Personal Health Train Architecture with Dynamic Cloud Staging

SN Computer Science, vol. 4, pp. 14

Abstract

Scientific advances, especially in the healthcare domain, can be accelerated by making data available for analysis. However, in traditional data analysis systems, data need to be moved to a central processing unit that performs analyses, which may be undesirable, e.g. due to privacy regulations in case these data contain personal information. This paper discusses the Personal Health Train (PHT) approach in which data processing is brought to the (personal health) data rather than the other way around, allowing (private) data accessed to be controlled, and to observe ethical and legal concerns. This paper introduces the PHT architecture and discusses the data staging solution that allows processing to be delegated to components spawned in a private cloud environment in case the (health) organisation hosting the data has limited resources to execute the required processing. This paper shows the feasibility and suitability of the solution with a relatively simple, yet representative, case study of data analysis of Covid-19 infections, which is performed by components that are created on demand and run in the Amazon Web Services platform. This paper also shows that the performance of our solution is acceptable, and that our solution is scalable. This paper demonstrates that the PHT approach enables data analysis with controlled access, preserving privacy and complying with regulations such as GDPR, while the solution is deployed in a private cloud environment.

Authors 5

1. Luiz Olavo Bonino da Silva Santos corresponding Aachen

   Leiden University Medical Center · University of Twente

   Affiliation as printed

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500 AE The Netherlands

   Leiden University Medical Center, PO Box 9600, Leiden, 2300 RC The Netherlands

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500, AE, The Netherlands

2. Luís Ferreira Pires

   University of Twente

   Affiliation as printed

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500 AE The Netherlands

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500, AE, The Netherlands

3. Virginia Graciano Martinez

   University of Twente

   Affiliation as printed

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500 AE The Netherlands

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500, AE, The Netherlands

4. João Moreira

   University of Twente

   Affiliation as printed

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500 AE The Netherlands

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500, AE, The Netherlands

5. Renata S. S. Guizzardi

   University of Twente

   Affiliation as printed

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500 AE The Netherlands

   Faculty of Electrical Engineering, Mathematics and Computer Science, University of Twente, PO Box 217, Enschede, 7500, AE, The Netherlands