Doctors are taught a simple rule early in their training: When you hear hoofbeats, think horses, not zebras. In other words, the most common explanation is usually the right one. But for the roughly 300 million people worldwide living with a rare disease, that rule can turn into a trap. Their symptoms often look like something ordinary – until years later, when someone finally realizes it was a zebra all along.
The problem isn’t the doctors; it’s the data. Every symptom, lab result, scan, or ER visit lives in a different corner of the healthcare system and nobody sees the full picture across time. When the clues are scattered, even the sharpest clinician might miss patterns hiding in plain sight.
When rare diseases hide behind familiar faces
Take acute intermittent porphyria (AIP), for example. Many patients live for years with unexplained pain, fatigue, and weakness – often told they have fibromyalgia, chronic fatigue, or anxiety. On average, it takes 10-15 years from first symptoms to the correct diagnosis. Only much later does genetic or biochemical testing reveal the true cause: a rare disorder of heme metabolism that triggers painful attacks when certain medications, stress, or hormones disrupt the balance.
Or consider Fabry disease — another master of disguise. Patients may spend over a decade bouncing between specialists from rheumatologists, to neurologists, to cardiologists because their symptoms look like nerve pain, autoimmune disease, or even multiple sclerosis. Studies show the average delay to diagnosis is about 14 years in men and 16 years in women. Only when someone connects the dots (skin changes + kidney problems + subtle heart thickening) does the real story emerge.
Then there’s transthyretin amyloidosis (ATTR), which can quietly damage the heart over years. One of its earliest clues? Carpal tunnel syndrome – that tingling wrist pain that sends you to an orthopedist. Most people get surgery and move on. But in hindsight, that was the first breadcrumb. Research suggests it takes about 6-8 years on average before ATTR is correctly identified, long after the first symptoms appear.
In each case, the evidence was there – it was just scattered across years and specialties.
Why it’s so easy to miss
These aren’t isolated failures. They reflect a system that’s built around encounters, not continuity. Rare diseases are not invisible, but the healthcare system often sees them through a keyhole. Here’s how it works:
- Each visit is a snapshot. A lab result here, a symptom note there, maybe an MRI five years later – but never viewed together.
- Different data lives in different systems. Your labs might be at one hospital, your imaging at another, and your genetic report in a lab portal no one checks twice.
- The “common first” mindset. Because rare diseases are, by definition, rare, the odds seem stacked against them, until you realize those odds reset every time a new doctor starts from scratch.
AI: one common thread to stitch it all together
Imagine a longitudinal, multi-modal patient record – one that brings together every clue across the years: blood tests, imaging reports, pathology results, doctor’s notes, even data from wearable devices. AI is the key that unlocks our ability to stitch these pieces together and start making diagnoses that are accurate, not just easy.
Instead of thousands of disconnected dots, you’d get a timeline that tells a story. A patient with years of abdominal pain, dark urine, and episodes triggered by stress or medication might instantly flag an AIP pattern. A string of orthopedic notes mentioning carpal tunnel in both wrists, followed years later by rising cardiac biomarkers, could trigger a prompt for ATTR screening. That’s the power of seeing the whole patient.
Here’s what that AI-powered vision looks like in practice:
- Connect the timeline. Pull together labs, imaging, procedures, and doctor’s notes into one unified patient journey.
- Teach the system what to look for. Build a digital “fingerprint” library of rare diseases – combinations of features that tend to cluster together over time.
- Blend rules with learning. Some patterns come straight from medical literature (“bilateral carpal tunnel before 60 plus thickened heart walls – think ATTR”). Others can be learned automatically by algorithms scanning thousands of historical cases.
- Keep the human in the loop. Doctors still make the final call, but now they’re alerted to patients who deserve a second look.
Every year shaved off the diagnostic odyssey means fewer irreversible complications, fewer misdiagnoses, and fewer clinicians and families left wondering what they missed. A unified, longitudinal approach doesn’t just speed up diagnosis — it changes, and in some cases saves, lives. It gives clinicians the full canvas instead of a handful of puzzle pieces. And if it works as intended, it helps more patients hear their doctor say: “We finally know what this is and we can treat it.”
About David Talby
David Talby, Ph.D., MBA, is the CEO of John Snow Labs. He has spent his career making AI, big data, and data science solve real-world problems in healthcare, life science, and related fields. John Snow Labs is an award-winning AI for healthcare company providing state-of-the-art software, models, and data that power the world’s leading pharmaceuticals, academic medical centers, and health technology companies. Creator and host of The Applied AI Summit, the company is committed to further educating and advancing the global healthcare and AI communities.
