Kaivo - Pseudonymised Health Data Warehouse for Leadership Analytics and Permitted Research

Platform home, leadership overview, pipelines with run details, the research request desk and a request detail with 7 tabs, from permit and data specification to outputs and messages.

25 nightly pipelines across ingest, pseudonymise, model and marts, drawn against the 06:00 freshness deadline. When an ingest fails, the steps that depend on it are skipped and shown as such.

Care guarantee waits, primary care T3, bed occupancy, 30-day readmissions and patients waiting for further care placement, each with its target, trend and 13 months of history, from aggregated marts only.

14 research data requests across all 10 statuses, with Findata and county permits, plus 12 source systems, 25 pipelines, 7 quality issues and 13 staff users.

Every Purpose Has Its Own Rules

Under the Act on the Secondary Use of Health and Social Data, the county may generally use its own data for its own knowledge management, while research, development and innovation generally need a data permit. The product has to know the purpose behind every access and every project.

Permits From More Than One Body

Findata grants permits when data comes from several controllers, private providers or Kanta; when only the county's own data is used, the county as controller may in some cases decide the permit itself. Project access must stay inside the permitted data, purpose and validity.

Pseudonymised Is Still Personal

Under GDPR, pseudonymised data is still personal data. Only aggregates that pass disclosure control count as anonymous, so permitted data stays in a secure processing environment and every query, export request and key operation must be logged.

Numbers Nobody Agreed On

Waiting times, bed days and readmissions were counted differently across the old organisations, and the care guarantee limits in the Health Care Act have changed more than once. Leadership needs one definition per measure, with an owner and dated rule versions.