Population-level NDJSON export. Full cohort portability.
Security
Your data. Your rules.
🔐
BAA signed before any data flows
HIPAA Business Associate Agreement with every customer and subprocessor.
🛡
SOC 2 Type I (target Q3 2026)
64 controls monitored continuously via Trust Protocol. Type II target Q1 2027.
☁️
Deploy in our cloud or yours
Standard SaaS, dedicated GCP project, or on-premise.
🔒
Encryption everywhere
AES-256 at rest. TLS 1.2+ in transit.
🚫
Zero PHI in AI processing
Clinical intelligence runs inside GCP. Never sent to external models.
⏹
Revoke anytime — 30-day deletion
Offboard = purge. Certified destruction on request.
Early Results
What the data shows
23%
Dropout reduction in pilot cohort
ML-powered early intervention identified at-risk patients 30 days before discontinuation. Operators who acted on alerts retained significantly more patients.
Pathriva internal pilot · 10,000 synthetic patients · 18 archetypes
0.99
AUC on dropout prediction model
XGBoost classifier trained on 20+ clinical and behavioral features. Cross-validated across 100K patients with realistic dropout patterns.
5-fold cross-validation · SURMOUNT/STEP-calibrated data
"The platform identified 2,145 high-risk patients in our 100K cohort within seconds — the kind of signal that would take weeks to surface manually."
— Internal validation, Pathriva engineering team
Resources
Transparency. Built in.
Open methodology, live compliance scores, and a free ROI tool — because trust is earned, not claimed.