AI-based next-generation anonymization technology developed in Finland enables versatile and high-quality research utilization of health data.
Bayer, a pharmaceutical company, is currently undertaking a three-year development project in Finland called Future Clinical Trials, which utilizes new technology from a Finnish privacy-enhancing technology company, VEIL.AI. The project uses new methods of personal information anonymization to produce high-quality anonymized data while ensuring privacy protection.
Health data contains sensitive personal information. According to Finnish law, they can only be processed in secure and audited environments. Results taken out of these environments must be anonymous, in other words, a form that cannot be used to identify individuals.
Finland provides an internationally favorable operating environment for health research activities. Two factors contribute to this in particular including comprehensive patient and health data registers, and legislation on the secondary use of social and health data. Finland is the first European country to have this legislation in place.
One of the challenges in using health data has been the quality of anonymization and verification of anonymity.
“Traditionally, anonymization is often done at the expense of quality, resulting in significant information loss,” says Tuomo Pentikäinen, CEO of VEIL.AI.
The newly developed technology now enables the prioritization of certain variables over others.
“For example, we can define that blood pressure is an important variable in this dataset, but height and weight are not. Artificial intelligence takes into account the user’s priorities and selects anonymization methods based on that among numerous options. From a health technology perspective, this can be seen as a breakthrough,” says Pentikäinen.
The result is anonymous data that can be analyzed in the same way as individual-level personal data.
“This is a significant achievement. In our study, we could draw the same conclusions from anonymized data as from traditional pseudonymized, individual-level research data,” says Jussi Leinonen, the person responsible for the project at Bayer.
The research project by Bayer and VEIL.AI has been granted permission by Findata, the Social and Health Data Permit Authority. Findata has also provided the data, offered a secure environment for data analysis, and verified the anonymity of the results.
“Finland’s operating environment for the secondary use of health data is considered unique in Europe and globally. This project is an example of how Finland leads internationally in better utilization of health data,” says Johanna Seppänen, the director of Findata.