Dr. Andrew Trister

While nobody could have predicted then, Dr. Andrew Trister’s journey becoming one of the world’s leading experts in AI in healthcare began when he was just a toddler. He was introduced to programming by his great-uncle, an engineer, and had access to computers in his formative years. “My great-uncle brought computers into the house when I was around 3. It was kind of brutal — all I wanted to do was play Frogger, but he insisted I learned how to program it,” Trister says.

Coming from a small family, he spent summers in Israel with his three cousins, one who passed away from a posterior fossa brain tumor when Trister was 13 years old.

Trister says, “I saw the aftermath of surgery and the last stages of his life; it really changed my outlook on where I wanted to contribute. I wanted to learn more about medicine and see what ways I might be able to make an impact.”

He chose to study computer science and neuroscience as an undergraduate and went on to earn his PhD and MD at the University of Pennsylvania, believing that understanding the nervous system’s biological and computational aspects would be crucial in advancing medical treatments to help people like his cousin.

“When I was looking for a school for residency, I was focused on the proximity of an engineering school,” explains Trister. “My undergraduate thesis advisor had a lot of strong collaborators at the University of Washington. Dr. Mark Grudine had a similar path at Penn and came to the University of Washington for radiation oncology and had built a tremendous research program. “

Trister wanted to continue his multidisciplinary approach to medicine and saw the University of Washington School of Medicine’s Radiation Oncology residency program as an ideal fit, allowing for easy collaboration between disciplines and providing strong mentorship.

“So many amazing people like Grudine and Dr. George Laramore, who cared deeply about mentorship, who cared about the work we were doing, combining clinical care and research, were at the UW,” Trister says.

His time here proved to be formative, enabling him to apply his combined knowledge of computer science and medicine to real-world clinical challenges. His experience with patients facing geographic barriers to care, coupled with the institution’s strengths in medicine and engineering, helped shape his ideas for AI-driven solutions in healthcare.

“My idea was to learn how to deliver care, as one should in residency, and there was a great opportunity to learn from master clinicians and then to focus on the ability to translate the research,” says Trister.

The Growth of AI

The development of Trister’s career mirrored — and often intersected with — the rapid growth of AI in healthcare. He witnessed the field’s evolution from rudimentary systems to the rise of deep learning in the mid-2000s. He experienced firsthand the excitement surrounding AI’s potential to revolutionize healthcare, followed by the realization of the significant challenges ahead. Issues such as data biases and the need for robust, generalizable models shaped his perspective on the future of AI in healthcare.

“The idea was to harness this opportunity that we suddenly had. Bioinformatics was starting to determine something about precision. We were building AI models and starting to understand the differences that you might be able to discover with a computer,” says Trister.

During his residency, Trister focused on using AI to personalize cancer treatment, especially for glioblastoma. He saw AI’s potential in capturing the complex interplay between a patient’s anatomy, molecular features and treatment response, a concept he called “radiogenomics.”

“They used to say that as a resident, you generate this third eye in radiation oncology, where you have to see where things are heading,” says Trister. “You have to know what’s happening with the cancer. I said we should capture that; we should be able to build models that do exactly that.”

During his residency at the University of Washington, Trister took on a postdoctoral fellowship in systems biology at Sage Bionetworks and began exploring the practical applications of AI in healthcare, using wearable technology to collect data for clinical studies and overcome geographical barriers to patient participation.

Since 2015, he has held various roles at Apple and the Bill & Melinda Gates Foundation, and in 2023 he joined Verily, an Alphabet, Inc., company (also known as a Google company). As its chief medical and scientific officer, he leads population health initiatives and advances precision health strategies.

A Proactive, Personalized and Equitable Healthcare System

“AI has the potential to revolutionize medicine by augmenting the capabilities of physicians and creating a more proactive, personalized healthcare system,” says Trister.

By helping people manage their health and access care remotely, Trister believes that AI-powered tools will become increasingly commonplace and seamlessly integrated into our daily lives and healthcare experiences. He envisions a future where AI has the potential to transform healthcare from a reactive model to proactive health management, where AI-powered tools enable continuous monitoring of people’s health, allowing for early detection of diseases.

Additionally, AI-driven telemedicine platforms could expand access to care, particularly for patients in remote or underserved regions.

AI can also have a profound impact on personalized treatment, Trister says. Through its ability to analyze vast amounts of patient data, AI algorithms can tailor treatment plans to an individual’s specific characteristics and medical history. He sees AI systems — trained on large datasets of medical images and patient records — assisting healthcare professionals in making more accurate diagnoses, ultimately improving patient outcomes.

By automating routine tasks, healthcare providers can focus more on complex patient needs. Trister predicts a shift in patient attitudes as individuals increasingly seek out healthcare providers who use AI to enhance their diagnostic and treatment capabilities.

He says, “Today, if a doctor were to tell a patient, ‘I’m going to use AI to do this diagnosis and to do this whole exam,’ I think most patients would ask, ‘what are you talking about? I came to see you.’ In the not-too-distant future, this dynamic will flip around and the patients are going to seek out the doctors who use these tools. This is the cutting edge — this is the way forward.”

Potential Pitfalls and Ethical Considerations

While he’s optimistic about the future of AI in healthcare and its potential to revolutionize the delivery of care, Trister cautions that we must also be mindful of the potential pitfalls.

“We need to be cautious and address ethical and practical challenges, and realize the importance of human oversight, collaboration and thoughtful consideration of data bias and representativeness to ensure that AI is used responsibly and equitably in healthcare,” he says.

If implementation of AI is not handled carefully, there is a risk that AI could exacerbate existing healthcare disparities.

“The current healthcare system in the United States, and even globally, is broken and unsustainable. Introducing AI into a flawed system without addressing the underlying issues could lead to unequal access to AI-powered healthcare, further disadvantaging those already facing barriers,” Trister states.

For example, concierge doctors with more resources might readily adopt AI tools, while those reaching underserved communities might not, creating a scenario where individuals with greater means benefit disproportionately from AI advancements in healthcare, while those in underserved communities see little to no improvement.

Along with access, another major roadblock is underrepresentation of certain demographic groups in training datasets, which can lead to inaccurate outcomes.

Trister says, “AI models that are trained on biased or non-representative data can perpetuate and exacerbate existing healthcare disparities. Addressing this requires ensuring data sets are diverse and inclusive and developing methods to measure and mitigate bias.”

Privacy and data security are other concerns. In addition to bias and inequities, the use of AI in healthcare necessitates the collection and analysis of sensitive patient data. Ensuring data protection measures and transparent data governance frameworks is essential.

Navigating the Path Forward

By advocating for AI systems that are equitable, accessible and genuinely beneficial to all, Trister emphasizes the need to work with patients and healthcare providers in the development and implementation of AI tools, making sure that real-world needs and concerns are part of their development.

“Ultimately, the success of AI in healthcare hinges on its ability to augment, not replace, the crucial role of human empathy, judgment and connection in the delivery of care,” A collaborative effort involving research institutions, government, the private sector and philanthropic organizations is needed to effectively navigate the path forward for AI in healthcare.

Trister sees the potential for public-private partnerships and non-competitive collaborations to address common challenges and accelerate progress, including implementing regulations.

“Right now, nobody in government wants to be known as the wet blanket on what has captured everyone’s imagination. However, establishing appropriate regulations and ethical guidelines for AI in healthcare is essential to ensure patient safety, privacy and equitable access,” says Trister.

A Hopeful Future

“The idea of the proactive health system, where we can keep people healthier longer, is an exciting one” says Trister.

By integrating AI into healthcare practices, providers will be able to leverage technology to improve diagnostic accuracy and deliver more personalized patient care, empowering providers to deliver the best care possible.

Trister believes that the future of healthcare lies in the collaborative effort between human skill and AI, ensuring better outcomes and a more efficient, compassionate approach to patients’ well-being.

“It’s really important for us to recognize that the next generation is going to come and really lean in on AI and instill the idea of humanity, focusing in on how we build a system that supports the multidisciplinary approaches that medicine has always done.”

Written by Patsy Cadwell