Will radiologists be out of jobs in the next few years? Can artificial intelligence (AI) systems and devices replace doctors as diagnosticians? What are the ethical considerations? These were some of the key questions debated and discussed at the ‘AI in Healthcare’ conference organized by Insight Exchange Network at the Boston Harvard Club on January 18th and 19th.
Two distinct groups were in attendance—clinicians and computer scientists with AI serving as the bridge between them. Healthcare administrators representing medical centers and executives from health technology companies were also present on both days.
On Day 1 a fast paced start up showcase had entrepreneurs presenting AI powered solutions for a range of healthcare issue—from disease detection through eye scans (Optina Diagnostics) to a virtual nurse assistant (Angelcare) and more (see list with links below).
Keynote speaker Anthony Chang brought an intersectional perspective of a pediatric cardiologist care as well as a data scientist well versed in the intricacies of machine learning. The information packed presentation provided an overview of the impact and developments in AI and his work at the Children’s Hospital of Orange County where he is Chief Intelligence/Innovation Director and Medical Director, Heart Failure Program.
Chang delineated the two systems of thinking, methodogical thinking typical of AI systems and the emotional/intuitive thinking characteristic of human thinking. Ultimately human to human collaboration between clinicians and data scientists is critical.
The first keynote panel, moderated by Carlos Rodarte of Volar Health, included panelists Tatyana Kanzaveli, CEO and founder of Open Health Network, Sanjay Joshi, Chief of Technology Healthcare and Life Sciences at H20.AI, Marty Kohn, MD, consultant in clinical informatics and health policy and Kamaljit Behera, an analyst from Frost & Sullivan.
The panel discussed the role of AI in healthcare as AI becomes increasingly “intelligent” and capable of taking human like decisions. Specifically they discussed:
• How AI, serving as a clinical decision support system makes possible better decision making by clinicians for individual patients
• How AI makes it possible to deliver personalized precision medicine
These are exciting and promising developments that will shift the practice of medicine as well as medical education. These developments also raise questions regarding the new social contract emerging between the patient, researcher and clinician.
On Day 1 presenters highlighted the ethical challenges and opportunities for AI in Healthcare (Dr John Mattison, Assistant Medical Director and Chief Health Information Officer, Kaiser Permanente,), reviewed opportunities to bring AI into the workflows of physicians and patients at home (Dr. Kenneth Mandl, Director, Computer Health Informatics Program, Boston Children’s Hospital), and discussed the importance of data and “datathons” in bridging the Health Divide Data (Leo Anthony Celi, Principal Research Scientist, MIT).
The underlying question in all deliberations and discussions was the role of AI. Would it augment or replace existing systems and human services?
On Day 2 Dr. Michalski, Executive Director, MGH & BWH Center for Clinical Data Science presented the Center’s current projects that showcased the transformative power of emerging AI based technologies. He also discussed the challenges of implementation.
With possibilities come challenges. This was discussed by the keynote panel earlier in the day, moderated by Varesh Prasad, a Ph.D. candidate in Medical Engineering and Medical Physics at the Harvard-MIT Health Sciences & Technology program. The panel included Dr. Rowland Stewart, VP of Clinical Integrated Solutions at Mercy Hospital, Dr. Sven Van Poucke an Anesthesiologist and Emergency physician, Zol Genk, and Finale Doshi-Velez, an Assistant Professor of Computer Science at Harvard University. Here again the panelists emphasized the necessity for collaboration between data scientists designing the algorithms and the physicians providing the care. While there are use cases of successful implementation of AI powered systems, the panelists urged caution given the inevitable presence of bias in the data and the time lag in the availability of data. Computers, big data, algorithms are all tools. Powerful tools. They can identify patterns that are not evident otherwise. AI working with clinicians can reduce errors overall.
The one certainty in all scenarios discussed is the presence of humans and an uniquely human quality—Intuition. Human beings’ capacity to be intuitive and to provide context, ensure that human doctors cannot be replaced. Picture this, two superpowers working together, AI and the “gut feeling” that is the physician’s existing superpower. The future is promising.
Photo credit: Insight Exchange Network
Krina Patel is a writer/illustrator and educator committed to building provider-patient relationships. Dr. Patel’s doctoral research on the body and cognition at Harvard University and her experience adopting and promoting technology in education brings her to her current work in the health and technology sector. Follow her on Twitter, @positivelylearn.
Send this to a friend