Healthcare Advancements in a Digital World: Hello Artificial Intelligence
Innovation Posted Sep 28, 2018 by Jennifer Torode
This week, I attended an MIT Enterprise Forum of Cambridge panel discussion with a few of the CHEN PR crew. The topic of the evening was artificial intelligence (AI) in healthcare.
The astute panel included IT Physician Lead, Senior Director of Business Intelligence and Clinical Research Informatics Jonathan Bicknel, MD of Boston Children’s Hospital; Director of Machine Learning Research Hunter Elliott of PathAI; Research Fellow Catherine Kreatsoulas, Ph.D at Harvard School of Public Health, University of Toronto; and Associate, Electrical and Computer Technologies Dan Rudoy, Ph.D at intellectual property law firm, Wolf Greenfield. Moderating the panel was Phillip Machnik of Patient Keeper.
Over the last several years, we’ve seen an influx of terms such as artificial intelligence, machine learning and deep learning and there’s a lot of confusion around what they all mean and how they correlate to one another. To explain without getting into the weeds, at the core of AI and machine learning is an algorithm, which is a set of rules that are followed when solving problems. In the healthcare context, algorithms and software are used to approximate human reasoning in analyzing complex medical data without direct human input. If you want to really dig in and learn more about these terms, Data Science Central provides clarity and context.
In today’s rapidly changing and technology-driven world, the healthcare sector has embarked on the journey of exploring and implementing AI to maximize efficiencies across practices such as drug development, diagnosis processes, personalized medicine and patient monitoring and care. Some institutions such as Massachusetts General Hospital, Memorial Sloan Kettering Cancer Center, the Mayo Center and others, have already developed AI algorithms and tech giants like Google and IBM have also developed AI specifically for the healthcare sector.
AI being able to diagnose a patient through algorithms had an opposing side, though. Dr. Bicknel said at times, he was able to predict what was wrong with his patients before any machine could by simply looking at them. While this remains true at times, Hunter Elliot was optimistic and said the goal is to work to a point where “algorithms provide doctors with new information about patients previously unknown, and then that will help AI show its true value instead of just being stuck doing the same thing doctors can already do faster.”
The panelists spoke enthusiastically about the amazing benefits that AI can bring to medical professionals and patients, but they also realize the journey is not a simple one. With such vast amounts of complex data to sort through, not enough skilled professionals can keep up with the demand, and with privacy and regulation laws, it gets complicated.
Mr. Elliot from PathAI, a company that’s developing AI software aimed at helping pathologists be more efficient and accurate in diagnosing disease, noted that AI requires specific blended skillsets—medical and mathematical—and there are not too many professionals that possess both, which is a challenge. In fact, there was a gentleman in the audience who was in his last year of medical school and he told Elliot he wanted to steer his career towards machine learning and asked what Elliot looks for when hiring. Elliot said he looks at those individuals that possess strong mathematics backgrounds.
Ms. Kreatsoulas shared many examples throughout the discussion and one of her stories stood out. She said her friend who analyzes data for a living, wore a medical device to monitor a health concern. When she asked the device company to see the data so she could analyze it, she was told they could not provide as they own the data that was collected.
There was a doctor in the audience who said he’s practiced medicine for decades, taught numerous medical classes throughout his career and helped hundreds of medical students prepare for medical board examines. He said that he could walk into an emergency room (ER) and just by sight, he could tell which patient would die, who would live, who needed immediate attention, etc. He asked the panel if they thought there would be a way to have a computer in the ER that could scan the room of patients and produce data that would be able to provide such an analysis. One panelist grinned and said that we’re definitely not there yet and that capability is a bit far off in the future.
Admittedly, science and math were never my strong suit. There were times during the discussion where I found myself smiling as I could not grasp some of what they were discussing amongst themselves. I felt a bit like a fish out of water. I was blown away at how incredibly committed, passionate and intelligent the panelists were and grateful there are people like them in this world.