Mapping Finnish medical terms into the OMOP Common Data Model started as research collaboration in 2020. The exciting possibility to use numerous OHDSI-developed tools in cohort-building, case-control comparisons, visualisations, and analytics expanded quickly across all Finnish University Hospitals, University of Helsinki, and the Finnish Institute for Health and Welfare (THL). Due to the common need to improve interoperability between Finnish University hospitals, we further strengthened the collaboration to make sure that we will perform OMOP-mappings consistently across the country. During this process we realised that we would benefit from an even deeper collaboration and we therefore continued to align our procedures accordingly. We are now in the process of formalising our partnership with an agreement signed by the FinOMOP host organisations (five university hospitals and THL). FinOMOP has regular update-meetings of 50-60 active contributors of the consortium and key stakeholders. In addition, multiple groups such as steering, mapping, quality assurance, and analytics groups work closely together. Thanks to funding from the EHDEN project, we were able to develop the first versions of our OMOP databases.
When constructing a frequently updated copy of patient data into the OMOP CDM, a robust transformation process for hospitals’ non-standard data structures is crucial. In the Finnish hospitals, we have tens of operational patient IT-systems from which data is extracted for secondary use into a data lake. We developed a process whereby the relevant row-level data is recognised and combined through person ID and then mapped by the OMOP CDM ETL process. In the University Hospitals of Helsinki, Tampere, and Turku, we now have fully operational OMOP databases including all the patients that have been treated in the hospital (catchment population 3,2 M, 60% of the Finnish population). The age of different data sources varies, but since 2010 (electronic prescriptions since 2013) the overall coverage for secondary and tertiary care data is good. Patient data that has been mapped into OMOP so far include given clinical and pathology diagnoses, diagnostic and therapeutic procedures, laboratory values, drugs that have been administered in the hospital and electronic prescriptions. THL has thus far mapped secondary care (in- and out-patient visits) nationwide for given diagnoses and procedures since 2015.
This fall we started our first pilot study in three therapeutic areas for the Finnish Medicines Agency (FIMEA). The pilot rigorously tested the Finnish data providers’ network, licensing authorities, and the Finnish act on secondary use of patient data. The act passed in 2020 and was often referred to when the preparation of the regulatory framework for the European Health Data Space started. Patient data was shown to be valid, deep, and comprehensive. Based on our experience, however, the new legislation has slowed down the effective collaboration of university hospitals and did not fit well into some specific features of federated study approaches. The FinOMOP consortium has just recently been accepted as a new potential data partner into the Darwin EU network, which supports EU decision-making. In this project we will further explore and develop best practices for high-quality RWD studies.
On behalf of the FinOMOP consortium
Tarja Laitinen, MD, PhD
Director of Research at Tampere University Hospital