MedRight: AI-Powered EHR Workflow Optimization

The Challenge

Healthcare workers using EHR systems like Meditech and Cerner face critical workflow failures that compromise patient care. Nurses cannot quickly identify urgent patient needs among administrative noise, persistent "red clock" alerts obscure real priorities, and emergency situations force teams to abandon digital systems entirely for paper charting.

"How might we create an AI-assisted EHR interface that intelligently organizes clinical information, streamlines routine workflows, and remains reliable during emergency situations while keeping all clinical decision-making with healthcare providers?"

Research & Discovery

Drawing from my clinical nursing experience on a medical-surgical unit and user interviews with healthcare colleagues across Fraser Health (Meditech) and Vancouver Coastal Health (Cerner), I identified critical pain points:

Key Findings

• Interface Navigation: Excessive clicking and cumbersome workflows force nurses to become system administrators rather than focusing on patient care

• Alert Fatigue: Overdue medication alerts mask new critical alerts, creating safety risks

• Temporal Task Management: Previous shift errors compound indefinitely with persistent indicators that cannot be cleared

• Emergency Abandonment: Staff resort to paper charting during critical situations because EMRs are too slow and unreliable

Impact on Patient Care

• Critical medication timing gets buried in poor interface design

• Nurses spend valuable time navigating cluttered interfaces instead of providing direct patient care

• Transcription errors increase when reverting from paper back to digital systems post-emergency

Solution Approach

AI-Powered Clinical Intelligence that analyzes multiple data streams to create real-time priority rankings, helping nurses instantly see what actually needs attention right now.

Core Features in Development

1. Dynamic Clinical Priority Intelligence - Real-time patient acuity analysis

2. Intelligent Alert Categorization - Context-aware notifications that reduce noise

3. Smart Task Clustering - Logical groupings that match clinical workflows

4. Temporal Context Management - Time-based alerts that reflect clinical reality

Success Metrics

• Time to Priority Identification: <30 seconds to identify most urgent needs

• Alert Relevance: >85% of high-priority alerts require immediate action

• Emergency Response: System remains functional during critical situations

Early Concepts

Real-World Applications:

• Shift start prioritization with AI-generated patient summaries from the last shift, tasks and alerts

• Administrative task grouping to prevent workflow disruption

• Emergency-resilient interface design for critical care situations

Technical Safeguards

• Clinician Override: Nurses can always manually reprioritize AI suggestions

• Transparency: Clear indicators showing why AI ranked something as high/low priority

• Learning Loop: System learns from nurse actions to improve future prioritization

🚨 What AI Does NOT Do:

• Make clinical recommendations or suggestions

• Predict patient outcomes or care needs

• Override clinical judgment or hospital protocols

• Automatically complete clinical documentation

The AI in this scenario will serve purely as an organizational tool to surface relevant information. All clinical decision making remain with licensed healthcare professionals.

Next Steps

Currently developing:

• Interactive prototypes demonstrating AI-assisted workflows

• Detailed user journey mapping for emergency vs. routine scenarios

• Technical implementation considerations for EHR integration

This case study leverages authentic clinical experience to address real healthcare workflow challenges. Full documentation including detailed research findings, user personas, design iterations, and interactive prototypes available for discussion.

Contact me to learn more about this project and see the complete case study.