The business
Our client is a regional Midwest health system with several hospitals, clinics, home care and its own health plan. Its mission is to improve the health of its communities, and to care for each patient across every setting.
The challenge
One view of each patient, and a clear path to AI
Data in many systems
Clinical, health plan and operational data sat in separate systems, so each report started with another data pull.
Applications that were hard to connect
Older applications held key records but could not share data easily with new platforms.
Population health needs one view
Care gaps, high-risk patients and value-based contracts all depend on the same trusted data about each patient.
The right home for each AI use case
Patient data is protected, so each AI use case needs a clear decision on which model it uses and where it runs.
The solution
What we are delivering
- Application modernizationOngoing
- Data foundation on Microsoft FabricQuarter by quarter
- AI use case prioritizationNow
- A data strategy and roadmap
- Agreed with leadership, setting the path from a shared data model to population health use cases.
- Modernized applications
- Older applications updated or rebuilt, so they share data with the data foundation and are ready for AI.
- A data foundation on Azure and Microsoft Fabric
- Clinical, health plan and operational data brought into one governed model, released in stages on a quarter-by-quarter roadmap.
- Built for population health
- One view of each patient that supports care gaps, high-risk patient lists and value-based care reporting.
- Reports and dashboards
- Leaders and teams plan and track from the same data, with less reliance on reports from each source system.
- Patient data protected
- Access limited to what each role needs, patient data de-identified where needed, and every access logged.
- AI use cases scored
- Each idea scored on value, data readiness and risk, with a recommendation on whether it runs on a frontier model in the cloud or an open model on-premises, where patient data has to stay inside the health system.
Progress
Where the program stands
Data foundation
Clinical, health plan and operational data in one model on Microsoft Fabric.
Application modernization
Applications updated in stages, alongside the data foundation.
AI use case prioritization
The first AI use cases are being chosen, with a model and a place to run for each.
First AI use cases live
Built on the same governed data, with a person approving anything that affects a patient.
What’s next
The health system will put its first AI use cases into production on the same governed data. Each one will run on the model that fits it: a frontier model in the cloud for complex reasoning, or an open model on-premises where patient data has to stay inside its walls. See how we score AI use cases and choose the right model.