Your EHR Sucks
I won’t sugarcoat it: your EHR sucks. And if you’re a clinician reading this, you probably don’t need me to tell you that. It’s the system you’re stuck with, the one that slows you down, buries you in unnecessary clicks (fight me if you want about clicks not being a good measure of an experience, but first listen to every clinician speak in those terms), and leaves you wondering why we put people on the moon decades ago but still can’t design an EHR that works.
It’s not just frustrating—it’s dangerous. Poorly designed EHRs contribute to cognitive overload, user error, and burnout. A study in JAMA Network Open found that bad EHR interfaces can impair decision-making, leading to errors that harm patients. Clinicians are spending nearly two hours on documentation for every hour of patient care, according to the Annals of Internal Medicine. That’s not progress; that’s malpractice by design.
So let’s call it what it is: EHRs aren’t serving clinicians. They’re failing them. And, by extension, they’re failing patients.
The real problem: EHRs weren’t designed for you
How did we get here? The truth is, EHRs weren’t designed with clinicians in mind. They were built to meet billing requirements, regulatory mandates, and administrative needs. That’s why so many of them feel like glorified filing cabinets (and usability/accessibility dumpster fires!) rather than tools to support actual patient care.
The result is a system where you’re forced to navigate irrelevant tabs, sift through cluttered data, and deal with non-specific alerts that scream “click fatigue.” And let’s not forget the specialty mismatch—using a generalist EHR for a nuanced specialty is like trying to perform heart surgery with a Swiss Army knife. Sure, it technically works, but would you want to be the patient?
What’s missing: Context and design leadership
Here’s the thing: we have the technology to make EHRs better. AI, machine learning, and natural language processing are advancing at breakneck speed, but what’s missing is a shift in mentality. The problem isn’t just technical; it’s a failure of design leadership.
EHRs don’t need more features; they need better design. They need to understand the context of the user—what kind of clinician they are, the patient they’re treating, and the task at hand. Instead of dumping every piece of data on the screen, they need to dynamically adapt, showing only what’s relevant and receding the rest into the background.
This approach, sometimes called “sentient design,” is about creating systems that feel intelligent and responsive—systems that work for the clinician rather than forcing the clinician to work for them.
What would a better EHR look like?
Imagine an EHR that actually supports you. Here’s what that could mean:
Context-Aware Data: A cardiologist doesn’t need to see the same information as a pediatrician. A context-aware EHR would prioritize the right data for the right clinician at the right time, reducing cognitive overload and cutting down on wasted clicks.
Proactive Assistance: Instead of making you dig through menus, the EHR would surface the labs, imaging, or treatment protocols you need based on your workflow. It would act more like a clinical assistant than a filing cabinet.
Streamlined Alerts: No more meaningless pop-ups. Alerts would be relevant, timely, and actionable, helping you focus on what matters without drowning you in noise.
Burnout Reduction: By aligning with your workflow and cutting unnecessary steps, a better EHR would give you more time for actual patient care—and maybe even your life outside of work.
Why aren’t we there yet?
The technology exists, so why hasn’t this revolution happened? The answer lies in the status quo. Major EHR vendors dominate the market with bloated, one-size-fits-all systems (Epic, Cerner, Meditech, Allscripts; I see you). There’s little incentive for innovation when these systems are entrenched in large health systems and protected by contracts that prioritize compliance over usability.
But here’s the kicker: healthcare is reaching a breaking point. Burnout rates are climbing, and the industry is losing talented clinicians because they’re tired of working in systems that seem designed to frustrate rather than facilitate care. Something has to give.
The case for “sentient” design in EHRs
Enter sentient design—a philosophy of building systems that adapt to the user, not the other way around. At its core, sentient design is about context. It’s about using AI to deliver information that’s relevant to the moment, tailoring the interface to the clinician’s specialty, and supporting decision-making in real time.
It’s also about empathy. A sentient EHR isn’t just smart; it’s considerate. It understands that clinicians are humans, not data-entry machines, and that the goal isn’t just to log information but to improve outcomes—for both the patient and the provider.
A call to action
Here’s the bottom line: your EHR sucks because it wasn’t built for you. But it doesn’t have to stay that way. Healthcare leaders, technologists, and designers need to stop tinkering around the edges and start reimagining what EHRs could be.
It’s time to demand better. Not incremental updates, but a fundamental redesign. One that puts clinicians at the center, uses context-aware technology to reduce friction, and embraces the principles of sentient design to create tools that feel intuitive and empowering.
The tools are here. The need is clear. The question is, who’s ready to lead the charge?
Here are some folks doing cool things with EHRs today
Notable Health: These guys are using AI to automate administrative tasks, like documentation and data entry. they use voice recognition and machine learning to streamline workflows.
Canvas Medical: Canvas integrates AI to provide clinical decision support and predictive analytics. Their system can suggest potential diagnoses and treatment plans.
DrChrono: Maybe the closest to sentient design principles I’ve seen out there for EHR, DrChrono has a mobile EHR with things like customizable forms and real-time patient data access. They focus a lot on adaptability and user-centric interfaces.