Doctors Who Code Blog Post
Why I Finally Cancelled My UpToDate Subscription After 12 Years
I just did something that would have been unthinkable five years ago: I cancelled my UpToDate subscription.
For those keeping score, I’ve been an UpToDate subscriber for over twelve years. It was the clinical reference tool that lived in a perpetually open browser tab, the resource I’d consult multiple times daily for everything from rare complications to drug interactions. It was expensive, sure, but it was worth it.
Was.
The Slow Decline
The truth is, I haven’t been getting my money’s worth for a while now. Each year, the subscription cost crept up—another $50 here, another $75 there. Meanwhile, my usage plummeted. Not because the content got worse, but because better alternatives emerged.
The ChatGPT Watershed
The inflection point came about 2.5 years ago with ChatGPT. Suddenly, I could have a conversation with medical literature. Instead of clicking through hierarchical menus and scanning walls of text, I could ask specific questions and get synthesized answers. Yes, early GPT models hallucinated and made errors—but the paradigm shift was undeniable.
Traditional medical databases were built for the search era. But we’ve moved beyond search.
The Final Straw: Medical-Specific LLMs
What sealed the deal for me were purpose-built medical AI tools—particularly Open Evidence and EvidenceMD. These aren’t general-purpose chatbots retrofitted for medicine; they’re built from the ground up for clinical decision support with proper citations, evidence grading, and study synthesis.
These tools do what UpToDate does, but with:
- Natural language queries instead of keyword search
- Dynamic synthesis instead of static articles
- Real-time literature integration instead of periodic updates
- Conversational follow-ups instead of hierarchical navigation
UpToDate’s interface, once revolutionary, now feels like browsing a medical textbook in a world where we have AI research assistants.
What This Means for Physician-Developers
For those of us building at the intersection of medicine and technology, this transition isn’t just about personal tool choices—it’s a preview of what’s coming for all of clinical practice.
We’re watching the early stages of a fundamental platform shift. The database era is giving way to the AI era, and the tools that dominated 2010s medicine won’t necessarily dominate the 2030s.
The question isn’t whether UpToDate will survive (it probably will, through institutional subscriptions and organizational inertia). The question is whether the next generation of clinicians will even understand why we paid hundreds of dollars annually for what amounts to a searchable medical encyclopedia.
Am I Going to Miss It?
Honestly? No.
That’s the most telling part. I thought I’d feel more nostalgic about closing this chapter. But the reality is that better tools exist now, and they’re only getting better. Holding onto UpToDate in 2025 would be like insisting on a pager when everyone else has a smartphone.
The medical software that once felt indispensable now feels… antiquated. And that’s okay. That’s progress.
What Are You Using?
I’m curious: how many physician-developers out there are still using traditional medical databases as their primary reference? Have you made the switch to medical LLMs? What tools are in your clinical stack these days?
Let’s talk about what the post-database era of clinical practice actually looks like.
C. Onyeije, MD