A FIELD GUIDE TO HEALTHCARE AI
Source-linked research, practical decision cases, and working briefs for people bringing AI into healthcare.
Independent. Educational. Built for the whole health system.
Perspective: Whole system
A WHOLE-SYSTEM VIEW
Connect clinical, operational, administrative, technical, governance, and research questions around the same decision.
In the collection: Ambient AI Scribes in Clinical Practice: A Randomized Trial
Explore the whole system →FROM A SOURCE TO A DECISION
A finding is a starting point. Keep the evidence, its limits, and the question for your setting together.
Showing Evidence context
University of California, Los Angeles; NEJM AI · Nov 2025 · Clinical trial
Original publicationWHAT THE SOURCE REPORTS
Nabla reduced time in notes versus control; DAX showed no significant reduction.
Three-arm randomized trial; 238 physicians.
One health system; short follow-up.
A QUESTION FOR YOUR SETTING
Which local outcomes justify continued investment?
IN THE COLLECTION · Whole system
Evidence report · Mar 2025
Participants reported possible workforce benefits; financial effects remained uncertain.
Keep in view: Early-adopter experience, not controlled economic evidence.
Open for scoped reviewOriginal evidence ↗Clinical trial · Apr 2025
Intervention encounters had lower mortality hazard and shorter stays; unanticipated ICU transfers increased.
Keep in view: Results depend on documentation and response workflows.
Open for scoped reviewOriginal evidence ↗Framework · Oct 2026
The release describes 57 draft controls across 11 domains.
Keep in view: Draft guidance; version may change.
Open for scoped reviewOriginal evidence ↗Published-study synthesis
Study question: do ambient scribes change documentation time? Editorial prompt: establish local pilot continuation criteria.
Still unresolved: Long-term financial returns were not established.
Explore this decision →Published-study synthesis
Editorial decision question: Could a shared display of nursing-derived risk signals improve our escalation process?
Still unresolved: Local training and escalation ownership are not documented here.
Explore this decision →YOUR NEXT DECISION
Evaluate
Consider study design, population, comparators, local relevance, and the information that is still missing.
A working document for the problem, people, evidence, risks, and next steps.
Begin here. Sign in when you are ready to save privately and collaborate.
CONTRIBUTE CONTEXT
Contributions are assessed, attributed, and connected to the part of the record they can improve.
A publication, evaluation, or report worth examining.
The workflow, people, and practical limits a publication may miss.
Help make a record more accurate, useful, and complete.
Relevant experience is welcome across all six perspectives. Submission is not automatic publication.
Editorial method, attribution, review scope & disclosures →