Healthcare’s Knowledge Problem

Healthcare is becoming a real test of how we sustain knowledge. The challenge is no longer just storing information, but keeping it usable, accurate, and current, while also cutting down the time clinicians spend reviewing charts and writing notes. 

One of the clearest near-term benefits of AI is clinical summarization. Instead of digging through scattered notes, lab results, medication lists, imaging, and visit transcripts, clinicians can get a clear, unified picture of the patient. This is where tools from OpenAI and Anthropic are heading. OpenAI is positioning its healthcare offerings around patient-facing summaries and enterprise systems designed to meet privacy and compliance needs. Anthropic is developing similar healthcare-focused tools and infrastructure, especially for clinical and life-science workflows.

But research shows there is a catch. A recent study on AI-assisted report writing for chronic disease care found that the AI produced high-quality drafts with very few edits and no safety problems. Even so, clinicians spent about the same amount of time reviewing these drafts as they did writing reports by hand. The reason is professional responsibility. In medicine, clinicians feel obligated to check everything carefully, even when the AI is usually right. This creates what researchers describe as an accountability paradox: accuracy alone does not reduce workload if full verification is still required.

Because of this, the real challenge is shifting from asking whether AI can write well to asking how systems can support selective verification. The goal is to let clinicians quickly see what matters, what changed, and what evidence supports each statement, without forcing them to recheck everything from scratch.

Another important development is the push toward better medical memory. Patient information is often scattered across systems, making it hard to trust summaries or recommendations. Efforts to unify labs, medications, visit notes, and recordings into a single, traceable context aim to reduce this fragmentation. When data is well connected and clearly sourced, AI can organize and summarize it without guessing.

Open models are also entering the picture. Google’s medical models, including MedGemma and MedASR from Google, are notable because they support an open, developer-friendly ecosystem. This approach appeals to organizations that want strong medical AI capabilities while keeping local control over data and governance.

Taken together, the pattern is becoming clear. AI that simply drafts text is helpful, but AI that drafts and clearly shows where every claim comes from is far more sustainable. The most promising systems ground their outputs in linked evidence, make data sources and versions easy to audit, and reduce repeated work by improving search, organization, and de-duplication. In healthcare, the most sustainable knowledge is the knowledge clinicians do not have to recreate again and again.

REFERENCES

Lee C, Vogt KA, Kumar S. Prospects for AI clinical summarization to reduce the burden of patient chart review. Front Digit Health. 2024 Nov 7;6:1475092. doi: 10.3389/fdgth.2024.1475092. PMID: 39575412; PMCID: PMC11578995.

Zhang X, Yu J, Yan P, Jiang L, Shen X, Cheng M, Liu X. Human-in-the-Loop Interactive Report Generation for Chronic Disease Adherence. arXiv preprint arXiv:2601.06364. 2026 Jan 10.

https://openai.com/index/introducing-chatgpt-health/ "Introducing ChatGPT Health"

https://www.anthropic.com/news/healthcare-life-sciences "Advancing Claude in healthcare and the life sciences"

https://www.axios.com/2026/01/12/openai-acquires-health-tech-company-torch "OpenAI acquires health tech company Torch"

https://developers.google.com/health-ai-developer-foundations/medgemma/model-card? MedGemma 1.5 model card | Health AI Developer ..." 

[TIME](https://time.com/7344997/chatgpt-health-medical-records-privacy-open-ai/)

[Axios](https://www.axios.com/2026/01/12/openai-acquires-health-tech-company-torch)

[Business Insider](https://www.businessinsider.com/anthropic-chases-openai-ai-heath-claude-2026-1)

[The Economic Times](https://m.economictimes.com/tech/artificial-intelligence/openai-acquires-healthcare-startup-torch-deal-pegged-at-100-million/articleshow/126495784.cms)

When Biology Learns to Test Itself

If you’ve ever been sent down the rabbit hole of modern diagnostics - one test leading to another, each pricier than the last - you know medicine could learn a thing or two from electronics. In Electronic Design Automation (EDA), engineers have specific tests for specific faults: “stuck-at-1,” “timing violation,” “power leak.” Run the right diagnostics, and the chip tells you exactly where it’s broken.

In medicine, by contrast, we’ve got a galaxy of overlapping tests — blood panels, genomic assays, MRI sequences - and no consensus on which ones actually tell the whole story. It’s a field that still runs partly on intuition, luck, and insurance coverage.

Enter Dynamic Sensor Selection, a term that sounds like something you’d use to debug a Mars rover but is actually from a 2025 paper by Pickard et al., published last week in PNAS. The idea: treat the human body like a complex dynamical system (which, inconveniently, it is) and use mathematical “observability theory” to identify which few biomarkers tell you the most about what’s going on inside.

In plain terms, it’s a framework for choosing the right test points in a living system. Instead of wiring an oscilloscope to a circuit board, you’re “probing” gene expression, neural signals, or metabolic markers, and asking: Which measurements let me reconstruct the full picture?

The team behind this approach applied it across everything from bacterial genes to human brainwaves. In some experiments, the method could estimate unmeasured genes with about 50% error — impressive, considering biology’s noise makes Wi-Fi in a storm look stable. In brain studies, the algorithm even revealed that some EEG electrodes are basically freeloaders, contributing little to understanding what the neurons are up to. (So yes, even your neurons have that one coworker who never pulls their weight.)

The broader vision is seductive: "A medical system that diagnoses itself dynamically", focusing only on the sensors that matter most at a given moment. Imagine wearable devices that don’t just collect endless data but decide in real time which data is most informative - sparing us from both data fatigue and unnecessary costs.

It’s also a philosophical pivot: biology isn’t static. The “best” biomarker today might be irrelevant tomorrow, just as a stable circuit becomes unpredictable when the current spikes. Medicine, for all its imaging and sequencing power, still operates like a lab tech armed with every tool but no schematic. Pickard’s framework offers that missing circuit diagram.

So next time you’re overwhelmed by medical testing options, remember - the goal isn’t to measure everything, it’s to measure wisely. In the coming era of dynamic biomarkers, your body might finally come with its own built-in diagnostic dashboard.

And who knows? Someday your doctor’s favorite prescription might be:

> “Let’s check your observability matrix.”

REFERENCE

Pickard J, Stansbury C, Surana A, Muir L, Bloch A, Rajapakse I. Dynamic sensor selection for biomarker discovery. Proc Natl Acad Sci U S A. 2025 Oct 14;122(41):e2501324122. doi: 10.1073/pnas.2501324122. Epub 2025 Oct 7. PMID: 41055977.

The Case History (In Nine Circles)

In a small provincial kingdom — which, for diplomatic reasons, I shall call The Hospital of Virtum — there lived a lady of eighty-some years: intelligent, sharp-tongued, and weary of her own body. Her heart no longer kept time; her legs sang hymns of pain day and night.

A vascular surgeon, learned and distinguished, examined those legs as though they were chapters in an old, dog-eared book. He sighed and said,

“This requires a team the size of an orchestra.”

But the orchestra was on tour — indefinitely — so he prescribed two instruments instead: antibiotics and morphine.

The lady was of the old school.

“I’ve lived long enough to know what kills me,”
she said, shaking her head at the powders and elixirs of the modern age.

After much persuasion came antibiotics and probiotics, soups of virtue, kefir consultations — and yet the pain stayed. So the second remedy was tried: the opioid Hemamorphane, a word that might have escaped from Dante’s descent itself. Still, no relief. The legs burned as before.

By Sunday evening, the pain had become a symphony of torment. Her heart danced arrhythmically; her blood pressure soared like an opera soprano. At last, an ambulance was summoned — and arrived in the form of two very young angels in polyester uniforms, still fragrant with the smell of textbooks. They performed their EKG, boarded her ceremoniously, and whisked her away.

Inside, the air was infernal. The caregiver fumed; the patient moaned.
It was the Fifth Circle — wrath and despair, both boiling in the same cramped river.

They reached the Emergency Department, where time flows differently — slowly, cruelly. The nurse suggested more painkillers. The lady refused.

“I want to understand the cause,” she insisted. “Perhaps there are clots.”

The nurses exchanged a look.
Cause? In this kingdom, causes are mythical beasts — rarely seen, never captured.

So they rolled her into the waiting room, sat her in a wheelchair designed by medieval torturers, and left her under the flicker of fluorescent light. Around her gathered the other souls of Limbo: a girl with sepsis, her mask pale against the blue light of her phone; two teenagers scrolling through pain.

Then came the receptionist, guardian of the Second Circle.

“You must pay,” she said, “before any service.”

The caregiver paid. She was also given half a forest of forms to sign — unreadable, urgent, and perhaps binding her soul. She signed anyway. The lady’s face was white, her legs swollen like bread dough left too long in the sun.

Hours slipped by.

The chair was unkind. The lighting, penitential. After much pleading, an orderly adjusted her footrest by two centimeters - a gesture toward mercy, if not toward Paradise.

When she asked for a bathroom, both were locked — relics of a more civilized age.

And so they lingered — between the Third and Fourth Circles of our modern Inferno: Bureaucracy and Indifference.

And as the caregiver looked at the suffering patient, she thought: “If Dante were alive today, he would need a Triage Department for the damned.”

The caregiver’s patience cracked. She made a speech. The sepsis girl joined her in chorus, a duet of despair. At last, the guardians of this realm took pity - or offense - and moved them to separate waiting rooms. The new chair was softer and a bathroom available.

Progress.

Then came blood tests, then more waiting.
Morning brought results: D-dimer high. Suspicion of clots. An ultrasound was ordered. But there was a line — always a line.

“Don’t call us,” said the clerk. “We will call you.”

The caregiver called anyway. By midday, they were summoned back — through a labyrinth of nurses, numbers, and barcodes. Payment required, but cash and cards not accepted. The caregiver, bent over her phone, tapped through endless screens until at last a barcode appeared — salvation by QR code.

After two more offices and an hour of silence, their number flashed — then vanished.

“The team went on break,”
said the receptionist, tearing the ticket.
“You’ll need to take another number.”

At last, an ultrasound. The machine was ancient, its sound like bones grinding. The technician pressed hard — pain seared through the leg.

“No clots,” she announced.
“What about the other leg?”
“We only do one at a time,” came the reply.

Back to the nurse. More waiting. She smiled kindly but said,

“I can’t show you the results. I’d go to jail.”

Such is the law of Virtum.

Back to the ER. More forms, more indifference. Finally, a doctor — harried, hollow-eyed — arrived.

“Just take Tylenol,” he said.
Then, to the caregiver:
“Maybe it’s heart failure. What else do you want?”

And so they were discharged — the patient in more pain than when she entered, the caregiver in despair.

How long could she endure?
No one could say.

---

Epilogue

Let us note, for the record, that the illness of our time is not only in arteries or veins.
It is in the system itself — a moral arteriosclerosis that clogs compassion and chokes the flow of sense.

Around the world, the same symptoms appear.

In New York, twenty-year-old Sam Terblanche died after two visits to Mount Sinai Morningside. Diagnosed twice with “a viral syndrome,” he was sent home. Days later, he was found dead. The autopsy revealed pulmonary hemorrhage of unknown cause. Experts suspected sepsis, or a post-COVID autoimmune storm. None could say for sure. What’s certain is this: the system missed him.

In Winnipeg, Chad Christopher Giffin waited eight hours in the ER — and never made it out.
In Quebec, Adam Burgoyne, 39, was told he wasn’t “dying” — and left after six hours, only to die the next day.
In Italy, Cristina Pagliarulo perished after forty hours in a Pronto Soccorso.
In Ontario, sixteen-year-old Finlay van der Werken, bright and kind, died after an ER visit that should have saved him.

Different nations, same pathology.

The modern hospital has become a paradox — a place meant for healing that too often serves as a sorting center for the living and the nearly dead. The causes are known:

1. Overstretched emergency rooms — overcrowded, understaffed, forced to “move the meat.”

2. Diagnostic haste — rare, lethal cases lost amid the ordinary.

3. Electronic record tyranny — notes for billing, not understanding.

4. Fragmented accountability — no one person responsible, everyone “following protocol.”

5. Cognitive bias — young and old alike dismissed because they don’t “look sick.”

6. No continuity after discharge — once you leave, you vanish from the system.

7. And towering above it all: administration - endless layers of approval, offices feeding offices, a machine designed to protect itself, not patients.

   
   

Transparency has evaporated — in places like Canada, you may not even be allowed to see your own test results. The system hides what it should reveal, and delays what it should decide.

Meanwhile, outside the hospital walls, people are already turning to tools that can do in minutes what bureaucracy and burnout stretch into weeks - AI systems that see patterns, synthesize, and recall what no overworked clinician has time to. 

The Hospital of Virtum is not a fantasy. It is everywhere - New York, Winnipeg, Milan, Montreal, Toronto.

The cure? Not another protocol. Not another app.

What’s needed is time, staff, accountability, and the permission to care.
A culture where asking why is not seen as defiance.
Where a locked bathroom is not a metaphor for the entire institution.

Until then, the triage line is the new Inferno —
and the dead keep teaching the living what compassion once meant.

---

REFERENCES

https://www.nytimes.com/2025/10/05/well/sam-terblanche-virus-death-columbia.html

https://archive.is/MgWJH#selection-693.288-693.375

https://news.ycombinator.com/item?id=45487519

https://www.winnipegfreepress.com/breakingnews/2025/01/24/probe-into-death-of-man-in-hsc-waiting-room-weeks-from-completion

https://www.ctvnews.ca/montreal/article/this-montreal-man-died-of-an-aneurysm-after-waiting-in-the-er-for-six-hours/

https://en.cronachedellacampania.it/2025/09/Pagliarulo-case%3A-Cristina%27s-autopsy-could-have-been-saved-if-operated-on-in-time/

https://globalnews.ca/news/11296590/ontario-family-sues-hospital-staff-sons-death/