Population Risk Stratification Engine

About

The Population Risk Stratification Engine from CareCORE Analytics uses gradient boosted trees and transformer-based clinical NLP to score patient cohorts across 12 chronic condition pathways. It ingests FHIR R4 Patient, Observation, Condition, and Claim resources and outputs composite risk scores, care gap flags, and social determinant signals via FHIR R4 RiskAssessment resources. Deployed across 14 ACOs and 3 MSSP participants, it has identified 23% more high-risk patients than traditional claims-based stratification alone, enabling earlier outreach and reducing preventable ED utilization by 18%.

Key Features

• 12 chronic condition risk models (cardiovascular, diabetes, COPD, oncology, and more)
• FHIR R4 RiskAssessment output with explainability scores
• Social Determinants of Health (SDOH) integration via FHIR Observation
• Cohort segmentation by risk tier, specialty, and care gap type