Two hours on a Friday
What four months of chemo taught me about real human connection in healthcare
It was Friday January 6, 2023 and I found myself sitting in the waiting room of a hospital. The oncology department. My oncologist was running an hour late. Two days earlier I had learned that my cancer was back. It had metastasised. A 5cm malignant mass was now spreading across the lymph nodes in my belly. I would probably need chemotherapy. It was almost 5pm.
I was watching the other people in the waiting room. So this is what cancer patients looked like. There was nothing special to see. Just regular people. Some looked just as nervous as I was. Others looked bored, probably because they had been in that waiting room countless times by now.
She was finally ready with the previous patient and called me in. She could have easily kept the conversation brief, explaining me the essentials of the treatment plan. It was the end of her week. I’m sure she had plans. Instead, she sat down and actually talked to me for two hours.
Two hours.
She walked me through every detail. What the chemo would feel like. What the side effects would be. What weeks would be hard and what weeks would be easier. What I could eat, when I would lose my hair, how my energy would shift. What it would mean for the people around me. She answered questions I didn’t even know I had yet. It was the first time my conversation with a doctor did not feel rushed. She wasn’t rushing. She was just there, making space for me. On a Friday at 6 PM.
I left that room calm about what was ahead. I was in good hands.
And that exact feeling carried me through four months of chemotherapy. The treatment wasn’t pleasant but I was never really anxious, worried, or stressed. Because that single two hour conversation had given me a feeling of safety and trust in my doctor. The kind you can only build human to human, face to face, in real time.
What it means for my work in digital health
I run a digital health company. We build software that helps hospitals monitor patients remotely. I spend my days thinking about how technology can make healthcare better, more efficient, more scalable and more accessible. I believe in it deeply.
And yet, when I think back to that Friday, I don’t think about technology. I think about a person who chose to make time for me. Who understood that the most important thing she could do in that moment wasn’t to prescribe the right drugs, it was to make me feel safe enough to handle what was coming.
That can’t be automated. I don’t want it to be automated. No chatbot, no matter how empathetic its language model, no matter how well-trained on oncology literature, could have done what she did. Because it wasn’t just information she gave me. It was her time and attention.
It reminds of a research study I published ten years ago (Fox et al., 2015). Me and my colleagues at Stanford compared the effects of agents versus avatars: virtual representations controlled by a computer or a human. We did a wide sweep of the scientific literature and collected all studies that has studied these effects. We especially looked for studies that just manipulated the perceived agency: so actual virtual representations where always controlled by a computer, but half of participants were told that they were interacting with a human. We contacted the individual researchers to ask for access their original data, followed by a detailed statistical analysis of those data.
It turns out that believing a digital representation is controlled by a human makes a significant difference. When we believe we are interacting with another person, the interaction has a much more profound impact on us than when we believe we are interacting with an AI agent. This shows me that AI cannot replace everything.
The space I hope AI creates
And that’s why this effect is especially interesting for me these days. I don’t want AI to replace that conversation, even though it probably could. But I do want AI to protect it.
Because the thing is, my oncologist spending two hours with me on a Friday evening is rare. Most doctors or nurses don’t have that time. Not because they don’t care, but because their days are crushed by administrative work and documentation. Or patients that are actually doing fine. Every minute spent on paperwork is a minute not spent sitting with a scared patient.
This is why I am excited about AI in healthcare. Not AI diagnosing rare diseases from a photo or replacing radiologists. But rather the boring stuff. AI that drafts a discharge summary so a nurse gets 20 minutes back. AI that pre-fills forms so a doctor doesn’t spend their evening doing admin. AI that handles the triage questions so that when a patient truly needs a human, the human is actually available.
I want AI to automate everything that doesn’t need a real person, so that nurses and doctors have the time and energy for the patients that actually need them the most.
A terrible trade
I know this sounds simple, and maybe even obvious. But I think we’re at risk of getting it backwards. There’s so much energy right now going into AI that replaces human interaction in healthcare. Chatbots as the first point of contact, or AI therapy, or automated follow-up calls. And I understand the economics because healthcare is expensive, there aren’t enough doctors and nurses, something has to change.
But we should be very careful about what we choose to give up. The two hours my oncologist spent with me that wasn’t inefficiency. That was an important part of the treatment.
I sometimes worry that we’re building a healthcare system where everything is optimized except the thing that matters most. Where every metric improves — response times, throughput, cost per patient — while the space for real human connection disappears.
That would be a terrible trade.
What I want us to build toward
I’m fine now, by the way. Three years out, clean scans.
But I still think about that Friday. Not about the diagnosis. I think about my oncologist. About the fact that she was there. About the way she made something hard feel bearable. Not through medication or protocols. Through presence.
That’s what I want to protect. That’s what I hope we’re building toward. Not AI that replaces the doctor. AI that gives the doctor or nurse back a Friday evening.
Fox, J., Ahn, S. J., Janssen, J. H., Yeykelis, L., Segovia, K. Y., & Bailenson, J. N. (2015). Avatars versus agents: a meta-analysis quantifying the effect of agency on social influence. Human–Computer Interaction, 30(5), 401-432. Download PDF
Brilliant - I write about similar in general practice https://amcunningham72.substack.com/p/redesigning-general-practice-by-accident?utm_source=share&utm_medium=android&r=6f6pv
Powerfull @Joris Janssen! Fully aligns with what I wrote last week: we have to liberate people from the work that doesn’t need humans. And thus creating space for them to be meaningfull. As did your oncologist.