A data sharing system architecture for scientific purposes for a Dutch health care environment

Abstract

Data sharing between researchers in different health care organisations can contribute to more insights and better scientific results. Nevertheless, sharing data between organisations is not as simple as it sounds. The patient is the data owner and regulations, both European and Dutch, restrict organisations, since they can’t just give data away to another organisation. This thesis identifies two strategies for data sharing, namely with consent of the patient or after anonymization of the data. To share detailed data about patients with a third party, the patient must be fully informed and must give her/his consent. To achieve this, the patient must understand its data. For the understanding of data, metadata can help. With already in use ontology libraries, the patient can read and understand its own data. To anonymise the data, this thesis argues, the best solution is to contract this transfer and bind the third party to what it can and cannot do with the data. In that sense, the data set can be more loosely anonymised, which contributes to a better utility for the researcher. The main contribution of this thesis is that it gives a modelled architecture system to let health care organisations share their patients’ data with other research organisations. The system takes data transfer, semantics and law into account. This thesis provides health care organisations a route map to construct a data sharing system, which is validated by experts. The validation of this thesis happened by interactive sessions, where experts could bring input and discuss the various aspects of the proposed architecture. Together with the experts, we optimised the system.