The essence of the PHT approach is that the research question ('trains') travels to the data sources (‘stations’) rather than data from various sources having to be transported to the research question. So, sensitive data remains where it is. The PHT can transport workflows for complex research questions; ‘Which data elements are most predictive of survival after lung cancer given all data in the Netherlands?’; or more specific for a citizen: ‘Which data stations contain data about me?’ Besides such federated learning, a second main principle underlying the PHT is that health data must be FAIR (Findable, Accessible, Interoperable and Reusable).
The PHT infrastructure connects different FAIR data stations and secures and governs that connection ('track'). Stations can range from hospitals and health insurers to research cohorts and smart phones and from very large databases to small personal lockers containing the data of one person. Each station has its own set of house rules governing what a visiting ‘train’ is allowed to do with their data. Ranging from ‘nobody can do anything’ to ‘anyone can use this element as they wish’. Already more than 40 organizations worldwide have agreed to using the PHT infrastructure and give access to their data.
A core part of the PHT architecture is its holistic governance i.e. the definition of interoperability requirements and standards AND the legal, ethical and societal framework needed for using such sensitive data. In fact, PHT can be seen as an implementation of a set of rules ('afsprakenstelsel'), and a technical implementation of the Medmij afsprakenstelsel. The infrastructure (stations, trains and tracks) is being built by the different public, not-for-profit, and for-profit organizations. Convergence of all the components into one interoperable infrastructure, and further development and revision of the 'afsprakenstelsel' requires a significant amount of central coordination and funding.
The main research that is facilitated with the PHT is research to advance preventive and participatory health and health care and predictive and personalized medicine. PHT will also facilitate research in the domain of social sciences and humanities on the ethical, legal and societal implications of such an infrastructure. Finally data, computer and mathematical science is being conducted in the field of federated learning, privacy-preservation, (cyber)security, proven trust and artificial intelligence.