Clinical Judgment Is Becoming More Valuable, Not Less
I spend a lot of time in the EHR just looking for information. Before I can write a note or make a decision, I’m digging through prior notes, lab results, imaging reports, and messages from other team members.
I’ve used Epic at different health systems, and much of the interface feels familiar wherever I go. But the information I need upfront changes depending on whether I’m seeing someone in pulmonology clinic or caring for them in the ICU. I still spend too much time searching for what matters to the patient in front of me.
A LinkedIn post recently got me thinking about how the EHR interface could better reflect those specialty-specific needs. I can’t remember who posted it, but if it was you, thank you for inspiring this discussion.
I think more of our clinical work could move into workspaces built around how we practice, with the EHR continuing to hold the official record. These platforms would still need to integrate with the EHR, but they could become our “home base” for reviewing information and preparing notes. I’ve already experienced a version of this at the VA.
In this article, I’ll share my experience with a platform layered on top of the VA’s EHR, explain how our information needs change across specialties, and explore whether specialty-specific workspaces could become our new clinical home base.
My Experience at the VA
During residency, I worked at the VA and used its ancient EHR, CPRS. I’ve discussed my frustrations with it in Why the VA’s EHR Is Still Stuck in the Stone Age, but for much of my time there, I reviewed patient information in a third-party platform called Avicenna.
Avicenna pulled information from CPRS into an interface that made it easier to review notes, labs, imaging, and medications. I could also draft my progress notes there, then copy and paste them into CPRS when I was done. CPRS still held the official record, but much of my chart review and note writing happened in Avicenna.
At the time, I didn’t think much about that arrangement. I just knew it made my day easier.
During my last year of residency, the VA hospital where I worked stopped using Avicenna. I was back to working directly in CPRS, and I noticed how much more effort it took to review the same information.
That experience has stuck with me. As AI expands what developers can build, I think there’s an opportunity to create better clinical workspaces around existing EHRs. Avicenna had already shown me how much a better interface could help, even while the underlying EHR stayed the same.
Different Specialties Have Different Needs
The information I want upfront has changed as I’ve moved through different clinical settings. Some needs overlap, but what I’m looking for—and how I want it organized—depends on the work I’m doing.
Primary care: In residency clinic, I wanted a summary of our last visit and any visits or hospitalizations since then, along with medications, immunizations, and cancer screening status. I also wanted relevant labs—A1c, urine albumin-to-creatinine ratio, lipids, CBC, and BMP—easy to find before the visit.
Hospital medicine: On the inpatient side, I wanted vital sign and lab trends, recent imaging, and a summary of the hospital course and consultants’ recommendations. Active medications and one-time doses, like Lasix or pain medication, mattered too, along with progress toward discharge and any remaining barriers.
Pulmonary medicine: In pulmonology clinic, I want the last visit summary, current pulmonary medications, and pulmonary treatments we’ve already tried. I also need the latest pulmonary function tests and chest imaging. In my experience, PFTs are often buried in separate PDFs, making it difficult to quickly find results and compare them over time.
Critical care: In the ICU, I want vital sign and lab trends alongside sedation, vasopressor, and analgesic infusions; ventilator settings and blood gases; fluid balance; and ECMO settings when applicable. I also need antibiotic history and prior culture susceptibilities, a day-by-day clinical summary, goals-of-care discussions, and progress toward extubation and ICU discharge.
I’ve had to piece these views together by searching through different parts of the chart. I’d like the information I routinely need brought together in one place, organized around the setting I’m working in.
You can extend this to other specialties. A cardiology workspace, for example, could bring together prior catheterizations, echocardiograms, stress tests, relevant labs, and guideline-directed medications. The underlying record could stay the same while each specialty gets a view that makes its work easier.
Dashevsky’s Dissection
I think some of the most useful improvements for physicians could come from building around the EHR. My experience with Avicenna showed me that the underlying record could stay the same while the work of reviewing it became much easier.
What I’m envisioning is a specialty-specific workspace where we review information, prepare notes, and coordinate care, connected to the EHR that holds the official record. In pulmonology clinic, that might mean seeing PFT trends, chest imaging reports, and prior treatment trials together. In the ICU, it might mean seeing ventilator changes alongside blood gases, sedation, and fluid balance without piecing that view together ourselves. Sanara Health is one example of a company building in this direction for critical care (built by fellow Huddler co-founder Harrison Kaplan).
Specialty-specific EHRs already exist, and my experience doesn’t mean every Epic or Oracle implementation looks or works the same. But I still see room for tools that bring the information we routinely need to the foreground, organized around our specialty and clinical setting.
The integration would need to go further than what I had at the VA. Copying a note into CPRS was manageable, but a workspace we rely on throughout the day needs current data and a reliable way to send our work back to the record. If we still have to check both systems to confirm a medication change or whether a note saved, we’ve added work.
AI could help summarize a hospital course or pull relevant information from scattered notes. We’d still need to see where that information came from, when it was updated, and what the summary might have missed. An interface that makes information easier to find also needs to make it easy to verify.
I’d want these products judged on time spent searching, duplicate work, missed information, and whether they help us care for patients. Coding support may help sell the software, but I want the clinical workspace designed around what we need at the bedside.
In summary, I think the EHR could remain our official record while more of our daily work moves into specialty-specific platforms. I’d welcome that change if it gives us a clearer view of the patient and fewer steps to act on it.
