Between AI chatbots, social media influencers and websites offering medical advice, it can be difficult to parse the vast amount of health information online. It also presents a challenge to physicians and other healthcare providers who are fielding questions from patients based on what they’ve read online.
Kristin Whitman is the head of information and learning services at the Oregon Health & Science University Library. Fabrice Jotterand directs the university’s Center for Ethics in Health Care. They join us to offer advice on how to navigate chatbots and other online resources in health-related situations.
Note: The following transcript was transcribed digitally and validated for accuracy, readability and formatting by an OPB volunteer.
Dave Miller: From the Gert Boyle Studio at OPB this is Think Out Loud. I’m Dave Miller. Between AI chatbots, social media influencers, and websites offering medical advice, it can be difficult to parse the vast amount of health information online. This is an issue for the general public. It also presents growing challenges for healthcare providers who are often fielding questions from patients based on what they’ve read online. Kristin Whitman is the head of learning and information services at the Oregon Health & Science University Library. Fabrice Jotterand is a philosopher and medical ethicist who directs the university’s Center for Ethics in Health Care. They join us now to talk about wading through healthcare information and misinformation. It’s great to have both of you on the show.
Kristin Whitman: Thanks for having us.
Fabrice Jotterand: Thank you for having us.
Miller: Kristin, first, before we talk specifically about AI and disinformation or misinformation, I just would love to understand the two different professional worlds that you both come from. What does your job entail as a medical librarian?
Whitman: Medical librarianship entails working with health information. So when you think of a librarian, frequently people think of books. We are more engaged with medical research that’s often published in journals or coming from health organizations, health agencies and we help clinicians, students, faculty, and even patients engage with that information.
Miller: And Fabrice, what role do philosophers or ethicists play in a medical school, in a university hospital?
Jotterand: Well, I would say there are two dimensions. One is address ethical issues and be available to clinicians, to staff, and to students if they encounter ethical issues. The other is to think philosophically about some of the challenges encountered in healthcare in general. For instance, the introduction of AI in healthcare raises tremendous philosophical questions. Not just ethical questions, but also philosophical questions in terms of expertise, in terms of agency. So you have these two dimensions that an ethicist will be engaging in.
Miller: It seems, Fabrice, like one of the challenges of wrapping our heads around AI both from a policy matter or from a technological standpoint is that it’s changing so fast. What are your constants when you think about AI as an ethicist given the speed of change?
Jotterand: So it’s always a kind of risk-benefit calculus, right? You have to address the risk associated with the implementation of AI in healthcare. But you cannot have benefits without risks, so you have to look at also the benefits, and what are the benefits of implementing AI in healthcare? And so this is a constant. But the issue here with AI is it’s not a question of kind, it’s not a new kind of technology, but it’s a question of degree. Because AI can drastically change healthcare, drastically change how we think the role of clinicians in hospitals, and then also in terms of how we’re going to educate the next generation of healthcare
professionals.
Miller: Do I understand you correctly there, that you’re saying that the questions that are arising from AI are in some ways fundamentally different than other questions that have arisen from previous technologies? Medicine, for 200 years, has been the successive history of new technologies. But you’re saying that this is somehow truly new.
Jotterand: Well, yes, it’s truly new in the sense, it’s not the first time that we introduced technology in medicine. In the 70s or 60s, 70s, 80s, you have the introduction of computers, and this changed the paradigm of the clinical encounter. It’s not just the clinician and the patient, but suddenly you have a screen and the clinician turning his head toward the screen and typing. But with AI, the issue is you’re going to have in the future, and maybe it’s already occurring, you’re going to have AI making decisions for clinicians, and we know that AI is more accurate in terms of diagnosis, but medical practice is not just about diagnosis, so you need the human component. But my worry is about allowing these machines to make decisions for human beings, for patients, for clinicians, undermining the agency of the clinician, but also to a certain extent making decisions for the patient.
Miller: Kristin, I want to turn to the information side of this. How has AI already been affecting your work as a medical librarian?
Whitman: AI has really changed the way that we interact with information. So, AI chatbots, for example, you ask a question in natural language, it gives you a very confident sounding answer. And if you’re not familiar with the kinds of errors that these tools can make because of the way that they’re created and trained, and the way that they work, you may be led into believing a claim that is not actually true. It could be a completely false or just a misrepresentation of the source information that that chatbot is working from.
Miller: What are you hearing from your medical colleagues, Kristin, about the way their patients are using AI right now?
Whitman: Medical colleagues are so used to Dr. Google, right? And so they’ve already…
Miller: For 20-plus years.
Whitman: Yes, yes.
Miller: I’m dating myself, actually about close to 30 years. I think that’s the Google anniversary, closing in on it, yeah.
Whitman: Totally. So, we’ve been for a long time, we’ve been able to search our symptoms and convince ourselves that we have something really of issue. Doctors have been seeing us for a long time. And now it’s just a similar kind of issue where people have been relying on chatbots and coming in and bringing in these potential things. So it’s a similar conversation where we have to talk to patients, go back to, where did you find this? What was the actual source that the chatbot was working from, who posted that information? What was their motivation? Are they trying to sell you something? Is this just one person’s experience? How can we learn more and work with that patient to kind of figure out how to continue to keep trust with that patient while sort of introduce these information literacy concepts?
Miller: Those are all obviously, those are great questions to ask, but I don’t imagine that by the time the conversation that I’m imagining is happening, this is in some patient room, in say, a primary care provider’s office. At that point, my guess is that many patients couldn’t answer any of those questions. They’ve come in, they’ve gotten good or bad advice, good or bad news in this scenario from a chatbot. I’m curious what kind of conversation, at that point, you would like clinicians to have.
Whitman: Well, I do understand the time constraints, right? But what librarians frequently recommend for basic information literacy is stopping, taking a step back and thinking about those sources. So, introducing very basic questions like what tool did you use, let’s talk briefly about how that tool works, is the beginning of a conversation that I would think I would recommend. And also, it may be helpful to identify that emotion is part of this conversation. What you found online is making you feel hope or making you feel fear.
Can we put a pin in that and remember that wherever we encounter this information, if it’s online, if it’s on social media, that someone may have posted it with the intention of making you feel emotion, maybe the intention of making you buy something. So if we can just start from a place of, I recognize that you’re feeling hope or fear, let’s take a step back and get some distance from that and attempt objectivity and start thinking about motivation behind what’s been posted online. That’s what I’d recommend.
Miller: Fabrice, what do you think people are looking for when they engage with a chatbot and seek medical information from some AI creation?
Jotterand: Well, it’s just practical, right, but I think one of the problems we have in our current healthcare is the introduction of electronic healthcare records. And so you have an obligation to post the results of tests online and so a patient will have, let’s say it’s Friday afternoon, 4 o’clock, and you have a test and you get an email, you have a text, go on MyChart and check what you have the result.
Miller: Right, and then you see that that some particular level is very elevated. It’s in the red.
Jotterand: Yes, your cholesterol is 250, but if you don’t know about medicine, then you’re going to have anxiety. So what you’re going to do, you’re going to search online, you’re going to try to find a solution and try to interpret what is given to you. And so this is the dehumanizing of healthcare from my perspective. And so when you use technology, I was listening to the previous segment, we want quality, not volume, we want decreased cost, you use these technologies, it’s very efficient. You might reduce cost, but there is a human cost that we need to assess, because intrinsically, clinical practice is a human endeavor. Once you remove that, you reduce the role of the physician to a technician, basically. And so that’s the problem that I see that is emerging in healthcare.
Miller: So, what is your advice then about how the people in the building where you work, the frontline healthcare providers, can increase the trust that their patients have in them?
Jotterand: That’s the challenge, because to build trust you need time. I think on average a physician will spend seven to eight minutes per patient. So then you have nurses, they will spend more time, but let’s talk from the perspective of a physician. So it’s very difficult to build trust. Moreover, you might go one day to the same clinic and you might have another clinician. So that’s very difficult to build trust. And then there’s a particular context of our culture. We don’t lack expertise anymore. We live in a post-truth era where it’s very difficult to have meaningful conversations, difficult conversations.
So once you understand the context, one, you understand that people, they don’t like experts. So you have to use a different perspective and build trust and hear the patient, respect the patient, then the patient will trust you and then you can use your expertise. Because they will come with articles, with facts, and they will push back. So I would advise clinicians to recognize that particular context and to find strategies to build that trust and to use the expertise so that you can debunk misinformation, but that’s very difficult because of the particular context in which we talk about healthcare facts.
Miller: Kristin Whitman, how do you feel about what Fabrice just said, that we live in a post-truth era? How do you feel as somebody who focuses on helping people get accurate information?
Whitman: Well, it’s an important question, and I almost feel a little bit of hope about the introduction of chatbots for information research if people are able to gain a little bit of literacy about how they work. Because you can, when you prompt a chatbot, tell it, “use high-quality medical sources.” You can say, “show me a balanced picture of the evidence.” If you don’t do those things and you let it make choices for you, there’s this element of it wants to make you happy.
It’s going to try to anticipate what you’re hoping to hear and maybe fulfill that, right? But if you have literacy to know that it is choosing sources for you, it may not be choosing the best sources, but you have the power to click those links in its answer and go investigate those sources and then decide what you feel about those sources. Is it a trusted news outlet? Is it one person’s experience? Is it someone trying to sell you something?
And if you explicitly instruct that chatbot to give you a balanced picture of the evidence, that actually can sometimes be helpful when you are looking at one claim that may have been posted by someone who’s looking to hook you or make you feel emotion and give you a broader picture. It’s just that this inherent element of it’s trying to please you, that’s what worries me, but people don’t have a basic literacy about what these things really are and what they really do.
Miller: Kristin Whitman and Fabrice Jotterand, thanks very much for your time. I appreciate it.
Whitman: Yeah, thank you.
Jotterand: Thanks.
Miller: Kristin Whitman is the head of learning information services at the OHSU library. Fabrice Jotterand is a philosopher and a medical ethicist and the director of the OHSU Center for Ethics in Healthcare.
Miller: Finally, today, our managing producer Sheraz Sadiq joins me in the studio to read some of your recent feedback. Hey, Sheraz.
Sheraz Sadiq: Hey, Dave. Last week we had a wide-ranging conversation about menopause, and we asked people on Facebook what helped or hurt them most when they were going through it.
Miller: RA Adel wrote, “What helped? Getting educated on my own body and hearing from others experiencing the same things. 45 years of life, and I had no idea. Why is this secret so well kept?”
Sadiq: Adrienne Hass wrote, “No one spoke about it. So nice that now there’s an open discussion. Finally, the experience is being acknowledged and welcomed and shared so that women have resources that weren’t openly available in the past.”
Miller: Jessica Bardsley wrote, “The increases in allergies have been terrible. With perimenopause, I’ve developed chronic hives in response to apparently nothing and everything. Estrogen is apparently involved in the control of mast cell and histamine production. So when estrogen levels drop, histamine increases. It’s completely changed my daily life. I’m taking high levels of antihistamine daily and not being able to enjoy being outside as much as I’m more sensitive to environmental triggers. The hormone replacement therapy doesn’t seem to help.”
Sadiq: LA Starcevich wrote, “As someone who’s had estrogen-positive breast cancer, learning that early research on hormone therapy has been revised and that there are safe options has helped.”
Miller: Michelle Gila wrote, “I resisted hormone therapy for years. Now I know you’ll have to tear that estrogen patch and progesterone cap from my cold, dead hands. I will never go back. I do wish I’d started during perimenopause. I wish I would have just walked out and found someone who would listen to me. Lots of body changes and pain and discomfort could have been avoided if only. So now I tell everyone who will listen, men too. Most men know women. Everyone benefits from being well informed, and science continues to improve treatments and safety, though it’s been miserably slow for half of the world’s population.”
Sadiq: Just a few days ago, we had a conversation with Portland music producer Larry Crane. And we asked folks on Facebook, if you make or produce music, what part of the creative process do you love the most?
Miller: Mike Grabarek wrote, “Finally finding an idea I truly love after weeks of frustrated noodling. It might not sound fun, but it is so worth it.”
Sadiq: Ryan Miller wrote, “The composing process itself. There’s almost a letdown when I finish something I was fired up about.”
Miller: And Hakeem Hasworth wrote, “Writing music is a form of nonverbal therapy. For me, it’s a monologue, but I’m the audience and the speaker simultaneously. It’s therapeutic when the hands communicate what the heart is feeling, and often surprising.”
Sadiq: A couple of weeks ago, we had a conversation about the Olmsted design philosophy and public parks in Portland. We asked about your favorite parks.
Miller: Dave Dyk wrote, “Gresham’s main city park, home to the Gresham Japanese Garden. Truly something special in the region. If you haven’t seen it, take a ride out the Springwater Trail and check it out.”
Sadiq: Amanda Briles wrote, “Gotta say, Mount Tabor. How many places can you hike on an extinct volcano? We have tons of great parks for sure, but this is my favorite.”
Miller: And Sydney Casper wrote this, “I love so many of our parks here in Portland. Washington Park has just about everything all in one place. A person can spend a day exploring the gardens, find a little tucked away pocket of grass surrounded by trees and flowers for a nearby private picnic, stumble upon a concert or a play being performed, or climb up on the empty stage and put on an impromptu performance of their own. Play tennis, practice archery, visit the zoo, head off for a hike through the forest, and it’s all right here in the city. That is pretty hard to beat.”
Sadiq: Thank you so much for all your comments and suggestions. We always read them and so appreciate them.
Miller: You can leave more on our Facebook page. We are at OPBTOL.
Sadiq: You can also send us an email at thinkoutloud@opb.org or leave us a voicemail at 503-293-1983.
Miller: Thanks, Sheraz.
Sadiq: My pleasure.
“Think Out Loud®” broadcasts live at noon every day and rebroadcasts at 8 p.m.
If you’d like to comment on any of the topics in this show or suggest a topic of your own, please get in touch with us on Facebook, send an email to thinkoutloud@opb.org, or you can leave a voicemail for us at 503-293-1983.
