Artificial intelligence (AI) has taken the country by storm in recent years, especially in the health care space. During the Association of Cancer Care Centers (ACCC) 51st Annual Meeting & Cancer Center Business Summit in Washington, DC, Douglas Flora, MD, FACCC, executive medical director of oncology services at St Elizabeth Healthcare, and editor in chief of AI in Precision Oncology, spoke about the challenges and opportunities posed by AI for oncology care providers.
In this interview, Flora explains how training AI to be used as a “personal assistant” for providers can help give them back more time to spend face-to-face with patients, also emphasizing the importance of protecting patient data.
This transcript has been lightly edited for clarity; captions were auto-generated.
Transcript
What specific challenges have you encountered in implementing AI solutions in oncology care, and what are some important factors to keep in mind for successful integration?
We’re early adopters, but everyone wants a little guidance. We’ve proceeded very cautiously around patient health information; we really want to make sure that’s kept sacrosanct. It’s our responsibility to our patients not to share their things, so a lot of the early discussions were governance, cybersecurity, make sure that PHI [protected health information] is protected. Those 18 things that we’re not allowed to share with anyone else, a lot of the vendors wanted access to. We had to really work out, what are our rules, what are we most comfortable with, and once we established that, everything could proceed from there.
How do you envision the role of AI evolving in oncology over the next 5 years, and what steps do you think health care organizations should take now to prepare for these advancements?
I hope everybody’s getting involved. I hope people are educating themselves. It is here. We will not put this back in Pandora’s box.
I think it’s a responsible approach for us—as leaders, as providers, as physicians—to learn this like we did immunology or genetics. We hadn’t studied that since second year medical school, and we’re using it all day, every day now. So that’s one.
Number 2, I would say, as we see these things evolve from just pattern recognizers and sort of data machines to things that really do think at the level of a human or even brighter than humans, I hope we use it for good. I don’t want these systems to exist so I can see more patients a day. I want to be able to watch the daughter of my patient if she’s getting hives because I’ve shared too much bad news, or watch the husband getting sort of sweaty and ashen because we can’t deliver that much hard stuff in one visit, and I can’t do that when I’m staring at my computer screen or looking down at my keyboard. That humanistic part of the approach for me is what I’m most excited about in the next 5 years.
Looking beyond that, I think every physician is going to need the help. It’s going to be a personal assistant for each one of us to stay current in the medical literature, to know everything about our patient without having to collect it from 10 different sources, and then I can make better decisions for my patients. So, I’m excited to have the help.
