It’s late December 2019 at the Ritz Carlton in Laguna Niguel, California, and the revolution that AI-inspired technology is having on healthcare, continues apace. From Anthony Chang’s virtuoso performance in presenting an AI in Healthcare Primer, to the closing comments – attendees were treated to a wide ranging cornucopia of topics, presentations, speakers, and entertainment.

AIMed19 continued to dazzle as over 500 clinicians, entrepreneurs, industry representatives, data scientists, innovators and academics from around the world gauged the progress of AI
in healthcare. Specialties covered included: of course radiology and pathology, but also cardiology, cancer care, pediatrics, medical education and oral health with the latter winning the exciting Shark Tank competition (see page 64 for full story).

The first day kicked off with a reflection by AIMed founder Anthony Chang, recovering from his own recent experience of being a critically ill patient. He went on to introduce Christine Schweer and her son who, with the aid of a voluminous beach bag, vividly described their excruciating experience with an unknown, chronic and life threatening disease where the family has to be the bearer of all information in search of a diagnosis and therapy. These personal reflections gave a compelling start to the meeting, keeping the focus not only on emerging AI technologies but, importantly, on the needs of patents. Throughout the meeting, many speakers reminded us to never forget that the data comes from and belongs to the patients. They have paid a price for our data. This charges us with an ethical responsibility to respect their data and to learn from their data to help others. This patient focus was expanded to encompass diversity and an international outlook in the application of AI in healthcare.

The opening keynote speaker, Ted Shortliffe, a true pioneer in the application of AI to healthcare, reminded us that we are now fifty years into AI in medicine. From expert systems and early computational aids to the widespread adoption of big data, data science, machine and deep learning, AI in medicine has been through the valley of disappointment. Currently, in an explosive phase of interest and growth, we still have problems with integration, interoperability, terminology standardization, data ownership and others. Ted’s long perspective reminded us we stand on the shoulders of giants and must persevere.

Kathy Jenkins then led a panel on adoption. Throughout the meeting this topic was addressed by many, including Arta Bakshandeh, who, with others, reminded us of the necessity to keep it real, do not over-offer, be honest, focus on solvable problems, test multiple data science solutions for any given problem, learn from our mistakes and remember the ROI, in health, in reduced disease burden, in patient satisfaction as well as decreased costs. To drive AI adoption in healthcare, deep collaboration is necessary among healthcare systems, physicians and industry to understand what is needed and what will work. Finally, broader approaches, including cognitive sciences will be required. Vimla

Patel described the fusion of deep learning AI, symbolic AI and explicability as the next important phase in the future of AI decision making.

Transformation of many areas of medicine is well underway, as Yigal Samset and Kevin Maher discussed in cardiology. They highlighted the importance of building a system of algorithms, applications, care protocols, focusing on episodes of care and weaving these into a culture-changing system that captures data while iteratively creating a virtuous cycle of continuous improvement. A focus on anticipatory treatment melding omics, image interpretation, radiology pathology and clinical data into early, even pre-disease identification is necessary to drive the personalized health care that AI promises. Several speakers throughout the meeting called for cultural transformation of health care that AI will require to be fully integrated in our patients’ care. Throughout the meeting, it was stressed that the need for extensive, coordinated involvement throughout medicine will be necessary for this cultural transformation.

Several speakers addressed the explosion of AI in healthcare and its consequences. These include great growth, but confusion, risk but great hope and numerous challenges from FDA regulation, legal and regulatory aspects to insurance questions. Eric Fish, Lynda Chin and Jesse Erhenfeld, Chairman of the AMA, explored regulatory aspects of AI. The doctors’ dilemmas around AI were discussed: Does it work? Will I get paid? Worse, will I get sued? Will it work in my practice? Can AI fundamentally change healthcare? Attendees were reassured that medicine has always adopted new technologies and will grow along with AI. Success will require focusing on the quadruple aim: better outcomes, improved patient satisfaction, lower cost, and improved caregiver’s experience.

This review would be woefully incomplete without mentioning the comprehensive kaleidoscope of workshops and abstract presentations.

Numerous early morning and luncheon workshops with such topics as: From Data to AI; Cognitive Computing and How Clinicians Think; Getting Clinician Adoption; NLP in Healthcare; Deploying MI in Health Care; and Blockchain and Cybersecurity were all fascinating, informative and generative. These gathered likeminded people from industry, academia and data science worlds to discuss and explore topics from diverse perspectives

often revealing insights, such as the need for cultural change, focusing on the economics of developing AI, and involving physicians at all stages from concept to applications at the bedside. The afternoon abstract presentations were informative and inspiring with tremendous contributions from
young students and investigators. The exciting three days left all attendees with eager anticipation for AIMed20.

Randall Wetzel is Founder of The Virtual Pediatric ICU