Normalized claim
Sepsis identification: 80% increase
AI insights have improved sepsis identification by 80%.
Virtua Health builds AI-enabled patient insights with Microsoft Copilot, Azure AI Foundry, Azure Machine Learning and Power BI
Use case typeClinical decision supportUpdated 5 days ago
Virtua Health is a healthcare provider serving the South Jersey and Philadelphia region across five hospitals and more than 400 health facilities. The organization uses Microsoft technology to deliver AI-powered patient summaries and predictive insights to clinicians and operational leaders at the point of care to reduce cognitive load and improve patient outcomes.
- Organization
- Virtua Health
- Industry
- Healthcare
- Location
- United States
- Published
- July 2026
Reported outcomes
+80%
sepsis identificationOther quantified impact
58-81%heart failure follow-up rate1 daysaverage patient stay97%heart failure model accuracy+84%patient identification−80%sepsis false positives
Strategic outcomes
Speed & agilityClinicians can review patient information much fasterOther strategic outcomeEarlier treatment and improved patient outcomesEmployee experienceReduced cognitive load on staff
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
Heart failure follow-up rate: 58-81% increase
Heart failure follow-up rates have increased from 58 percent to 81 percent.
Correction recorded Jul 26, 2026 · Previously: % increase · Stored the 58% to 81% rate transition as a 23 percentage-point increase.
Normalized claim
Average patient stay: 1 days decrease
And average patient stays are reduced by 1 day.
Correction recorded Jul 26, 2026 · Previously: 1 days decrease · Corrected the singular duration unit.
Normalized claim
Heart failure model accuracy: 97%
The new model achieves 97 percent accuracy, lifting patient identification by 84 percentage points.
Correction recorded Jul 26, 2026 · Previously: 97% increase · Marked 97% as an absolute model-accuracy level, not an improvement.
Normalized claim
Patient identification: 84% increase
The new model achieves 97 percent accuracy, lifting patient identification by 84 percentage points.
Correction recorded Jul 26, 2026 · Previously: 84% increase · Preserved percentage points as a separate unit from percent change.
Normalized claim
Sepsis false positives: 80% decrease
Compared to commercial models, Virtua Health’s model reduces false positives by over 80 percent.
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- Virtua Health
- Provider
- Microsoft
- Maturity
- Production
- Linked source
- Microsoft Customer Stories
The organization uses Microsoft technology to deliver AI-powered patient summaries and predictive insights to clinicians and operational leaders at the point of care to reduce cognitive load and improve patient outcomes
Customer identity supportedSource describes one deploymentMaturity supported
Primary read
Use case focus
Showing 3 of 3
- 1Clinical decision support
- 2Patient engagement
- 3HR support automation
- Reduce cognitive load on clinicians and operational leaders
- Make patient data easier to act on within clinical workflows
- Improve patient outcomes and staff efficiency
- Predict sepsis and heart failure earlier with better accuracy
- Integrated Power BI with Teams and added Copilot as a natural-language layer over the patient data view
- Built a custom transformer-based model in Azure AI Foundry to convert unstructured patient notes into structured heart-failure predictions
- Built a custom XGBoost model in Azure Machine Learning and integrated it into the EPIC clinical workflow for sepsis prediction
- Created an HR virtual assistant with Copilot Studio
- Improved sepsis identification by 80%
- Heart failure follow-up rates increased from 58% to 81%
- Average patient stays reduced by 1 day
- Heart failure model achieved 97% accuracy and lifted patient identification by 84 percentage points
- Sepsis model reduced false positives by over 80%
Architecture
Virtua Health built a unified patient data and semantic layer, then used Microsoft Copilot over Power BI for at-a-glance summaries in Teams. It also deployed a custom transformer-based model in Azure AI Foundry for heart-failure prediction and a custom XGBoost model in Azure Machine Learning for sepsis prediction integrated into EPIC clinical workflows; Copilot Studio powers an HR assistant.
Sources & evidence1
Evidence: Medium65/100Evidence strength
- Customer explicitly identified
- Deployment status explicitly supported
- Primary source available
- Quantified outcome available
- Technical implementation details available
Type: Customer StoryPublished: Jul 23, 2026Publisher: MicrosoftEvidence: PrimaryConfidence: High
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
No published comments yet.