Taming “Rogue Notes”: How Memorial Healthcare is Building an EHR foundation for AI Success
The promise of Artificial Intelligence in healthcare hinges on one frequently enough-overlooked factor: the quality and consistency of clinical documentation. At Memorial Healthcare, a strategic overhaul of their Epic EHR system, led by CIO Dr. Sam Weiss,is demonstrating how to “tame rogue notes” – those inconsistent,free-text entries – and lay the groundwork for a future powered by AI. this isn’t just about cleaner charts; it’s about patient safety, clinician efficiency, and future-proofing the institution’s data ecosystem.
The Problem with “Rogue Notes” & The Power of Simplification
For years, healthcare organizations have grappled with the challenge of balancing detailed clinical documentation with the demands of a busy workflow. The result? Clinicians often resort to workarounds – dot phrases, copy-pasted text, and inconsistent data entry - creating a fragmented and unreliable record. This impacts everything from accurate billing to effective care coordination, and critically, hinders the potential of AI.
Dr. Weiss and his team recognized that complexity was the enemy. Their core beliefs? Make the safest path the easiest path.
This translates into a focus on simplification. Rather of relying on clinicians to remember to document critical details, the team implemented smart, data-driven prompts within the clinical note. A single button click can now route discharged patients to the appropriate primary care provider based on geography and insurance, streamlining a previously multi-step process.
“Don’t underestimate the value of a well-placed single button click,” Dr.Weiss emphasizes. This seemingly small change dramatically reduces errors and improves adherence to best practices.
Balancing customization with the Epic Foundation
A key challenge was navigating the tension between legacy customizations and the standardized Epic Foundation build. Many organizations accumulate years of bespoke modifications,making upgrades and data sharing difficult.
Memorial Healthcare adopted a purposeful strategy: prioritize alignment with Foundation whenever possible. They established a standing review process to identify areas were Foundation now offers robust support, justifying the investment in “refueling” – reverting to the standard build.
This isn’t about stripping away functionality. It’s about recognizing that staying close to Foundation unlocks future capabilities,facilitates participation in larger data ecosystems,and minimizes costly rework down the line.
However, complete rigidity isn’t the answer. Hard stops in clinician workflow are reserved for situations where patient safety is paramount. More frequently enough, just-in-time prompts and carefully placed constraints provide sufficient guidance without disrupting clinical flow.
Real-Time Monitoring & Proactive Intervention
To ensure ongoing success, the team implemented real-time monitoring of sensitive processes. For example, documentation related to violent restraint now triggers immediate outreach from quality teams if inconsistencies or omissions are detected.
This combination – thoughtful constraints, measured adoption of Foundation, and continuous monitoring – allows memorial healthcare to raise standards without alienating clinicians who value their autonomy. it’s a delicate balance, but one crucial for long-term success.
Thinking in Horizons: Preparing for the AI Future
Dr.Weiss encourages his peers to adopt a long-term perspective. He urges them to consider not just what prevents errors today, but what choices will enable a smooth AI-assisted workflow in five or ten years.
“Framed that way, documentation design is not a series of one-off fixes but a strategic investment in how the health system will practice medicine as automation matures,” he explains.
This forward-thinking approach means building a foundation of consistent, structured data that AI models can learn from – without requiring expensive retrofits later.
Key Takeaways: Building an AI-ready EHR
Here’s a practical guide,distilled from Memorial Healthcare’s experience,for organizations looking to optimize their EHR for AI:
* Embrace Dynamic Documentation: Move beyond static dot phrases and leverage data-triggered documentation that adapts to the clinical context.
* Embed Required Fields: Place mandatory fields within the clinical note, blocking sign-off only when absolutely necessary. This minimizes disruption and ensures completeness.
* Real-Time Support: Route documentation fallouts to quality teams in real-time for “at-the-elbow” support, addressing issues proactively.
* Align with Epic Foundation: Prioritize alignment with Epic Foundation to improve data comparability and prepare for AI integration.
* Favor Integrated AI: Choose vendor-integrated AI solutions that write natively into the chart, rather
Related reading