Aortic dissection is rare. It’s terrifying. It masquerades as back pain, abdominal agony, or a standard heart attack.
Senator Lindsey Graham’s sudden death from this condition brought a hidden fracture in American healthcare into sharp relief. Most people focus on the disease. I focus on time.
For decades, we’ve framed medical technology around a single question. Can an algorithm spot a tumor faster than a human radiologist? Can it read a chest X-ray without blinking?
Yes. It can. But that’s the wrong question.
I’ve spent over 20 years watching health systems deploy technology. I’ve talked to burnt-out ER docs and residents running on fumes. They never want to miss a diagnosis. The system isn’t failing because people are negligent. It fails because it is human.
And humans forget.
How clinical communication breakdowns actually kill
Here is how it happens. It isn’t malicious. It’s just busy.
A patient goes in for kidney stones. The CT scanner catches a small spot on a lung. Irrelevant to the kidney stones. The radiologist flags it. Standard protocol. Follow up.
The urologist fixes the stone. The patient goes home.
Now life resumes. School runs. Aging parents need rides. Work meetings. The follow-up for that lung nodule? It’s easy to lose. Especially when you’re exhausted from surgery and pain meds.
Six months later, the patient needs another scan. Maybe a car accident. Maybe just pain. The radiologist sees the spot. It’s grown. Again. Flagged. Recommended action noted in a chart that nobody checks unless something explodes.
The emergency team focuses on the immediate trauma. The lung spot? It sinks into the digital ocean of “past medical history.”
Years pass.
The pain gets worse. Another scan. This time, it’s not a spot. It’s advanced cancer.
Nobody ignored this patient. The radiologist was correct. The ER doc treated the acute issue appropriately. The system did exactly what it was designed to do. Process the acute event.
The failure wasn’t diagnosis. The failure was memory.
Dr. Andrew Ibrahim from Viz.ai told the House Energy and Commerce Committee something that stuck with me. He spoke about his father’s stroke. Because Dr. Ibrahim was there. Because he knew what to say over the phone before the ambulance even arrived. The outcome was good.
Then he said this:
Not everyone has a medically trained son.
We need systems that work even when you don’t have a specialist on speed dial. That is the core issue with healthcare AI implementation today. It’s not about replacing the doctor. It’s about replacing the silence.
What proactive AI monitoring looks like
Let’s rewrite the lung spot scenario. Same patient. Same kidney stones. Same incidental finding on the lung.
The AI system picks it up. It doesn’t just print it in a report that sits on a server.
It generates an evidence-based follow-up path.
If the patient misses the next scheduled imaging? The system notices.
It triggers an alert in the patient portal.
No response? It flags the primary care provider.
Still no contact? An administrative assistant calls. Schedules an appointment.
The patient forgets. The system doesn’t.
The follow-up scan happens early. The nodule is found. It’s suspicious, but manageable. Treatment is straightforward.
No chemotherapy. No radiation battles. No missing school runs.
The patient drives their kids to soccer. They help their mom. They live their life.
This is what operationalizing preventive health means. It isn’t sexy. It isn’t a robotic arm doing surgery. It is a boring, relentless, automated net that catches patients when they fall through the cracks of their own busy lives.
Andrew Menard, a healthcare executive who beat cancer, gets it. He automates his life. Why? Not because he likes robots.
Listen to your body. Trust your instincts. Push.
But he knows instincts fail when you’re tired. Technology handles the mundane tracking so humans can handle the care.
Why time and context matter more than accuracy
We ask if AI can diagnose better. We should ask if it can communicate better.
Lindsey Graham’s death wasn’t about a misread test. It was likely a cascade of timing, recognition, and resource mobilization. In an ideal world, a system would recognize the risk factors, flag the urgency, and mobilize specialists before the patient even hit the floor.
Currently? We rely on luck.
Luck that a specific doctor looks up. Luck that a specific nurse asks a specific question. Luck that you have a family member who knows medical terminology.
Healthcare shouldn’t require luck.
The future isn’t machines replacing physicians. That’s a sci-fi distraction. The future is building rails. Iron tracks that ensure information moves from point A to point B without getting buried under the weight of daily human error.
Clinicians shift. They go home. They sleep. They have families too.
When they clock out, the care shouldn’t stop. It should just change hands. From human to machine. And back to human when the patient returns.
The goal isn’t perfect accuracy. It’s resilience.
It’s making sure that the ordinary moments—the school bus, the parent’s birthday dinner, the quiet morning coffee—aren’t stolen by a finding that was noted, forgotten, and then fatal.
We build AI to remember what we forget. So we can go back to being human.
