HealthNet reduces 30-day readmissions by 22%

{ TODO: replace with real customer story. This is a Phase 2 mock. }

The challenge

HealthNet’s care-coordination team was prioritizing post-discharge follow-up based on a simple risk score. It missed complex cases where risk emerged from the combination of diagnoses, medications, and prior admissions — exactly the kind of signal a relational model can learn.

The Langsat approach

• Trained on de-identified EHR data across 8 hospitals
• Model consumes the graph of patients, encounters, diagnoses, and prescriptions
• Output feeds the care team’s existing discharge dashboard

Outcomes

• 22% fewer 30-day readmissions among flagged patients
• Care coordinators now see a risk driver breakdown, not just a score
• HIPAA-compliant deployment in customer’s private AWS account

{ TODO: add Chief Medical Officer quote and compliance details. }