October 24, 2025
2 min read
Key takeaways:
- About 66% of patients found a custom-tailored AI platform to be useful in addressing questions prior to total joint arthroplasty.
- This platform can be used by surgeons to interact with patients at any time.
DALLAS — Preliminary results presented here showed use of a custom-tailored AI platform may be a feasible option to address preoperative patient questions for total joint arthroplasty.
“As we move into this age of AI, it is becoming more and more robust. We’re trying to make sure we eliminate as much hallucination or misinformation by teaching models what needs to be known correctly,” Karlos E. Zepeda, DO, MS, clinical research fellow at Hospital for Special Surgery, told Healio in an interview at the American Association of Hip and Knee Surgeons Annual Meeting. “Because it’s self-learning, as long as you have some human oversight with this, I think we will do well. This helps and serves a need within the health care system to ensure that patients are getting the right information at the right time and, ultimately, leading to better outcomes for their health.”
About 66% of patients found a custom-tailored AI platform to be useful in addressing questions prior to total joint arthroplasty. Image: Adobe Stock
Zepeda, Eytan M. Debbi, MD, PhD, and colleagues at Hospital for Special Surgery created and validated a custom-tailored AI platform to address preoperative questions from 32 patients scheduled to undergo TJA (knee = 16, hip = 16) based on information preapproved by four arthroplasty surgeons. Primary outcomes included patients’ satisfaction and perceived effectiveness of the custom-tailored AI platform, with satisfaction and perceived effectiveness of the custom-tailored AI platform in the postoperative phase serving as secondary outcomes. According to the abstract, survey metrics included patient utilization rates and ratings on understandability, usefulness and likability using a Likert scale.
“One of the concerns we had with the study was that 50% of the cohort was older than 65 years, meaning that, with AI being a new technology, accessibility was one issue that we were hypothesizing might be a problem,” Zepeda said. “The results from our study, however, indicated that around 85% of the patients were able to access it with no issue.”
In addition, Zepeda said 66% of patients found the custom-tailored AI platform to be very useful, and around 84% of patients understood the information provided to them and 80% reported it as a tool they would recommend to others.
“This is a conduit. It is a right arm for the surgeons to be able to still interact with their patients at any given time so they feel more prepared and more at ease throughout the surgical experience,” Zepeda said.
In an analysis of submitted questions, Zepeda and colleagues found surgery logistics (22.8%), postsurgical concerns (22%) and activity guidelines (12.9%) were the most frequent themes, followed by postoperative expectations (10.4%), presurgical concerns (9.5%) and medications (5.4%).
According to Debbi, one of the limitations of the study was that the custom-tailored AI platform did not have personal health information on the patients included in the study. However, he said this limitation presents as a future step on providing more specific information to the AI platform to enhance outcomes.
“We’re also looking toward beyond just a chatbot where right now patients are typing in questions to an algorithm,” Debbi, hip and knee replacement surgeon at Hospital for Special Surgery, told Healio. “We want to have a phone call answering service where it sounds like a live agent but it’s an AI chatbot that’s answering their questions. Even though we found that some of our elderly patients were fine with typing in questions, it could be easier for them to speak with somebody.”
For more information:
Eytan M. Debbi, MD, PhD, and Karlos E. Zepeda, DO, MS, wish to be contacted through mediarelations@hss.edu.
