Patients have been searching their symptoms online for years, but generative AI chatbots and algorithm-driven social media feeds have fundamentally changed how those conversations play out in the exam room.
Amna Jamshad, DO, an emergency medicine administrative fellow at Wellstar Kennestone Hospital in Marietta, Ga., frequently sees that shift play out in the emergency department.
“Probably in the last two, three years is when I’ve had a huge influx of patients coming in with, ‘Here’s what I think is going on. I want you to tell me if that’s true, and order these tests,'” Dr. Jamshad said. “So it’s a little bit more direct, and they’re more involved in their own care, which I think is a huge bonus as being a patient, as well. This is what I would want to advocate for myself.”
Many physicians consider the elevated patient engagement to be a win — especially as trust becomes harder for physicians to establish and maintain with their patients. One recent study from Richmond-based Virginia Commonwealth University found that Generation Z participants who follow physician-influencers on TikTok viewed them as more accessible and credible than their own physicians.
“At the same time, I think it also feeds into the AI psychosis or the AI anxiety, where AI is kind of dictating what you think you should do and kind of feeding into what you think is already going on,” Dr. Jamshad said, acknowledging the flip side of this shift. She underscored the echo-chamber nature of these tools, in which patients may be presented with information that further validates their concern as opposed to applying a medically-trained diagnostic eye.
“As much as I think AI can be a tool and used correctly, it can definitely help you, I think it also perpetuates what you’re already thinking, so it just makes it worse,” she said. “And sometimes it takes time to convince them against what AI is saying.”
Social media compounds the dynamic. Recommendation algorithms tend to surface more of whatever content a user has already engaged with, so a patient who starts with one concern can quickly find a feed full of material that reinforces it.
Dr. Jamshad described one scenario that has become familiar.
“I’ll have patients come in and say, ‘Hey, I don’t really want to say much because AI is saying exactly what I think is going on, and you should just read my AI prompt,'” she said. “And that’s one perspective. But at the same time, I want to hear it from you.”
Taking a history directly from the patient remains a cornerstone of the clinical encounter.
“As much as I can just read the AI prompt, the fact that I’m getting the history from you in the primary source, that’s very important to me,” Dr. Jamshad said. Her approach to closing the loop also runs through the chatbot.
“At the end of the workup or whatever it is, I’ll go back to it and be like, ‘This is why I think AI is incorrect, and I want you to ask AI to tell you why what we think is going on is more appropriate as opposed to what the initial concern was, and why this is an adequate workup,'” she said. “And then suddenly the AI prompt or the response changes completely.”
The exchange, she said, shows “it’s more of a tool, and it’s very much dependent on what you put into it.”
The care setting adds another layer of complexity to this development.
“I think it’s become more of, ‘I’m coming in because I already know what tests I want, and I want you to order them,'” Dr. Jamshad said. “So, it creates a level of disconnect between what patients expect and what I can give them.
“It’s one thing if you’re going to an outpatient clinic and you’re like, ‘I want a whole-body MRI.’ That’s an option if that’s what you want,” she continued. “But in the emergency department, I have to go by clinical guidelines and what’s recommended and the literature to kind of support what we’re going to do for the next steps. I think there’s a blending of those two things going on, which is I think creating the challenge from both the patient and provider relationship.”
Those conversations take place against a backdrop of broader distrust and misconceptions of how physicians and the border healthcare system function as a whole.
“One of the things that I do see more commonly nowadays is people come in with preconceived notions about healthcare, and I think it’s unfortunate where healthcare has been headed overall, but it’s become more antagonistic against providers or groups,” Dr. Jamshad said. “But at the end of the day, we’re all on the same side. We’re all trying to take care of the patients, and I think that’s really the core of it. And the fractures that it’s creating is just hurting the relationships between the provider and the patients.”
For example, physician compensation varies widely by employment model, and the industry has spent years moving from fee-for-service toward value-based arrangements. That nuance rarely reaches patients.
“[O]rdering an MRI is not going to give me a higher paycheck,” Dr. Jamshad said. “I think it creates a challenge [for the] relationship when patients think that based on the tests I order, is how much I get paid, which is not correct.”
For Dr. Jamshad, the place to start is physician education.
“I think we’ve kind of gotten to the point that not only are patients coming in with AI training, I think that should be part of physician education to kind of be able to handle these circumstances,” she said. “We have patient education, patient encounters and things like that, and this should just be part of it. I’ve perpetually heard, if you aren’t using AI, it’s being used against you. And I think that is the mindset you kind of need to approach it with healthcare, but both patients and physicians.”
At the Becker’s 32nd Annual Meeting: The Business and Operations of ASCs, taking place October 29-31 in Chicago, ASC leaders, surgeons and healthcare executives will explore strategies to drive growth, enhance operational performance, navigate reimbursement challenges and prepare for the future of ambulatory surgery. Apply for complimentary registration now.
