The technology of VEIL.AI enables the secondary use of sensitive data.
VEIL.AI is a technology company established in 2020 specialised in the protection of privacy and anonymisation of sensitive data. VEIL.AI helps its clients to utilise sensitive health data in research, innovation and development and to support new discoveries, patient care and better decision-making in healthcare. The company is a spin-off originating from the University of Helsinki, the Institute for Molecular Medicine Finland (FIMM) to be exact, and its employees have extensive prior experience in working with health data. In their work related to various technologies used in privacy protection, they observed several related limitations such as the poor quality of anonymised data and the slowness of the process. Traditionally, anonymisation occurs at the cost of quality, leading to the loss of a lot of information and an inability to utilise the data from the perspective of research needs.
“Actually, that was where the idea came from for developing this innovation and that was the origin of our company. In the TUTLI project, we continued exploring the opportunities for the commercial application of the idea, at which point Tuomo Pentikäinen, one of the company’s founders and its current CEO, joined the team”, explains VEIL.AI’s Key Account Manager Lauri Kuronen.
Data utilisation has emerged as a strong part of the business of biotechnology and health care organisations. However, this involves several challenges, as health data contains sensitive personal data. This data may be only processed in secure and audited user environments. Under valid legislation, the results obtained from the environments must be anonymous, i.e. provided in a format that does not enable identifying individuals. The VEIL.AI Anonymization Engine provides a solution for protecting the anonymity of health data. The solution can be used to produce extremely high-quality, secure, anonymised, synthetic and pseudonymised data, providing new opportunities for the utilisation of sensitive data.
The technology developed by VEIL.AI enables marking certain variables as having more priority than others. This results in data that is anonymous but still lends itself to analysis similarly as with personal data on the level of individuals. Thanks to anonymisation procedure, sensitive data can be shared, joined and utilised safely in research, innovation and development projects.
“VEIL.AI’s solution helps making the secondary use of sensitive data possible. The first use basically refers to things such as the use of a patient records system by physicians in their day-to-day work, where all identifiers all intact. Meanwhile, in research, there is a long tradition in using pseudonymised data where any identifying data in the raw data, such as age, name and social security code have been replaced with codes. However, there is a problem with pseudonymisation concerning reversibility, which enables returning to and re-identifying the original patients”, Kuronen continues.
“In turn, anonymisation continues to modify the data, removing direct and indirect identifying data, which results in data that is no longer subject to the GDPR (General Data Protection Regulation). This allows you to reuse the data and utilise it in various ways. It also allows retaining data quality at nearly the original level. Synthetic data is artificial data which can be based on, for instance, the anonymised data of 200 patients. This data can be used to create a synthetic data set of 2,000 patients mimicking the original, and the new data set can be used in developing and testing a machine learning solution.”
The company currently primarily offers its solutions to the pharmaceutical industry, hospitals, biobanks and other organisations producing and utilising sensitive data.
“We’re currently largely focused on hospitals, particularly university hospitals conducting a lot of research. Pharmaceutical companies are another important core client group as they currently invest a lot in business measures related to data and AI development. Of course, medical device manufacturers and other health tech companies are also interesting client groups for us. Our technology can be applied in any industry that makes use of sensitive data. These industries could include insurance, the FinTech sector and even the defence industry.”
“We also have presence in Denmark, particularly Copenhagen. VEIL.AI was the first Finnish company to be selected into a BioInnovation Institute programme, and as a result, the company founded an affiliate in Denmark. The Greater Copenhagen Area in Denmark and Southern Sweden form a major drug development cluster and we find that there are many future prospects there.”
More versatile data use
“Our value proposition for our clients is that we enable more diverse secondary use and utilisation of data. We can use the data available to the client or, in cases concerning multiple projects, we can facilitate data collaboration between different countries and organisations using anonymised data. We consider that this is, above all, an enabling technology. We do not produce or store data, so we do not serve as what is known as a data silo.”
The data usually comes from researchers with a data set on patients with a specific disease for research purposes, for instance.
“Our AI-based solution sets the VEIL.AI anonymisation process apart from others. For example, if the client has previously had a dataset with one hundred patients that has been anonymised and would like to add one more patient to it, they’ve had to do this addition to the original data set and re-anonymise the data. By contrast, we can also anonymise streaming data, i.e. use a certain process through which we can enter and anonymise data as a constant process”, Kuronen says.
Another example of the use of technology given by Kuronen is the Future Clinical Trials Project carried out with VEIL.AI’s reference company, the pharmaceutical company Bayer.
New methods for the anonymisation of personal data developed by VEIL.AI, which can be used to produce high-quality anonymised data while ensuring privacy protection. The Finnish Social and Health Data Permit Authority Findata granted the permit for the research project, submitting the data, offering a data secure environment for data analysis and ensuring the anonymity of the data. The result of the study was that the same conclusions can be made based on the anonymised data as from pseudonymised research data at the individual level.
According to VEIL.AI, the Finnish operating environment is rather well-suited to health research in an international comparison. Finland has adopted an Act on the Secondary Use of Health and Social Data and has highly extensive patient and health data registries. GDPR sets particular requirements for data use and these must be taken into consideration, for example, in the context of the individual’s privacy protection.
“Of course, the legislation valid in each country also has a specific impact. Work is currently underway for the European Health Data Space (EHDS) whose purpose is to draw up legislation at the EU level for the secondary use of data, for instance.”
VEIL.AI is happy to cooperate with Kuopio Health.
Kuopio Health enables new initiatives
“We have seen that Kuopio has good competence and high-quality research in the healthcare and pharmaceutical industry. The University of Eastern Finland is a high-quality university and we see a lot of potential in this environment. A nice group of people is also naturally an important factor. I was previously involved in Health Capital Helsinki, and during it, I frequently collaborated with Kuopio Health and Aki (Gröhn). Tuomi (Pentikäinen) also had a lot of contacts in Kuopio. This year, we have started to create more concrete collaboration with specific operators in Kuopio and wanted to also be part of the official ecosystem for that reason.”
“I consider Kuopio Health as a good platform that brings together public and private sector operators. They would not necessarily meet in other circumstances. Kuopio Health is also an unbiased party that aims to enable new initiatives and cooperation with its members and other operators, also on an international scale. It is also nice to be involved in where people meet and strike up conversation. If you can build something greater together, that’s the main thing.”
Lauri particularly finds that working groups play an important role in the ecosystem operations.
“The health and well-being data working group has particularly interesting discussions and relevant encounters. These help us think about the cooperation and also learn new things. On the other hand, we can raise awareness about what can be done with anonymised data and which opportunities it can bring”, Kuronen concludes.
