In May, the National Academy of Medicine announced that Ziad Obermeyer, acting associate professor of health policy and management at Berkeley Public Health, has been selected among this year’s 10 Emerging Leaders in Health and Medicine Scholars. Obermeyer has been recognized for his research and training at the intersection of healthcare and machine learning tools.
This honor is bestowed to early to mid-career professionals who cover a range of fields related to health and medicine. This year’s list of scholars include experts in areas ranging from engineering and economics to health policy and clinical care. Part of a larger initiative by the National Academy of Medicine, this award provides a space for emerging health researchers and practitioners to collaborate with others in the academy to advance science, policy, and public health.
“As the world faces the devastating effects of the COVID-19 pandemic, we are reminded of the importance of involving emerging leaders, who are poised to shape the future of health and medicine, in cross-disciplinary activities to tackle pressing challenges such as these,” said NAM President Victor J. Dzau in a statement. “I am delighted to welcome these extraordinary individuals who represent the next generation of leading scientists, healthcare providers, public health professionals, and policymakers into the National Academy of Medicine’s Emerging Leaders in Health and Medicine program.”
Obermeyer joined the UC Berkeley faculty in 2018. An emergency physician by training, he had already started research at Harvard at the intersection of algorithms and emergency medicine, focusing on how machine learning can improve diagnoses and limit physician error. By computing advanced data sets on patients’ histories, environments, and levels of risk, machine learning can mitigate the “trial and error” of emergency room testing and help doctors make predictions about patients, even for patients without clear risk factors.“Big data will transform medicine,” Obermeyer wrote in a commentary for the New England Journal of Medicine.
At Berkeley Public Health, Obermeyer formed a joint lab with Sendhil Mullainathan, the Roman Family University professor at the University of Chicago Booth School of Business. Together, Obermeyer and Mullainathan analyze how to improve the use of algorithms in the medical system.
Part of their research involves careful implementation of machine learning in healthcare, particularly when algorithms can replicate the same racial, socioeconomic, or gender-based biases that already plague the healthcare system.
In a seminal study published last fall in Science, a research team led by Obermeyer and Mullainathan found that a type of software program that determines who gets access to healthcare management programs favored healthier white patients ahead of black patients who are less healthy.
“Instead of being trained to find the sickest, in a physiological sense, [these algorithms] ended up being trained to find the sickest in the sense of those whom we spend the most money on,” Mullainathan told Berkeley News. “And there are systemic racial differences in health care in who we spend money on.”
The study suggested that fixing this bias could double the amount of black patients who are admitted into these programs. According to Obermeyer, the racial bias could be addressed by training the algorithm to determine risk based on other variables, such as the number of chronic conditions that need treatment of avoidable cost.
“Algorithms can do terrible things, or algorithms can do wonderful things. Which one of those things they do is basically up to us,” Obermeyer said. “We make so many choices when we train an algorithm that feel technical and small. But these choices make the difference between an algorithm that’s good or bad, biased or unbiased.”
The Science paper attracted the attention of many media and news outlets, including the Wall Street Journal, The Washington Post, Los Angeles Times, Wired, and others. In a November blog post in Health Affairs, the paper’s authors discussed the path forward for machine learning in healthcare, including the launch of a new research initiative to continue to address racial bias.
As one of the emerging leaders at the National Academy of Medicine, Obermeyer will bring this critical research of machine learning tools to the NAM community of experts. Alongside the other emerging leaders, Obermeyer will share insights from his research at the NAM Emerging Leaders Forum in Washington, D.C. in October.