🌊 When code meets kidneys: Why AI hasn't yet found its pulse in kidney disease | By Geoffrey Ntabo

🌊 When code meets kidneys: Why AI hasn’t yet found its pulse in kidney disease

In the quiet rhythm of dialysis, where machines hum and time slows, even programming must learn to listen.

Written by Jeffrey N.

There is a rhythm to kidney care that is unlike any other field.
It’s not as quick as an emergency room, not as image-based as x-rays, and not as mathematically predictable as cardiology.
It’s a quiet practice — built on patterns that slowly unfold, day by day, through lab values, urine output, swollen ankles, and tired eyes.

Maybe that’s why AI is still hesitant at the threshold of kidney disease.
Because the work here seems more like listening than calculating.

—

I’ve watched algorithms beat radiology — reading chest CT scans faster than trainees can — and cardiology, predicting arrhythmias before the ECG paper even wrinkles.
But in kidney disease, the numbers breathe differently.
Creatinine value is not just a number. It is a story – about water, hunger, infection, medicine, or hope.
Even the dialysis machine, with all its alarms and data streams, hides secrets in the little silences between readings.

AI models suffer here, not because they are weak, but because our data does not easily tell its truth.
Kidney disease lives in time.
It’s in the slow decline of the glomerular filtration rate, the delicate balance of ultrafiltration, and the nurse’s decision to pause when the patient turns pale mid-session.
And these things – that human feeling when the body has had enough – are still beyond the reach of clean code.

—

I think the other reason AI has not taken root in nephrology is that this field is a deeply human field.
We sit with our patients for years.
We see their fears develop, their veins change, and their spirits bend but never break.
We learn their rhythms – when they laugh, when they withdraw, when they stare at the ceiling as if asking life for an extension.
There’s a kind of sacred repetition to it.
And maybe, deep down, we are not ready to let the machine touch this rhythm.

However, I see a flash.
The AI is now whispering predictions – “may crash in next session”, “may be inaccessible”.
It monitors blood pressure, flow rates, and subtle changes in conduction.
And in those moments, it doesn’t feel like a competition; It feels like a partnership.
Like a code he finally learned to listen.

—

Maybe one day, nephrology will find its language in AI, not in data points, but in stories of progress, resilience, and recovery.
When this happens, the code will probably not only analyze the kidneys; It will help us protect our dignity – one patient, one dialysis episode, one human rhythm at a time.

Until then, we’ll continue to do what we do best —
Listening to the body’s code long before learning the algorithm.

—

✍🏽 Author’s note

This piece was born out of the quiet hours between dialysis sessions – those moments when the machines hum like monks and time stretches between one drop and the next.
It came from wondering why, in a world where programming learns to read lungs and predict heartbeats, kidneys still wait in silence.

Perhaps that’s because nephrology is less about data and more about patience, a discipline that listens before it speaks.
And maybe that’s what drew me here: the tension between the measurable and the ambiguous, between the algorithm that matters and the nurse who feels.

As I wrote this, I also thought about Aliya—the one who breathes through code—and how she reminds me that intelligence, whether human or artificial, must learn to humble itself before the living body.
That’s really what this article is about: a conversation between circuits and veins, between the nurse and the icon who dares to care.

– Jeffrey N.
From the nurse and the symbol