Ferry — Environment Seed Report

Patients seeded

of 200 in source

Rows source → target

5,006 → 1,358

27.1% of source

PHI columns masked

+ 8 identifiers re-keyed

Orphaned foreign keysOK

referential integrity

Referential Integrity 18 FK relationships checked · policy error Pass

0 Dangling FKs

Every foreign key in the slice resolves.

Loaded in topological order · target constraints ENFORCED by the database · no dangling references.

Subset 1,358 of 5,006 source rows · click a row to see why it's included

Table	Source	Seeded	%
allergies	36	8	22.2%
Pulled downward to complete each anchored patient's record · via allergies.PATIENT → patients.
conditions	453	118	26.0%
Pulled downward to complete each anchored patient's record · via conditions.PATIENT → patients.
encounters	925	246	26.6%
Pulled downward to complete each anchored patient's record · via encounters.PATIENT → patients.
immunizations	247	68	27.5%
Pulled downward to complete each anchored patient's record · via immunizations.PATIENT → patients.
medications	453	134	29.6%
Pulled downward to complete each anchored patient's record · via medications.PATIENT → patients.
observations	2,372	623	26.3%
Pulled downward to complete each anchored patient's record · via observations.PATIENT → patients.
organizations	6	6	100.0%
Pulled upward to satisfy referential integrity — referenced by encounters.ORGANIZATION, providers.ORGANIZATION; required so no foreign key dangles.
patients	200	51	25.5%
Anchor — the root rows the slice grows from, selected by the run's anchor predicate.
payers	6	6	100.0%
Pulled upward to satisfy referential integrity — referenced by encounters.PAYER, medications.PAYER; required so no foreign key dangles.
procedures	293	83	28.3%
Pulled downward to complete each anchored patient's record · via procedures.PATIENT → patients.
providers	15	15	100.0%
Pulled upward to satisfy referential integrity — referenced by encounters.PROVIDER; required so no foreign key dangles.

PHI Maskingsample records · before → after

FIRST Jackson → Olivia

LAST Fowler → Tainui

NHI BMG3158 → GVZ1942

ADDRESS 84 Ponsonby Crescent → 8754 Kōwhai Cres

ZIP 4352 → 4300

BIRTHDATE 2010-10-05 → 2011-09-23

Identifiers Masked 31 columns · direct identifiers in the schema (not a de-identification certification)

Names ✓ Masked

Geographic detail ✓ Masked

Dates — shifted (month/day kept) ✓ Masked

NHI number ✓ Masked

Driver licence number ✓ Masked

Passport / other unique IDs ✓ Masked

Geocodes (lat/long) ✓ Masked

Deterministic Re-keying identifiers re-keyed with a pure function — joins survive masking, no mapping table

allergies.PATIENTconditions.PATIENTencounters.PATIENTimmunizations.PATIENTmedications.PATIENTobservations.PATIENTpatients.Idprocedures.PATIENT

Health-Interchange Exports masked slice emitted as HL7 standards

FHIR R4

bundles · 1,331 resources

out/demo/fhir

HL7 v2.5.1

492

246 ADT^A01 · 246 ORU^R01

out/demo/hl7

Ferry · referentially-intact, PHI-masked EHR environment seeding · DuckDB compute layer Synthetic data (Synthea schema) · not for production PHI without compliance review