You’re in the doctor’s office, sitting on the itchy, wrinkled paper on the exam bed for longer than you’d like. Your family doctor finally knocks on your door, and you’re eager to explain your reason for booking the appointment. But as soon as you start speaking, the doctor is glued to his computer, typing rapidly and rarely looking up. They are taking notes while listening, but lack of connection leads to disconnect. Artificial intelligence may provide a solution.
Doctors are starting to use ambient listening and AI scribes to take notes during medical visits, freeing them up to better interact with their patients, The Philadelphia Inquirer reports. This follows a number of studies published in JAMA Network this year (February, August and October) that highlight the usefulness of such ambient AI scribes for physicians. However, some experts are raising red flags.
doing the grunt work of doctors
Bracken Babula, a primary care physician at Jefferson Health, and Dina Francesca Capalongo, an internal medicine physician at Penn Medicine, told the outlet that they use AI Scribe to help them take notes after asking their patients for permission to record their conversations. The tool records conversations and converts them into organized notes. This has helped both physicians change the dynamics in their exam room, allowing them to be more engaged in conversation and maintain eye contact.
These sentiments are supported by research. The above study involving 46 participants linked the use of an ambient authoring tool with greater physician efficiency, less mental burden of documentation, and an increased sense of connectedness with patients.
As a professional whose career once depended on intensive note-writing, I understand the extreme convenience of such tools better than most. When I first started in journalism, I rarely looked at my interviewees – I was too busy typing out their responses at the speed of light. When I finally got good transcription tools, it was life-changing. With permission, I now record conversations and then transcribe the audio file into text that I refer to when writing my articles. This gives me more brain space to be friendly, ask spontaneous questions, and respond appropriately to surprising answers.
Potential Problems with AI in the Exam Room
However, there is a difference between transcription and note taking. The latter requires understanding what information is important enough to keep. If my transcription tool detects an error and makes it into the published article, we issue a correction. But if an AI prescriber writes the wrong dose, I think it could have serious consequences for the patient. According to the Philadelphia Inquirer, the tools may also struggle with foreign accents and raise questions over privacy and medical malpractice issues. Additionally, health care systems have to figure out what to do when patients don’t want something recorded.
“It puts a lot of responsibility on health systems to approach that with caution and kick the tires,” bioethicist I. Glenn Cohen of Harvard Law School told the outlet. It seems that the final decision on the use of AI in the exam room will ultimately be determined by the guardrails and integration process implemented by health care systems.
It is certain that I and many of my former classmates would have given anything to have access to such a tool during my endless university lectures.