When Efficiency Trumped Narrative

Here’s an odd thing about modern medicine: We have perfected the art of medical storytelling for over 4,000 years but have spent the last few decades systematically destroying it. Doctors have always been storytellers first, from ancient Egyptian scrolls to medieval Islamic physicians carefully documenting case histories. Until we decided efficiency mattered more than narrative.

Today, we stand at another pivotal moment in medical documentation. Artificial Intelligence promises to transform how we record patient stories, potentially creating what I call the ‘(a)iRecord‘ – a new kind of medical record that’s AI-enhanced, automated, and possibly more removed from human experience than ever before. To understand this coming change, we must first look at how we’ve already transformed medical storytelling.

You can trace the beginning of the modern approach to medical documentation to a well-meaning doctor named Lawrence Weed. In the mid-20th century, he invented the problem-oriented medical record (POMR). His idea was simple: organize messy clinical information into neat, actionable problems. Of course, we know it looks in practice- rich human stories turned into standardized checklists. It’s like creating a detailed map of a neighborhood that somehow fails to capture what it feels like to live there.

The Cost of Oversimplification

The irony is that this oversimplification actually makes medicine less effective, especially when dealing with serious illness. When my patient with advanced renal cancer tells me about her husband and three teenage daughters, and her voice cracks at the thought of them living alone after her death – well, good luck capturing that in a checkbox. These elements aren’t merely emotional flourishes – they become part of the essential tapestry of knowing our patients as people, allowing us to help them make decisions that genuinely align with their lives and values.

Narrative Medicine

Rita Charon, who pioneered narrative medicine, puts it perfectly:

Most patients want us to know not just what’s the matter with them, but what matters to them.

Yet our current system seems designed to ignore precisely this distinction.

The AI Promise

Now, AI promises to make all this “better.” A colleague recently showed me his new AI transcription tool, beaming with possibility.

I can finally look at my patients instead of my screen! he exclaimed.

The research shows that LLMs can significantly impact medical note-making in cancer care by enhancing documentation efficiency and accuracy. Here is one example. Multiple tools are already on the market—from Microsoft/Epic’s DAX to Nuance Dragon Ambient to Abridge—each promising to transform how we document care.

Digital Shadows

Yet, something nags. Abraham Verghese cautioned us that electronic records create an ‘iPatient’—a digital facsimile that frequently eclipses the actual person in the bed. Now, we may face something even more seductive: the (a)iRecord. It’s flawlessly efficient, completely accurate, and somehow feels less real than the messy notes we used to write.

Reimagining AI’s Role

But here’s where it gets interesting: What if AI could do more than transcribe? What if it could help us become better listeners? During patient encounters, I often discover crucial information about values and preferences after the visit – reviewing notes and thinking, “I wish I had picked up on that earlier.” Imagine an AI system analyzing conversations in real-time, not just transcribing but truly listening – flagging statements that reveal key values: “I want to be there for my daughter’s wedding” or “I’m tired of going to the hospital.” In oncology, where aligning treatment with patient goals is crucial, this could transform care. An AI could track evolving priorities across multiple visits, ensuring our treatments truly match what matters most to patients. These summaries could also be shared with the patient, reinforcing the discussion.

Finding the Balance

The challenge isn’t protecting medicine’s art from technological encroachment. It’s remembering that medicine has always been a dance between tools and touch, between science and story. This dance becomes even more complex as we grapple with ethical considerations, such as ensuring AI systems respect patient privacy, address algorithmic bias, and complement rather than replace human judgment.

Healthcare professionals will need new skills to orchestrate this dance effectively. Beyond learning to use AI tools, we must develop the wisdom to know when to rely on technology and when to trust our human instincts. Perhaps most importantly, we need to ensure AI serves not just clinicians but patients themselves—giving them greater access to their health information and a stronger voice in their care decisions.

From papyrus to keyboards to AI, the question has never been whether to use new tools but how to make them serve the story rather than vice versa. Ultimately, medicine isn’t just about solving problems—it’s about understanding the human beings who have them. While AI can’t tell that story on its own, it might help us become what we’ve always aspired to be: better listeners, better observers, and, ultimately, better clinicians.