If ambient scribes are the version of clinical AI I am most cautious about, patient-facing visualization is the version I am most excited about.
Here is the asymmetry. A scribe processes the patient. A teaching tool processes the disease, and gives it back to the patient in a form they can actually understand.
I take care of patients with congestive heart failure. The mortality of CHF, properly staged, is worse than most cancers. And yet the explanation a typical patient gets when they leave the hospital is some version of “your heart is weak, take these pills and watch your salt.” This is not because doctors are lazy. It is because there is no time, and because the available tools — a printed handout, a 90-second YouTube link, a poorly drawn whiteboard sketch — are not good enough to do the work the conversation actually requires.
What I want for my patients is closer to what I have in my head. When I explain heart failure, I do not see a list of medications. I see a pump that is being asked to push too much volume against too much resistance, which causes the chamber to stretch, which causes the muscle to thin and weaken, which causes the kidneys to retain more salt because they think the body is dehydrated, which adds more volume, which stretches the chamber further. It is a feedback loop that lives in three dimensions, and it is what I am trying to interrupt with every dose of furosemide and every conversation about salt.
I cannot draw that on a napkin in the time I have. But a good 3D animation — anatomically correct, accurately scaled, narrated in the patient’s language — can. Studios like Nucleus Medical Media have spent years building exactly this kind of content, validated and used by hospitals across the country for patient education. The animations are not gimmicks; they are pedagogy that respects how human beings actually learn.
The teaching lab on this site is a curated collection of those visualizations, organized around the conditions and decisions I most often need to explain at the bedside: how GLP-1 medications change the body, what end-of-life care after a devastating stroke actually looks like, why muscle disappears in hospitalized patients and how it is rebuilt, how the immune system remembers a pathogen for a lifetime, what vaping is doing to a generation of lungs, the trajectory of alcohol-associated liver disease, and end-of-life care for chronic disease. Each one is paired with a short note from me — what I want the patient to take away, what the visualization gets right, and where the analogy breaks down.
This is what I think AI’s near-term role in medicine should look like. Not standing between the doctor and the patient, but standing behind both of them, holding up a clearer picture of the body so the conversation can be a real one.
— Jeremy Tabernero, MD · More notes · RSS · Get in touch