An increasing number of healthcare systems are hiring executives, typically physicians, to lead AI-related activities – often AI executives, fully at the executive level. Convoluted technology with enormous potential is exploding in healthcare and many industries.
Kaiser Permanente, which has 40 hospitals and 600 medical offices, hired Dr. Daniel Yang at the end of 2023 as vice president of artificial intelligence and modern technologies.
Yang, who is not chief, reports to Dr. Andrew Bindman, the chief medical officer. In this role, he provides quality oversight of all AI applications across the organization, including those used in clinical operations, research, education and related administrative functions.
Which recently published a framework of seven principles that guide an organization, helping to ensure the security and reliability of the artificial intelligence tools it implements. He also called on policymakers to assist ensure the responsible apply of artificial intelligence.
sat down with Yang to discuss his role at Kaiser Permanente, the top healthcare issues that could be impacted by AI, and the implementation of a modern AI-enabled clinical documentation tool for physicians and other healthcare system caregivers.
Q. Kaiser Permanente management created the position of Vice President of Artificial Intelligence and Emerging Technologies in overdue 2023 when they hired you. What were they looking for in a director and what did you propose?
AND. I joined Kaiser Permanente from the Gordon and Betty Moore Foundation, where I supported AI governance efforts in healthcare systems to ensure a thoughtful and judicious approach to procuring and implementing AI tools.
While working at the Moore Foundation, I began conversations with Kaiser Permanente’s Chief Medical Officer, Dr. Andrew Bindman, about creating a responsible artificial intelligence program for Kaiser Permanente. It was clear that Kaiser Permanente wanted someone who woke up and went to sleep thinking how artificial intelligence and other emerging technologies can assist address the challenges facing the US health care system.
These challenges may include access to quality care, affordability, and patient satisfaction.
What ultimately attracted me to Kaiser Permanente was the opportunity to further shape the intersection of patient safety, healthcare quality and modern technologies. I saw this role as a compelling opportunity to build a culture and program of responsible AI at Kaiser Permanente and, by working for one of the largest integrated health systems, as an opportunity to influence the responsible adoption of AI in health across the United States.
It also allowed me to return to my roots, i.e. internal diseases. I am still a practicing physician and still see patients requiring urgent care at Kaiser Permanente and the Department of Veterans Affairs. This hands-on experience helps me understand and empathize with the everyday challenges physicians face. I bring this empathy to my role every day, thinking about how to implement artificial intelligence and modern technologies in a way that improves patient care and minimizes risk.
Q. How would you describe your job description to colleagues interested in your role and to executives considering filling the position?
AND. When I look at other AI leaders or similar leadership roles being created in different healthcare systems, I find that I see a common set of denominators in these positions.
First, these leaders are often practicing clinicians.
Second, many of these leaders are not actually technologists. They have come to appreciate the importance of technology, whether their background is in quality, safety or research.
Third, these roles require some operational experience. In most cases, these leaders are faced with modern processes and programs that did not previously exist. They must imagine and chart modern agendas without pre-existing playbooks and find the resources to do so, which can be especially challenging given macroeconomic pressures related to rising health care costs.
And fourth, forceful communication skills are crucial ensure the implementation of artificial intelligence technologies in a scientifically sound and practical way. I consider myself an AI translator, bridging the gap between technical teams, clinicians and society. Although I do not have a PhD in artificial intelligence or computer science, my forte is adapting AI solutions to the practical needs of our clinicians and members.
Q. What do you think are the most crucial issues related to artificial intelligence in healthcare today?
AND. I think the real question is, “What are the most important issues in healthcare today?” I feel like a lot of leaders get this wrong. You have no problem with AI. Artificial intelligence is just one of many tools or technologies we can apply to assist solve real problems facing the American health care system today.
Some of these crucial, overarching challenges in health care include increasing delays in access to patients, rising costs of providing health care services, and provider burnout. These challenges are a consequence of the discrepancy between the increasing complexity of care delivery and the availability of trained physicians to provide care.
Artificial intelligence can assist make two fundamental changes needed to solve the supply-demand mismatch.
First, AI can assist patients better self-manage their care.
Second, AI can assist the health care system evolve from a 1:1 (1 doctor: 1 patient) care model to a 1: many care model, while maintaining or even improving the quality of care and patient experience. Using generative AI for personalized learning in the education sector is an intriguing model for us to learn from.
Finally, the healthcare industry must address concerns about the safety and quality of AI in healthcare. This requires a stalwart, accountable AI program that puts patient safety and quality of care above all else.
It will also enhance organizations’ ability to address ethical issues and potential bias in AI implementations. The most crucial thing is to ensure that AI tools do not perpetuate existing inequalities in healthcare. To achieve these goals, continuous monitoring and evaluation is necessary.
Q: What AI project did you oversee during your first year at Kaiser? What was the process like and what did you and your team achieve? What were the results?
AND. The Covid-19 pandemic has caused burnout among American physicians highest level ever. We wanted to better support our physicians and clinicians by reducing the administrative work associated with their daily lives, especially the time spent documenting clinical notes during patient visits. We have seen that rapid advances in the field of generative AI can assist solve this problem.
I helped lead the implementation of a modern AI-powered clinical documentation tool for our physicians and other clinicians across 40 Kaiser Permanente hospitals and over 600 physician offices. We think this one was the biggest implementation of ambient listening technology in the United States.
The tool, which requires patient consent, helps doctors and other clinicians safely take clinical notes during patient visits. Most importantly, it frees our doctors and clinicians from their keyboards so they can focus their attention on patients, not computer screens.
At Kaiser Permanente, we believe that artificial intelligence should never replace the judgment and knowledge of our physicians and clinicians. For this to succeed, we must evaluate each AI tool before implementing it to ensure we know how to apply it safely and effectively.
For example, when we first implemented this tool, I co-led our quality assurance testing to assess how well it performed across clinical specialties and in a messy real-world environment. These insights have helped us apply this tool responsibly, including educating physicians and clinicians on how to apply it effectively while minimizing risk.
As of September 2024, physicians and clinicians have used this tool to record over 2 million patient interactions. Doctors and clinicians say the tool allows for more meaningful interactions with patients. Patients also reported that the tool provided greater transparency during their visit.
For example, many doctors now dictate their results aloud during a physical examination, whereas previously they might have said nothing. To me, this is a great example of how modern technologies can assist our care teams provide the highest quality care to our members and patients.