At a glance
- Client
- A health insurerName withheld for confidentiality
- Industry
- Healthcare
- Service
- Data & Analytics
- Scope
- Claims data platform, risk model and care management tooling
The challenge
Care managers could only support a small share of members, and they chose whom to call largely from past claims. By the time someone appeared as high-cost, they were often already in hospital.
Claims, pharmacy and member data sat in separate systems.
Our approach
Bring data together
Claims, pharmacy, diagnosis and demographic data were combined in a secure analytics platform.
Predict risk
Models estimate each member's chance of an avoidable admission in the coming months.
Explain the score
Each score lists the main risk factors so care managers know where to start.
Fit the workflow
Prioritised member lists appear in the care management system.
The solution
Care managers now start each week with a prioritised list of members at rising risk, along with the reasons behind each score. Outreach happens earlier, when prevention still works.
Privacy controls and access rules protect member data throughout.
Results
- Unified claims and pharmacy data
- Forward-looking risk scores for every member
- Explainable risk factors for care managers
- Prioritised outreach lists in existing tools
Technology used
- Python
- Healthcare data models
- Explainable ML
- Azure Synapse
- Power BI
Facing a similar challenge?
Tell us what you are working on. We will walk you through how we would approach it, with no obligation.