Two decades ago, when I started practicing emergency medicine, patient documentation was a manual process. I wrote on forms of carbon copy which were then stored in the belly of the hospital, in the archives, as patient files. To obtain a laboratory test, an x -ray or a medication for a patient, I had to note the instructions on a piece of paper, track down the right nurse and verbally transmit the orders I placed. The phones were not intelligent and the fax was always a cornerstone of communication in the department. Each task required work and legs of the legs.
The advent of electronic health files has considerably improved the workflow. Suddenly, which has taken several steps could be completed in a few clicks. But, as suppliers, our eyes have become glued to the screens that helped us. We juggled in charge and medical notes while speaking to patients. Physically, we turned our backs on our patients and our faces to brilliant digital workflows in front of us.
Thus, many emergency services have turned to medical scribes. These trained professionals would document meetings for patients. Their presence was associated with an increase in the productivity of doctors, in particular during primary consultations, where productivity was noted increase by 25.6% according to a Multicentric randomized trial Posted in the British Medical Journal.
More importantly, having a scribe meant that I had more quality time with my patients.
Now, ambient listening technology led by AI pushes the concept of medical scribes a little further. Instead of counting on a person to document each word and each action, AI can capture and transcribe doctor-patient interactions in real time, without additional staff.
Today, when I enter a patient’s room, I no longer wear pen and paper or I have a trailer scribe with his laptop. My smartphone, equipped with an ambient listening technology, captures the conversation.
In the state I practice, only consent to a party is necessary to record. However, I present the process to the patients: “I use a tool that helps to document our conversation so that I can focus entirely on you – are you okay?” So far, no one has refused. Patients understand that documentation is a necessary part of the medical examination – in one way or another, the table should be completed. Many seem to appreciate that this AI assisted approach allows a more natural interaction and face to face – without distraction of a supplier typing on a computer or scribble of notes during the visit. As I come back to my office after having evaluated several patients, drafts of all my notes await me – ready for publishing and finalization.
I have used this system for almost a year, although I have been slow to adopt, which only turned to it intermittently at the start. When my hospital removed the medical scribes, I dictated my notes, believing that this is faster than the AI -based system. I first reserved an ambient listening for simple cases – those in whom I trusted the software to manage with precision. At the beginning, the process also involved an additional step: the copy of the note of an external platform in the medical file, a clumsy bypass solution that did not call me. But recently, an entirely integrated version has been introduced, incorporating ambient listening directly into the electronic health file. Once I started using this transparent system, I became a believer.
The advantages of this technology are clear – it is fast, effective and allows more time at bedside, focusing on what matters most: the patient. But, like all technologies, it is not without its limits. The AI does not always transcribe conversations perfectly, and the notes generated sometimes require reflected modification. According to my experience, however, the challenge is generally not a factual inaccuracy. When AI listens, it tends to make history well. The problem is more often a clinical accent – a question of identification with precision which elements of the meeting guarantee greater attention, especially in emergency medicine, where the ability to highlight what is most immediately threatening life or members is critical. I found that the solution is surprisingly simple: providing clear and verbal summaries to the patient and the ambient listener.
In fact, the system excels in capturing these summaries. Above all, when I take the time to repeat the story of a patient and to describe the treatment plan, I make sure that the patient and I have a shared understanding of the next steps. I articulate these summaries more deliberately because I know that the AI listens. That makes a better note. However, the real advantage, I find, is the detailed and formalized understanding between the patient and the doctor. I am able to reach a new level of clarity and mutual agreement in each meeting with a patient.
Looking towards the future, ambient listening technology focused on AI in health care could evolve in such a way as to anticipate clinical needs. For example, the AI can analyze my verbal synopsis and automatically generate relevant control sets, identify the necessary consultations or even recommend patient comfort measures. These capacities could further rationalize workflows, reduce the cognitive load and give me to me and my colleagues more time to focus on direct care. Such progression would naturally reflect the role of traditional medical scribes, which not only documented meetings but have also carried out various office tasks to support doctors.
Whatever to come, it is essential to approach any integration of the AI, ensuring that the complete AI, rather than replaces, the clinical judgment. I am thinking of the note generated by AI as a project written by a very attentive but inexperienced student: he captures the essence of history but sometimes lacks the hierarchy of the emergency that emergency medicine requires. I have never found dangerous or misleading errors, but I often revise the structure or accent to better reflect what is most clinically important. In this way, AI becomes a useful partner – listen, writing, even incentive – but always forcing the doctor to shape the final story.
For the moment, for me, the promise of ambient listening technology focused on AI is simple and immediate: it allows me to be more present with my patients and their families.
Iyesatta Massaquoi Emeli, MD, MPH, is a distinguished doctor and assistant professor in emergency medicine at Emory University.
