The Evolution of Clinical Documentation: From Pen and Paper to AI Scribes Reflecting on Four Decades of Change

By: Steven L. Sivak, MD FACP
June 4th, 2025

When I began my career in medicine over forty years ago, documentation was a pen- and-paper affair. We wrote notes in whatever space was available in the chart, often hurriedly between patients or at the end of a long day. These early notes were unstructured, highly individualized, and often difficult for others to read—illegible handwriting became a running joke in our profession. Yet they were also personal, and in many ways reflected the clinical reasoning and storytelling style of the physician. Over time, the practice of clinical documentation began to formalize. One of the earliest and most transformative changes came with the widespread adoption of the SOAP note format—Subjective, Objective, Assessment, and Plan. Introduced by Dr. Lawrence Weed in the 1960s, SOAP notes brought structure and a problem-oriented approach to documentation. This method not only standardized the way we wrote notes but also improved continuity of care and facilitated better communication among clinicians. As the volume of patients increased and practices became busier, efficiency became paramount. This led to the use of paper templates, especially in primary care and procedural specialties. These templates often included checklists and pre-written phrases to speed up documentation. While they saved time, they also began to reduce the individuality and nuance in clinical notes. Still, they represented a practical compromise in balancing thoroughness with time constraints.

Then came the electronic medical record (EMR). In the late 1990s and early 2000s, EMR systems began to replace paper charts in earnest. The promise was immense: centralized records, instant access, and improved safety. But for many physicians, the reality was quite different. EMRs were designed more for billing and compliance than for storytelling or clinical reasoning. Instead of focusing on patients during visits, we found ourselves typing furiously, clicking checkboxes, and navigating cumbersome interfaces. The burden of documentation grew exponentially. A 2016 study in Annals of Internal Medicine showed that physicians spent nearly twice as much time on electronic health records and desk work as they did on direct clinical face time with patients (Sinsky et al., 2016). This shift had a profound impact on physician burnout, which has reached epidemic levels. The joy of medicine—the human connection—was increasingly overshadowed by administrative demands.

To address these challenges, many turned to voice recognition dictation tools like Dragon. These systems allowed physicians to dictate their notes directly into the EMR, reducing typing time. While an improvement, they were not without issues: recognition errors, the need for proofreading, and incompatibility with some EMR workflows limited their effectiveness.

Enter the newest chapter: ambient listening and AI-powered medical scribing. These technologies use natural language processing (NLP) and machine learning to "listen" to clinical encounters and automatically generate structured notes. Companies like Nuance's DAX and Empathia.ai have developed tools that promise to return the physician’s focus to the patient. With an AI scribe, the physician can engage in natural conversation during the visit while the system constructs the clinical note in the background, often requiring only brief review and edits afterward.

This ambient AI documentation represents a profound evolution—not just in technology, but in philosophy. We are moving from the idea of documentation as a burdensome requirement to documentation as a byproduct of care. The early studies are promising. A 2023 Mayo Clinic study showed that AI-generated notes had higher quality and completeness scores than human-typed notes (Patel et al., 2023). More importantly, these tools have the potential to improve physician well-being by reducing after-hours documentation—what we’ve come to call "pajama time."

There are also implications for quality and safety. Structured, legible, and timely documentation improves care coordination and reduces medical errors. Ambient documentation systems can prompt physicians to include missing elements, improving compliance with clinical guidelines. They can also integrate with clinical decision support tools, offering reminders, alerts, or risk scores in real-time. However, these advances are not without concerns. Privacy, accuracy, and trust remain central issues. Physicians must be confident that the AI understands clinical language and nuance. Patients must trust that their conversations are secure and that sensitive information is handled appropriately. Transparency in how these systems function, and clinician oversight, will be critical.

From a patient experience perspective, ambient AI could be a game-changer. For years, patients have lamented the presence of screens between themselves and their doctors. Eye contact, empathy, and attentiveness are core to healing. When technology fades into the background and the physician can be fully present, trust and satisfaction rise. So, what have we gained and what have we lost in this evolution? We've gained legibility, standardization, and data accessibility. We’ve improved billing accuracy, public health reporting, and research capabilities. But we've also lost spontaneity, narrative depth, and to some extent, autonomy.

I am cautiously optimistic. If implemented well, AI scribing may restore the balance—freeing physicians to be the thoughtful, empathetic professionals they aspired to be. The journey from scribbled notes on yellow pads to ambient AI scribes reflects

not just technological progress, but a deeper commitment to aligning documentation with the values of medicine: clarity, compassion, and connection.

References  Sinsky, C., Colligan, L., Li, L., Prgomet, M., Reynolds, S., Goeders, L., ... & Blike, G. (2016). Allocation of Physician Time in Ambulatory Practice: A Time and Motion Study in 4 Specialties. Annals of Internal Medicine, 165(11), 753–760. https://doi.org/10.7326/M16-0961  Patel, R., Alaparthi, G., et al. (2023). Evaluation of an AI Ambient Scribe for Clinical Note Generation in Primary Care. Mayo Clinic Proceedings, 98(2), 135- 142.