AI for Predicting Patient Readmission

Project overview
A hospital group wanted to identify patients at risk of returning within 30 days of discharge. We built a risk model and a discharge-planning dashboard used by care coordinators.
Client challenges & requirements
Patient data lived across several clinical systems, and any tool had to meet strict privacy and clinical-safety requirements before it reached staff.
- Data integration across EHR, labs, and pharmacy systems
- Predictions needed clear reasons clinicians could act on
- Full audit trail and privacy compliance
Our solution
We built a compliant data platform, a readmission-risk model with explainable factors, and a dashboard that flags at-risk patients with suggested follow-up actions.
22%
Fewer 30-day readmissions
5
Hospitals live
30k
Patients screened
7 mo
From kick-off to rollout