Yes, I was a physician builder once. Briefly. I admire my IT and PB friends who can manipulate structures and build rules. And now, enter AI Build Agent. The game changes again
Epic shows us more of the guts of the build inside agent factory, simplifying those skills that Physician Builders struggled to gain from certificate courses.
AI getting into forms, reports, finding reporting workbench reports you need OR launching slicer and CHOOSING THE RIGHT DATA MODEL? Waat.
My frustration with reporting workbench has been: we have too many for me to figure out which is the right one to dig into to get my answer. Now, ask a question to the LLM-powered engine and it will search for you. Better yet, if there is not a reporting workbench report, it will FIND THE RIGHT DATA MODEL IN SLICER and run your first pass for you. OMG. Simplifying the complex.
OK, is this magic? Because, AI will now take a PDF or AN IMAGE YOU DREW ON A NAPKIN and make a smartform.
I will just leave this here. My brain is broken.
Oh, why don’t we just turn that into a registry? Sure…
What the heck is an intelligent visit?
ClinAC01 Clinical Advisory Council Welcome
Highlights: Rover Desktop: using Rover connected by cable to a display can drive the display to show and interact with HYPERSPACE. Waat.
Chart with ART for nursing. Able to capture voice conversation and fill in: Vitals, I/O, more than flowsheets: activity. Expressed education understanding. Planned discharge date. Whoooo.
Chart with Art inbasket reply will soon incorporate upcoming appointments, chart insight summary, and then re-draft a reply (version Nov 25). From narrative, be able to abstract SDOH into the circle.
Storyboard will soon be able to place Cologuard or FIT testing without leaving the storyboard (!)
DANbient, being trialed by Group Health, using Chart with ART will allow prompt customization of the note, resolving HCC alerts dynamically as diagnoses are captured in narrative in addition to capturing orders and diagnoses in the conversation.
Haiku will soon have Web hosted note editor: full Notewriter tools that is visible and editable on mobile (?!)
Too much to summarize. Thanks to our developer friends for another rush of new features.
Emmie interviews the patient ahead of time
The MA/RN can review the patient’s med records and updates IN LINE
Visit topics get an intelligent problem-based summary in line
Orders, diagnosis generate dynamically, not needing to pause to do so. Progress notes flow in, based on diagnoses.
All the pieces work together.
Things are happening, people.
Madison little delights. I love having hometown food in Madison, specifically Taiwanese food. This tiny place has been in business for years and I’ve loved dropping in here; almost never busy, and the flavors are authentic. Thanks to Steve Rotholz for a nice informatics chat during dinner. Cheers!
A pediatric DKA risk score!
It was great to see Dr. Danny Lee and Bala Kulandaivel at their poster. They did the hard math to create a predictive model that detects pediatric diabetic ketoacidosis, and displayed it on the multi-provider schedule and on in patient trackboard units.
Congratulations to their achievement! This is hard work, difficult math, really hard tuning to make the alerts relevant to busy physicians/providers in an acute care setting.
Reflecting on our own journey with the Epic Sepsis model and other deterioration scores, we found:
These are some details of our difficult journey to the acceptance of predictive alerts like this.
To Dr. Lee and team, great work on a new predictive model! and best wishes on avoiding some of the potholes on this really important journey toward a successful intervention! We will be eagerly watching!
Dr. Macdonald meets the challenge of the California need to consent patients prior to using an AI scribe.
California is a “two party state” where both parties must consent to any recording. This applies to AI scribes in the exam room, and can be a bother to manage correctly with every visit with every patient. For some organizations this is millions of times per year.
Dr. Macdonald’s big idea?
Set a smart data element as part of the question for any clinician to ask the patient “OK if my AI scribe writes my note today? If ok for your visit with me, OK if all your future visits with UCDavis clinicians can do the same?” (my paraphrase, his wording is of course better). If the patient answers “yes” which they do the vast majority of time, then the EHR stores that and shows it to all subsequent clinicians throughout the organization, so no need to ask again. ONE consent, ONE time, revokable by patient if needed.
SO SIMPLE. Thanks, Scott, for this innovation.
Eric Shelov and Margery Schonfeld from Childrens Hospital of Philadelphia tell us how Modern Notes are done. Listen up!
It is impressive what Drs. Shelov and Schonfeld have accomplished and also measured, in their efforts at improving progress notes, inpatient and out. They used disappearing help text, linking out just-in-time, Diagnosis Aware Notes, and have a plan to migrate to DANbient. Seriously, take a listen to this talk when it is published on userweb. Totally worth it. Some screenshots to whet your appetite.
How are inpatient AI tools coming along? Inpatient insights and draft hospital course both with great adoption and satisfaction.
PAC04
Pretty well, it turns out. St. Lukes and Inova colleagues show us: the tools work well and their front line clinicians love them. Thanks to our smart colleagues!
Agent Factor within Epic. Imaging vibe coding your own agent to connect data and take actions. Is it a dream? Yes. Is it coming soon? Also, yes.
PAC Welcome post #2, this deserves its own observation
AGENT FACTORY.
For example, have a high risk evolving situation, a measles case unintentionally exposing a host of staff and patients in a hospital? Create an agent to assess the exposure, follow procedure, notify and act! Low or no-code assembly. Epic is developing a workspace for organizations deploy agents within Epic. Available soon.
It is possible that this will transform physician informatics from coding deep in the Epic code to becoming vibe coders, using their deep knowledge of clinical workflows and best practice design to interact with Epic more efficiently and build sophisticated tools.
The future is here.
