'The potential is unbelievable': First chief data scientist of Duke Health talks intersection of AI and healthcare

Courtesy of Scholars@Duke.
Courtesy of Scholars@Duke.

As artificial intelligence takes the world by storm, Duke sees it as an opportunity to transform healthcare.

Michael Pencina, vice dean for data science, was recently named Duke Health’s chief data scientist, a leadership role created in the midst of rapid developments and intense discussions around AI technology. 

“I want Duke Health to be synonymous with trustworthy health AI,” Pencina said. “When people think about developing or evaluating algorithms, I want them to look at what we develop, come to us for guidance.” 

Merging data and health

Pencina, a professor of bioinformatics and biostatistics at the School of Medicine, said that his role was to facilitate a “partnership between the research side — the University’s School of Medicine — and the health delivery side.”

Duke’s creation of the new position was motivated by the potential benefits to medicine that the University’s research and education in data science could bring, according to Pencina. 

Duke Health also recently formed a major partnership with Microsoft to “redefine the health care landscape” through collaborative work on generative artificial intelligence and cloud technology.

According to Pencina, Microsoft saw Duke Health as a partner that could work to evaluate the development of trustworthy AI. With established governance for the use of AI in healthcare, Duke was also viewed as a national leader of AI applications. 

Road to chief data scientist 

The role of chief data scientist is new for Pecina, whose career spans decades as well as multiple campuses and organizations.

The School of Medicine regards Pencina as “an internationally recognized authority in risk prediction model development and evaluation,” especially in the field of cardiovascular disease risk prediction. 

In the past, Pencina has contributed to the Framington Heart Study, which is recognized by the National Heart, Blood and Lung Institute for “transformative discoveries related to the treatment of heart disease.” Pencina praised the Institute as having “pioneered risk prediction algorithms.”

With over 400 publications in peer-reviewed journals that have been cited over 111,000 times, Pencina brings experience and expertise to the ever-evolving fields of AI and medicine. 

Ethical and equitable AI

Duke AI Health, an initiative that Pencina directs, has ethical and equitable data science as its core, he said.

In the face of advancements as well as worrying possibilities in the field of AI technology, these values are a cornerstone of Pencina’s vision. 

“In 2019, a colleague of mine published this really important paper showing racial bias in an algorithm that was supposed to determine which patients need preventive health,” Pencina said. “Then the pandemic hit and what came to the forefront were major racial inequities in healthcare.” 

These instances of inequity motivated Pencina to prevent misuse of AI and ensure ethics in its implementation within health care. 

Pencina has since committed himself to actions such as hiring a health ethicist in February, as well as forming the Algorithm Based Clinical Decision Support Oversight to consistently evaluate whether newly introduced technology is equitable and safe. 

The future of healthcare 

With his experience in predictive AI, Pencina is both scared and excited about generative AI. He published an op-ed in The Charlotte Observer in August arguing for better guidelines to use AI correctly and ensure that AI is used to serve humans. 

“Artificial intelligence offers possibilities that test the limits of our imagination,” he wrote in the op-ed. “For many, these may exceed the boundaries of what is comfortable or even desirable, as the current call for a pause on AI development demonstrates.”

Pencina also believes AI can play a role in reforming healthcare systems through avenues like the documentation of meetings, authorization for insurance reimbursement, surgery arrangement and diagnostic imaging.

“The potential is unbelievable,” he said.


Winston Qian | Health/Science Editor

Winston Qian is a Pratt sophomore and health/science editor for the news department.    

Discussion

Share and discuss “'The potential is unbelievable': First chief data scientist of Duke Health talks intersection of AI and healthcare” on social media.