!""#$%& ()*"+,$-. /%0$%1 23 Coordinating e-Health Systems with TuCSoN Semantic Tuple Centres Elena Nardini Andrea Omicini Mirko Viroli Alma Mater Studiorum – Università di Bologna, Cesena, Italy {elena.nardini, andrea.omicini, mirko.viroli}@unibo.it Michael I. Schumacher University of Applied Sciences Western Switzerland (HES-SO), Sierre & Fribourg, Switzerland michael.schumacher@hevs.ch ABSTRACT Open and distributed application scenarios like e-Health sys- tems mandate for new coordination models and technologies. In particular, they require middleware providing coordina- tion and security services modelled with abstractions pro- moting run-time observability and adaptation. Along this line, in this paper we describe the architec- ture of the TuCSoN coordination infrastructure, and show its application to an e-Health application scenario. 1. INTRODUCTION Healthcare supported by software systems – in short, e- Health – is evolving quickly [7, 24]. Among the several e-Health research activities, research on Electronic Health Record (EHR) is particularly intensive [24, 43]. An EHR represents a set of medical information called fragments, which are stored in a digital format over different healthcare institutions. The introduction of EHR offers several bene- fits [26]: better patient safety, lower cost of health services, and improvements in healthcare audit and research. The main challenge in the EHR domain is to ensure interoper- ability among EHR fragments belonging to an environment that is distributed and open, and where the security support represents a fundamental requirement to protect the patient privacy [23]. Several efforts have been made in the EHR do- main in order to cope with such requirements, in particular: • Definition of standards for EHR-fragment format and communication such as Health Level Seven (HL7) [11] and Digital Imaging and Communications in Medicine (DICOM) [21]. • Definition of standards like openEHR [13] and CEN EN 13606 [15] promoting semantic approaches to face heterogeneity and dynamism of fragment formats. • Definition of specifications like the IHE (Integrating the Healthcare Enterprise) specifications [20], to build EHR coordination middleware able to coordinate EHR- fragment providers and requesters. • Definition of EHR coordination middleware based on the semantic tuple-space computing [5], such as the one promoted by Triple Space Computing (TSC) [1]. As shown by TSC, semantic tuple-space computing provides models for EHR coordination middleware which overcome the limits of the solutions proposed in the IHE specifica- tions. In fact, unlike IHE, TSC provides a solution based on semantic techniques as suggested by openEHR and CEN EN 13606, and models for building applications that sup- port more complex dynamics than the mere exchange of EHR fragments. On the other hand, like IHE, TSC pro- vides coordination media that cannot be tailored to the spe- cific application needs. Thus, any coordination law not di- rectly supported by the model has typically to be charged upon coordinated entities, thus increasing their complexity, especially in open scenarios. Moreover, TSC cannot cope with the evolution of applications over time, since it makes it difficult to adapt the coordination middleware in case of changes to application requirements. Along this line, in this paper we aim at providing mod- els and approaches extending the solutions proposed in IHE and TSC, in order to augment their effectiveness in building EHR services, in particular as far as interoperability is con- cerned. To this end, we draw our inspiration from TuCSoN [33], an open-source coordination infrastructure based on se- mantic tuple-centres [27], and promoting online engineering [35] through a middleware adaptable at runtime, thus coping with the dynamism requirements. In particular, this article is organised as follows. In Sec- tion 2 we briefly survey the existing approaches supporting interoperability among EHR fragments, and discuss their benefits and drawbacks. Then, in Section 3 we present an overview of the TuCSoN architecture. In particular, we de- scribe the key-features of the architecture: (i) the behaviour programmability and the semantic support of tuple centres — the coordination abstraction exploited by TuCSoN; (ii) the organisation and RBAC models, used respectively to model the system structures and their relationships, and the secu- rity aspects; and (iii) the online engineering approach, used to support the corrective/adaptive/evolutive maintenance of software systems. In Section 4, we show how to exploit the key-features of the TuCSoN architecture in order to extend the solutions proposed in IHE and TSC. Finally, Section 5 concludes, providing final remarks.