Episode Summary: Venture investing in AI healthcare applications has been on the uptick and is directly related to the subject of this week’s episode: just how the healthcare industry is (and isn’t) being impacted by innovations in AI technology. Guest Steve Gullans of Boston-Based Excel Venture Management talks about some of the various healthcare-related ML and AI applications that he sees being brought to light, and touches on which innovations have a better chance of getting blocked and redirected by parties of interest and those that have more promise in being accepted and rolled out sooner.
By the end of this episode, listeners will have a more clear picture of practical considerations in healthcare technology adoption, reasons that are often less about quality or potential of the technology and more about clarity on ROI for investors.
Expertise: Life science technologies and venture investments
Brief Recognition: At Excel, he focuses on life science technology companies with a particular interest in disruptive platforms that can impact multiple industries. Prior to Excel, Dr. Gullans co-founded RxGen, Inc., a pharma services company, where he served as CEO from 2004-2008. In 2002, Steve stepped in as a senior executive at U.S. Genomics to direct operations. In the 1990s, he co-developed the technology that launched CellAct Pharma GmbH, a drug development company. Steve’s experience with venture investing began in the late 1980s when he became an active advisor to small bio-techs and venture investors, including as a senior advisor to CB Health Ventures for 10 years.
Dr. Gullans is an expert in advanced life science technologies and was a faculty member at Harvard Medical School and Brigham and Women’s Hospital for nearly 20 years. He has published more than 130 scientific papers in many leading journals, lectured internationally, and co-invented numerous patents. He recently co-authored, with Juan Enriquez, the book Evolving Ourselves: How Unnatural Selection and Non-Random Mutation are Changing Life on Earth. Gullans received his B.S. at Union College, Ph.D. at Duke University, and postdoctoral training at the Yale School of Medicine.
Current Affiliations: Board Director at Cleveland HeartLab, Gemphire, Molecular Templates, NofOne, and Orionis; fellow of the AAAS and the AHA
Big Ideas:
1 – Healthcare AI Adoption is Resisted in Patient-Facing, Physician-Replacing Applications
1 – Healthcare AI Adoption is Resisted in Patient-Facing, Physician-Replacing ApplicationsIn healthcare, innovations that augment operation processes are more likely to be accepted than those that threaten to replace skilled human workers. For example, easing clinician workflow through smart scheduling/dictation is likely to be embraced, applications that are patient-facing (taking vitals) or physician-replacing (radiology image analysis) will likely be met with resistance and even fear. Gullans believes that these “resisted” technologies will be adopted first in healthcare environments that lack the human talent in the first place. With limited resources and no specialist doctors to feel threatened, an understaffed remote rural clinic may be more likely to use machine vision for x-ray diagnostics.
Turning Insight Into Action
Think about how your company’s innovation/technology affects key parties involved in the healthcare system—does it help patients but potentially hurt doctors in terms of job security? The more helps, the more likely the technology will gain traction in the nearer term. If certain parties are more “negatively” impacted, are there particular communication strategies or technology modifications that can help ease fears/smooth transition?
Interview Highlights:
The following is a condensed version of the full audio interview, which is available in the above links on Emerj’s SoundCloud and iTunes stations.
(3:18) Where have you seen AI and ML whittle its way into businesses and pitches form your perspective?
Steve Gullans: I focus generally in healthcare, though it’s broad ranging…almost anywhere you have lots of data. Our firm is more general, but in terms of my specialty — that’s where I sit. What we’re seeing is a convergence of a lot of the AI technologies and healthcare data generating platforms starting to talk to each other, with smart people on both sides.
(5:25) Give us a quick sense in the healthcare space…what’s the stuff that hasn’t gotten you excited versus the things that really feel like they have promise?
SG: The ones that are most exciting are the ones that have a team that include both health care professionals and AI professionals, because in general the AI people show up saying “as soon as someone gives us lots of data we’re going to find all the secrets”, and then the healthcare professionals don’t really know what’s possible — when you get them together you start to see some compelling applications…
(8:23)…the way in which it’s adopted is not going to be very natural, in terms of every area of the economy except healthcare; you really do have embedded physicians nervous that it’s going to take away the control and may make the wrong decision.
(10:58) It sounds like for AI’s adoption, you’re already really heavily considering where the pushback is going to be and figuring out where the least amount of pushback is going to be…it sounds like the threat to the specialists is pretty strong and we’d have to find ways around it…what are the ways around that?
SG: It is tough, but there’s always a few beachheads that are going to pay off; there are some applications right now where physicians don’t enjoy a particular kind of call or one where having some assistance can actually be a big benefit to everyone involved…I think what you’re going to see is very specific populations within a particular setting, such as calling a stroke in the ER as bleeding or non-bleeding, where there’s a life and death decision that’s very binary…but other applications will take a little more time, and obviously the AI is needed most where the decision-making is hardest, so there’s going to be reluctance to adopt it.
(16:02) You had mentioned drug discovery is a different pot of money…you have a massive incentive to do whatever it takes to get a drug out the door, and if that means one or two (people) have to get moved or laid off — if the company stock price rises, there’s less pachinko machine to deal with because it’s a pretty clear ROI assessment…do you feel similarly?
SG: I do…and it’s actually a confluence of a lot of factors. One is biological drug discovery technologies, which have been traditionally very poor signal-to-noise; that’s starting to change, so you actually have more reliable data right now…there’s also enormously large data sets that are publicly available to supplement what you have, and finally the computational powers I mentioned in the cloud in particular and GPUs…make it possible to do structures of molecules or inventory what you have in your screen in an hour or an afternoon, rather than a year.
(Note: Readers will a keen interest in drug discovery may want to read our full article on pharma applications of AI)
(18:21) Whether it’s understanding how to fill appointments in or project absences…you had mentioned things about managing patient flow too…that feels like a business intelligence application that we’re using in the healthcare industry, but that doesn’t feel to me like we’re inherently in healthcare?
SG: It is; in fact I’m involved in one particular project, where we’re pointing all these tools at business databases, and in this case what’s different about the biotech field — biotech lives and dies by intellectual property, and so you actually the way to use that as part of your rosetta stone, for looking for diamonds in the rough or opportunities that are undervalued, by looking through the patent histories and seeing what’s out there that people don’t appreciate, on a global basis…you could actually find things that were invisible before.
