Front page » News » Making better use of health data – new methods improve risk prediction and patient monitoring

Articles

Making better use of health data – new methods improve risk prediction and patient monitoring

Huge amounts of health data are collected, but utilized to a limited extent: fragmentation of the data and strict legislation have hindered the development of solutions. NHG joins the international Secure-e-Health research and development project where, together with partners, methods are created that can be used to process fragmented data without jeopardizing information security. This accelerates data-based health-promoting innovations, such as cardiovascular disease risk prediction models. In addition, new methods bring benefits to public health and competitive advantage in the global market.

Health data is sensitive information. Thus, it is justified that the law defines precisely how the data is used.

“In a scientific sense, however, this is a challenge, as strict regulation slows down the development of data-based solutions. For example, in order to transfer data for analyses, many kinds of agreements are needed”, says research team leader Mika Hilvo from VTT, Technical Research Centre of Finland.

The international Secure-e-Health project develops new methods for secure processing of fragmented health data. This lays the groundwork for the development of health-promoting services, where data are used in disease risk prediction and clinical decision-making, for example.

Federated learning and data harmonization play a key role

Data fragmentation refers to both the different locations of the data and the methods used to collect it. Federated learning, i.e. local processing of data, solves the challenges associated with transferring data. The health data itself is not transferred, but instead parameters are calculated from it, and they can be transferred securely. These parameters can be used, for example, to create predictive models that support doctors in decision-making.

Health data is collected, for example, by home monitoring methods and in various patient information systems. Information obtained from different data sources must be harmonized before it can be used in risk prediction. New, standardised data models provide means for this.

In order to protect health data, effective identity and user management is also needed to limit access to data precisely. In addition, we need new encryption methods and algorithms that guarantee data security even in the era of quantum computers. These methods are also being developed in the project.

Application area: risk predictions of cardiovascular diseases

The new methods will first be applied to cardiovascular disease risk predictions. Together with its company partners, VTT is investigating how the condition and rehabilitation of a person suffering from heart problems can be predicted using health data.

The heart patient’s condition is monitored at home using Bittium’s wearable device and Solita’s home monitoring system. This data is combined with the data obtained from Mediconsult’s patient information systems and to official register data acquired by Success Clinic. CSIT is responsible for health data security and access control. VTT is developing a machine learning model that predicts risks based on data. Nordic Healthcare Group is investigating how the new methods can be integrated into the patient’s treatment path. Success Clinic develops methods to model the changes in medication use. Eventually, the aim is to incorporate the new machine learning models into patient information systems.

Large scale health benefits and global business

There are great expectations related to health data, machine learning and artificial intelligence. With the help of new methods, these promises can now be redeemed by improving the prediction of risks and the identification of patients at risk. This can facilitate decision-making and the allocation of health care resources, as well as produce public health benefits and have a broad impact on people’s well-being even before getting sick.

“Finland is at the forefront of research of the new methods: we have unique health data reserves and top-class expertise in developing these innovations. Solutions based on health data have enormous potential in the international market. Finland has good opportunities to be in the front line of this development also globally,” states Hilvo.

The Secur-e-Health project started on January 1, 2023 and will last for three years. Five countries participate in the project. VTT coordinates the Finnish country consortium, which includes six companies (Bittium, CSIT Finland, Mediconsult, Nordic Healthcare Group, Solita and Success Clinic). In Finland, the project is financed by Business Finland.