the wall,” Diao said. “I’m grateful I didn’t say anything then. I still had a lot to learn.”

To catch up, he read. Dorothy Roberts’ book Fatal Invention on the misconception of race as a biological category was at the top of his Covid reading list. He spoke to medical students advocating for changes to the race-based kidney calculator in their hospitals and joined the grassroots Institute for Healing and Justice in Medicine meetings with other medical students who began to question the clinical equations they were taught.

As Diao continued his research, he had to carefully balance the 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 a young outsider who worked with teams of physicians and researchers who have been studying and using clinical algorithms for years.

“There needs to be a focus on data so that different people can use it to make policy decisions without feeling like they’re keeping their thumb on a scale,” Diao said. “But on the other hand, if you try to lean into it too much, people will think you don’t care about the basic issues.”

At the end of 2020, Diao was the first author on study in the Journal of the American Medical Association which quantified the impact on black patients if medicine were to immediately remove race from the existing eGFR calculator without otherwise changing it. 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 from accessing chemotherapy or drugs whose doses depend on kidney function – which some interpreted as advocating for not removing race from the calculator.

“It was really stressful trying to defend myself,” Diao said. However, when the National Kidney Foundation and the American Society of Nephrology convened a task force to re-evaluate the role of race in eGFR, Diao was able to make his case, testifying about the research alongside other medical students who advocated for removing race from the equation of their hospitals .

“He quoted some numbers,” Khazanchi recalled, “but he also talked about the challenges of talking to an individual patient and saying, ‘By the way, I take your race into account to determine the severity of your kidney disease.’ ”

The following year, Diao and others published: perspective in search of a better kidney function equation in the New England Journal of Medicine that would go beyond simply removing race from the existing eGFR calculator.

They wanted to move beyond the “false dichotomy” that pits these approaches against each other, said Manrai, co-author of the paper. “There are many other race-free equations and many other ways to change the equation to remove stratification by race that are more accurate and have different strengths and weaknesses.” A few months later, the task force issued his recommendation supports the two race-free approaches highlighted in this article by citing them directly. In practice, this decision led to thousands of black patients opting for this procedure waiting time on kidney transplant lists adapted.

Introducing the patient’s perspective into clinical algorithms

Race meant one thing where Diao grew up, in the suburbs of diverse Houston, where his parents worked as engineers for an oil company. But when he visited his dad’s hometown in China and gave his high school yearbook a tour, many people assumed that every dark-skinned student was of African descent.

Wanting to expand his global view of race, he traveled to the other Cambridge across the pond in 2022 to study health policy and expand his global view of race. “He really listens to a lot of people who have different points of view and actively seeks them out,” Manrai said. “He doesn’t just approach it from one community and one point of view. He really listens to a lot of people.”

Meanwhile, Diao continued to work on another race-based clinical algorithm. Inspired by the 2020 racial assessment, the American Thoracic Society has begun to reconsider its approach to pulmonary function testing, which has long assumed that Black patients have 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: “What are the positive and negative consequences for all patients of switching from a race-based lung function testing algorithm to a race-neutral algorithm?” based on an algorithm?”

Until Diao published another first-authored article in the journal New England Journal of Medicine this year the ATS issued a verdict: yes moving away from race-based lung function tests. However, the study provided critical context as healthcare systems figured out how to implement this recommendation.

Moving away from race-based tools would likely make black patients more eligible for workers’ compensation. With more accurate estimates of lung function, they could gain access to treatment and therapies. However, better diagnosis may lead to more invasive, potentially risky interventions, or even the abandonment of some surgeries. After the study was published, the Department of Veterans Affairs initiated an investigation impact of the change on disability benefits, expecting a smaller impact than the study predicted.

Through his work, Diao recently became the 22nd student in the history of Harvard Medical School to graduate with honors. Together with Khazanchi, she begins an internship in Boston, where she became interested in cardiology.

“Something clicked” in this specialty, Daio said. He loves the high stakes of helping a heart attack patient. But cardiology is also a particularly data-driven specialty – a place where he could see his knowledge of computer science and statistics make an impact. “The field is evolving really quickly and using all the data that is collected to benefit the patient,” Daio said. “It’s something I could really contribute to and be a part of.”

For now, he’s focused on helping patients avoid cardiovascular disease. In one of his latest publications in CAVITYDiao predicted an impact on statin eligibility with the likely adoption of another new race-free tool for predicting the likelihood of strokes and heart attacks. The new calculator, PREVENT, was created by the American Heart Association as part of a complete overhaul of the previous tool, with the goal of increasing its accuracy by taking into account patients’ BMI and kidney function.

It’s also for the first time on an American tool of this scaletries to refine its predictions by taking into account a patient’s social determinants of health, using zip code to estimate factors such as income, education and housing status. Diao played around with the new tool, trying out different zip codes in his hospital. “Boston has some of the widest differences in life expectancy among its neighborhoods of any city in the U.S.,” he said. By entering different ZIP codes in an area, “you get really dramatic differences in predictions.”

As the AHA considers whether and how to incorporate PREVENT into its clinical guidelines, the question arises whether ZIP code-based predictions are more accurate – “and if so, what do patients think about it?” Diao said. If patients are uncomfortable with their race being used to determine disease risk, how would they feel about their estimated income?

In one study he is currently working on, Diao and his colleagues “simply ask people,” he said: “You ask people how comfortable they feel when it is used in their care and when they feel comfortable when it is used in their care care”.

Still listening.