250,000+ patients. Zero data leaving.
The first federated chronic kidney disease study in German primary care, run with AstraZeneca across 60 GP practices.
Dr. Nicolas Conze · Medical Doctor · docport
Flower AI Summit 2026 · London, UK · 17 min
250K+
Patient records
60
Practices federated
0
Records moved off-site
Reach the same outcome
Two paths. Same first-class status. Both end with you running it independently - like docport now is.
Path 1 · Tools
Build it yourself
Open the recipe — Flower SuperGrid for orchestration, anonymization-by-aggregation, GDPR-compliant query control. Self-serve onboarding of practices as SuperNodes.
Path 2 · With us
Build it with our FDE team
Scoped Pilot to validate, or full FDE engagement to deliver. Includes the legal scaffolding, anonymization pipeline and SuperGrid setup docport used.
The build
A 2024 landmark study showed 86% of patients with lab-confirmed chronic kidney disease never receive a correct diagnosis. AstraZeneca asked docport whether their network of 60 GP practices in northern Westphalia could replicate that study regionally — without moving any patient data, and inside the boundary of GDPR, German professional law and medical device regulation. Working with the Flower team and a specialized law firm, Nicolas designed a layered legal-and-technical architecture: query control, anonymization by aggregation, transmission integrity, and full audit logging. Practices were onboarded as SuperNodes on Flower SuperGrid; the analysis ran across all 60 sites with zero records leaving any practice.
The result was not just a study. It was an operational loop: local dashboards for doctors, continuous benchmarking against peers, and federated analytics for research partners, all without raw patient data ever leaving the practice, executed in compliance with German privacy laws.
Onboarding practices as SuperNodes was as easy as running the federated analysis on the SuperGrid.
Dr. Nicolas Conze · Medical Doctor, docport
Stack
Flower SuperGrid · Flower Framework · Federated analytics · Anonymization by aggregation · Specialist legal review
Timeline
~12 months from kickoff to first published federated study, run in cooperation with AstraZeneca.
What docport is doing next
€1M EU and regional research grant secured for FlipIt — federated risk-prediction models for chronic kidney disease, run on docport's own network for the next three years. Independently led by docport.