Physicians are experimenting with AI tools that listen to their conversations with patients and generate written notes, freeing up time to focus more on patient care and less on documentation. Despite its potential, the technology is still evolving to overcome its current limitations.
Austin Bush, MD, demonstrates a new tool at Atrium Health in North Carolina that records doctor-patient encounters and generates notes for later review. This innovation is part of a broader effort to enhance the quality of medical visits and alleviate the administrative burdens that contribute to physician burnout.
Eric Poon, MD, MPH, acknowledges his tendency to be verbose during patient appointments, often running late as a result. However, in recent months, Poon has noticed a significant improvement in his schedule management and the quality of his interactions with patients. This change is largely due to an AI tool that records conversations and generates written notes, allowing him to be more present and responsive during visits.
At Duke Primary Care in North Carolina, where Poon practices, the implementation of these AI tools has shown promising results. These tools listen to doctor-patient conversations and produce concise, organized summaries rather than full transcripts. Early feedback from physicians indicates that these tools enhance the quality of interactions, reduce documentation time, and help mitigate burnout.
Poon, who uses a tool called DAX, finds that editing AI-generated notes is much quicker than creating them from scratch. Although not flawless, this method is a significant improvement over typing notes during patient visits or deciphering handwritten ones later.
The technology is still maturing, facing issues such as transcription errors, incomplete medical details, and trivial content. Additionally, there are concerns about the potential for increased workloads if time savings lead to pressure for more appointments. Questions also remain about the storage and use of recorded conversations.
Patricia Garcia, MD, from Stanford Health Care (SHC), is cautiously optimistic about the potential of these tools, describing her feelings as "hopeful."
How the Technology Works
Matthew Anderson, MD, at Atrium Health, seeks patient consent to use software on his phone to record their conversations, a practice that has so far met with patient approval. Anderson appreciates the efficiency: "I hit start, we have our visit, I walk out. Fifteen seconds later, I’ve got a draft."
These ambient intelligence tools analyze conversations, extract medically relevant information, and produce clinical summaries. Physicians then review, edit, and approve these summaries for inclusion in the electronic health record (EHR).
Various healthcare systems are testing different AI tools across multiple EHR platforms. These include Epic’s DAX by Nuance Communications, Oracle’s Digital Assistant, Meditech’s collaboration with Google, and eClinicalWorks’ partnership with sunoh.ai. Each system aims to improve efficiency and reduce documentation burdens.
Kenneth Harper, MS, vice president of Nuance, views these tools as "augmented intelligence" that supports clinicians without replacing them.
Physician Feedback
Physicians note several benefits and limitations of these tools:
Patient Conversations: AI tools help doctors focus more on patients and less on note-taking, enhancing the quality of interactions. S. David McSwain, MD, MPH, of UNC Health Care, emphasizes the importance of being able to concentrate on patients without the distraction of documentation.
Efficiency: Poon from Duke found that without the distraction of typing, his interactions with patients became more efficient and meaningful.
Accuracy and Comprehensiveness: While generally accurate, AI-generated notes still require careful review. Physicians worry about both excessive detail and the omission of important information. Personal anecdotes, while seemingly trivial, can be medically relevant and should be included selectively.
Time Savings and Burnout Reduction: Many doctors report that editing AI-generated notes is quicker than writing their own. This efficiency helps reduce fatigue and burnout. However, there is concern that time savings could lead to increased patient loads, potentially negating the benefits.
Privacy: Recording doctor-patient conversations raises privacy concerns. Clinicians must obtain patient consent, and there are questions about how recordings are stored and used. Ensuring the security and confidentiality of these recordings is paramount.
The Future
Doctors envision improvements as the tools learn to better understand human conversations, enhancing accuracy. They also desire more customizable editing options and the ability to assist with other tasks like prescription orders.
McSwain suggests that customization will add to the tools' capabilities, making them more flexible and personalized. Garcia from SHC believes these adjustments will solidify these tools as essential aids for many physicians, echoing a sentiment from an SHC doctor who finds their days significantly improved by the technology despite its current imperfections.
Share this post