Author: Janne Pullat, Head of the Health Data Unit at Metrosert
Investors ask what your evidence is based on. The regulator asks how the data has been collected and processed. The healthcare provider asks whether the results are transferable to their context. These questions do not concern analytical capability alone. They concern data structure, interoperability, and secure use.
The secondary use of health data is an integral part of health technology development. Registry data, electronic medical records, and other data generated in health information systems are used to assess clinical value, define target groups, and analyze impacts. The reliability of the solution depends on how thoughtfully the database is structured.
When data comes from several hospitals, registries, or countries, its substantive comparability must be ensured. Technical integration alone does not guarantee meaningful consistency. Diagnoses, procedures, and laboratory parameters must be coded according to common principles and documented in a way that allows the logic of the analysis to be explained later. Structured data exchange is supported by HL7 FHIR, which provides a standardized means of communication between clinical systems. At the level of analytical harmonization, the OMOP Common Data Model is often used, enabling data from different sources to be handled within a comparable framework. Such an architecture allows the creation of reproducible and defensible evidence.
Data quality is not limited to the values in the table. It is equally important to understand how the data was generated. What was the clinical workflow? Have coding practices changed over time? What transformations were applied before analysis? Metadata and a description of the origin of the data make it possible to assess whether the dataset is suitable for validating a specific solution. Without this context, the model may be technically correct but substantively difficult to interpret.
Information security is an integral part of development. Processing health data requires a controlled environment, clearly defined access rights, authentication, encryption, and activity logs. Secure processing creates the conditions for collaboration with data holders and partners and reduces the risk that development will stall due to regulatory concerns.
In the European context, the environment for secondary use is shaped by the European Health Data Space Regulation, which establishes common principles for data access and use. Solutions that take interoperability and security requirements into account early in the development phase are better prepared to operate in a cross-border data space.
Technical practice is moving in a direction where analysis is brought to the data rather than centralizing data in one place. Distributed analysis and privacy-preserving methods allow to combine knowledge from different data sources without centrally transferring raw data. This combines data protection principles and development needs into a coherent whole.
The secondary use of health data is not a post-project activity, but part of the solution architecture. Once the data structure has been harmonized, the origin documented, and security technically ensured, the answers to the initial questions also become clearer. The evidence is transparent. The analysis is defensible. Cooperation is possible.
The secondary use of health data is not a post-project activity, but part of the solution architecture.
If you are developing a health technology solution and want to assess whether your data architecture supports scientifically sound and regulatory-compliant validation, please contact us. We will help you map data flows, assess compatibility, and design a solution based on clear, documented, and secure data usage.
Janne Pullat
Head of the Health Data Unit
+372 5750 2344
janne.pullat@metrosert.ee
Used sources:
- Blom JMC, et al. Embedding equity in clinical research governance. Nature Medicine. 2026. doi:10.1038/s41591-026-04232-w.
- Everson J, Strawley C. Health interoperability across phenotypes of family physician practices. Journal of the American Medical Informatics Association. 2026;33(2):434–441. doi:10.1093/jamia/ocaf178.
- Souza R, et al. Combining federated learning and travelling model boosts performance and opens opportunities for digital health equity. npj Digital Medicine. 2026. doi:10.1038/s41746-026-02483-y.