The Digital Patient 220: ECU Health’s CMIO Dr. John Hanna: What Clinicians Build When Given No-code AI Tools, Why Most Health Systems Are Governing AI Backwards, and Is a Personalized EHR a Dream or Inevitable?
Feb 25, 2026
Dr. John Hanna, CMIO at ECU Health and physician-informaticist who runs hands-on AI workshops, discusses what clinicians build with no-code AI tools. He talks about clinician-built agents, real-world AI governance pitfalls, personalized EHRs versus standardization, monitoring AI performance in practice, and practical lessons from sepsis model deployments.
AI Snips
Chapters
Books
Transcript
Episode notes
Show AI Insights Inline Not As Popup Alerts
- Surface AI insights inline with the clinical conversation rather than as interruptive alerts to avoid annoyance and workflow disruption.
- Inline decision support that appears before a patient leaves the visit improves usefulness compared with late BPA-style pop-ups.
Bounded Personalization Will Replace One-Size-Fits-All EHRs
- Healthcare will likely move from one-size-fits-all EHRs toward controlled personalization where clinicians can customize interfaces within bounded guardrails.
- John calls this bounded flexibility and expects specialty-aware customization to evolve into individual personalization.
Monitor Problem Model And Workflow As One System
- AI monitoring must cover three intersecting lifecycles: the problem, the AI/model, and the clinician workflow — each requires different signals and owners.
- Treat monitoring as a pyramid: outcomes at the base, model performance in the middle, usage/interaction at the top.
