Finding Her Story Again

Lesson: When used with intention, AI can help us connect a fragmented system.

She was 82, wheeled into our nursing home late on a rainy Thursday afternoon. The transfer summary was sparse — two typed paragraphs and a med list. The admitting note from the sending facility was generic. No clear goals of care. No history of what led her here. Just another transition in a long, fragmented journey.

Her name was Mrs. M. And at first glance, she looked tired. Thin. Her hands were tremulous. She wore a faint look of resignation — the kind of expression patients carry after being passed between hospitals, rehabs, and nursing homes for too long.

There was a moment — one that’s become all too familiar — where I felt a familiar frustration rise: How am I supposed to make good decisions for her if I can’t even find her story?

In the past, I would’ve spent the next hour clicking through disconnected notes, scouring hospital portals, calling outside facilities, waiting on faxed records, and trying to reverse-engineer her past through guesswork.

But this time was different.

This time, I had help.

The Tools That Listened

I activated the ambient AI tool as I sat down with Mrs. M. It recorded our conversation, identifying my questions, her responses, the physical exam, the plan, and even the nuanced tone of our interaction. While I spoke with her — fully, attentively — it quietly documented in the background. Not word-for-word, but meaning-for-meaning.

Then I turned to the large language model integrated with our health information exchange. With a single request, it retrieved and summarized two years of care across four hospital admissions, three skilled nursing facility stays, and countless outpatient visits.

Instead of 30 PDFs and disconnected portal logins, I had a timeline. A story.

What I Learned

She had suffered a pelvic fracture two years prior after a fall on ice. She’d been hospitalized three times for infections — one of which resulted in sepsis and a prolonged ICU stay. She had completed rehab twice but regressed both times after being discharged too soon. She had been readmitted for a TIA. Her dementia had worsened after each hospitalization. She had undergone two rounds of chemotherapy for lymphoma but had missed follow-up due to transportation barriers.

None of this was in the transfer note.

But it was all there — buried in scanned progress notes, discharge summaries, and prior H&Ps across the region.

The AI summarized it in three paragraphs, followed by links to every source.

The Encounter That Mattered

Because I had the context, I didn’t need to spend our encounter checking boxes. I spent it talking with her.

I asked her what she remembered about her illness. She told me about the fall that started it all. How she had once been active, independent, cooking every night, and walking to church. How she started to lose pieces of herself after each hospital stay. How she missed her garden.

She didn’t remember every diagnosis. But she remembered the grief of losing her old life.

Because I knew her story, I didn’t ask irrelevant questions. I didn’t repeat labs. I didn’t put her through another CT scan “just to be sure.” I talked to her about what she wanted next. What quality of life looked like now. Whether she felt safe. And what she feared most.

And I had time to do it.

Because I wasn’t buried in the chart.

I wasn’t scrolling through PDFs.

I wasn’t recreating work that had already been done.

I was present.

What It Gave Back

Documentation fatigue is real. It creeps in not only as exhaustion, but as disconnection. We start to document to satisfy billing, not to understand. To survive, not to remember.

But this encounter reminded me that when used intentionally, AI can give us something back: time.

Time to be a doctor again. Time to listen. Time to think.

Ambient AI wrote the bulk of my note. The LLM pulled the records and organized them. All I did was review, refine, and re-engage.

The patient may not have realized that technology was working in the background. But she knew that I was fully there — with her. And that made all the difference.

Reflection

We often speak of AI as a threat to the human elements of care. And in some cases, it can be. But if we design it right — if we guide its use — AI can be a bridge back to presence.

It can free us from the tyranny of the inbox. From the endless clicking. From the documentation treadmill that steals our evenings and shortens our patience.

It can help us reclaim what we never should have lost: time with our patients. Context for our decisions. And the quiet satisfaction of knowing someone’s story before we try to change it.

Closing Thought

Mrs. M reminded me that people don’t arrive as “new patients.” They arrive carrying history — some of it spoken, much of it buried in systems not designed to be read by humans.

When the record is scattered, we lose the thread. But when AI helps us find it again, we rediscover not only the patient — but ourselves.

Because when you see the story, you see the person.

And when you see the person, you know what to do.

And that — always — lives between heartbeats.

Excerpted from Between Heartbeats and Algorithms: Reclaiming What Matters in Healthcare (American Association for Physician Leadership, 2025).