Model-based Design of Trustworthy Health Information Systems

Abstract

ions: Medical applications define a domain with unique abstractions. PHRs are structured using unique ontology’s. Care protocols need to adopt a large variety of abstractions used for describing and characterizing treatment steps, their coordination and their relation to patient states and to the actors of clinical environments. It means that design of HIS’s can be hugely simplified by defining well defined, domain specific abstraction layers in the design process that have intuitive meaning, and precise enough to perform formal verifications. Integration: Healthcare environments and the embedded HIS’s are extremely heterogeneous due to the many factors that influence design. A large range of design aspects, such as information flows, organization models, document models, activity and event models, coordination and timing models, policy models, deployment models and others need to be created and related. Integration and management of these interrelated, heterogeneous models is a major challenge. Model-based design environments facilitate this complex task with tools and methods. Automation: HIS’s need to be highly flexible to local needs, procedures and best practices. There is no chance for acceptance of technologies that impose strong restrictions and rigidity on the treatment process. This makes customization an extremely important requirement. Affordability of customizable HIS’s require high level automation in translating high-level models (that capture variations) into verified, operational systems. Model-based system generation and verification tools can provide solution for this challenge. We believe that various flavors of the rapidly progressing model-based software and systems engineering technologies, such as Model Driven Architecture (MDA) [5], the Eclipse Modeling Framework (EMF) [6], Model-Based Security Engineering (MBSE) [7] or Model Integrated Computing (MIC) [8] and others will play a key role in the future of HIS. Description of Papers The issue includes four contributions from researchers working on model-based software engineering and its application in HIS. A. Formally Defining Medical Processes The paper addresses the formal modeling of medical processes, or care protocols. The objective of the authors is to develop and test a method for precise process specification that has solid semantic foundation for conducting formal analysis of the models. The background of the proposed technique is the Little-JIL process definition language originally developed for software architecture modeling. Authors selected chemotherapy preparation and administration as a target application and have shown that their formal representation is able to capture the required complexity; properties can be verified using model checking. B. Model-Based Design of Clinical Information Systems The paper describes a new framework for the model-based design and implementation of clinical Information Systems. The Model Integrated Clinical Information Systems (MICIS) framework is implemented as a software toolkit that establishes high-level modeling abstractions for representing clinical information flows extended by a range of modeling aspects for the provider’s organization, documents, deployment architecture and privacy/security policies. The MICIS generators translate these high level models to an underlying SOA execution platform. C. Model-Based Security Analysis of the German Health Card Architecture The paper focuses on the methods used by the authors for analyzing the security of the German Health Card System. An essential aspect of any PHR implementation is security. The German Health Cards will include critical, highly sensitive data that will provide major help in healthcare delivery, but need to be carefully protected. The presented analysis of the security architecture of the system is based on UMLSec, the authors’ modeling language for security characteristics. D. Model-Based Assessment of Data Availability in Health Information Systems The core contribution of the paper is the development of a theoretically well established solution for assessing structural qualities of information systems. The authors propose an ontology for describing health information systems and show the practical impact of their method on information management.