Data Quality by Contract --- Towards an Architectural View for Data Quality in Health Information Systems

Clinical Information Systems CIS have become a pivotal appliance in modern healthcare systems. Their adoption and increasing integration is driven by expectations related to better health outcomes and cost effectiveness. In practice, however, a lack of data quality DQ is often referred to as a significant inhibitor, impeding the full realization of these benefits. Although many authors have reported on DQ related problems, attaining and sustaining DQ in CIS has been a multi-faceted and elusive goal. The current literature on DQ in health informatics consists mainly of empirical studies and practitioners' reports. Reports often focus on a particular issue or quality attribute but lack a holistic approach to addressing DQ 'by-design'. This paper seeks to present a general framework for clinical DQ, which blends engineering theories with concepts and methods from health informatics. We define a new architectural viewpoint for designing and reasoning about DQ in CIS. We also introduce the notion of DQ Probes for monitoring and assuring DQ during system operation. Finally, we validate our framework with a real-world application case study.