Einsatz und Evaluierung eines evolutionären IT-Konzepts für ein integriertes klinisches Informationssystem

For a clinical center it is increasingly important to successfully use information technology (IT). However, IT systems in a clinical environment are part of a complex socio-technical system. They only can be enduringly successful if they closely follow their users’ needs and if they are able to rapidly react on volatile requirements. Since 1997, an IT concept for the introduction of major parts of the Marburg health information system (HIS) has been elaborated. The concept is described in the first part of this thesis. Initially, due to time and cost restrictions, a so called "big bang" approach had to be chosen. Important base functionality had to be introduced rapidly throughout the clinical center. The "big bang" approach is associated with a high project risk. In order to minimize this risk, the "big bang" phase was limited as far as possible. For the further development of the HIS an incremental approach for demand-driven software evolution has been used. This approach shall allow to closely involve the users and rapidly and effectively implement changed requirements. The aim of this thesis is to assess the described approach concerning its success and sustainability when applied in practice. For this purpose an evaluation concept was elaborated starting from the fundamental IT success factors. It combines different qualitative and quantitative methods: measuring indicators of HIS development (extent of infrastructure, development and support times, system usage, and system response times), quantitatively comparing user satisfaction before and after using the incremental approach, interviewing users, as well as continuously monitoring process and documentation quality, and usability as part of software development. These studies were performed from 2000 to 2005, accompanying clinical parts of the HIS project. To evaluate IT interventions within an organization is closely intertwined with application development and user training. Therefore, a number of social and organizational factors aggravated the evaluation of the IT concept. These obstacles and possible weaknesses of the deployed methods are described and discussed. The evaluation results largely confirm the hypotheses. The "big bang" approach showed the expected drawbacks: functional deficiencies, insufficient alignment to processes and users’ needs, and moderate user satisfaction. The situation improved measurably when the incremental approach was used for the further development of the HIS. Combining these approaches proved to be feasible. With limited resources it was possible to build a HIS that offers broad hospital-wide clinical functionality and that has the advantages of an integrated system as well as enough flexibility and adaptability. This is also reflected in an acceptable user satisfaction, which increased significantly when using the incremental approach. Moreover, only the scores of those satisfaction factors increased that are assumed to be causally influenced by a participatory demand-driven software engineering process. The comparison of results from different analyses (user satisfaction, system response times, interviews, and user feedback) seems to confirm the assumption that these positive effects can be explained to a large extent by the applied model of software engineering and are not caused by possible confounders like higher system performance or higher commitment of IT staff. The experiences from clinical parts of the HIS project also show that effective and demand-driven IT support for clinical processes can be achieved on the basis of the Marburg IT concept. Furthermore, potential negative impacts of an IT intervention can be rapidly identified and eliminated by continuously monitoring quality indicators as part of an iterative participative software engineering process that is embedded into a comprehensive change management process. Finally, shortcomings of the applied tool are discussed and international research activities on multi-layered service-oriented architectures are described briefly with respect to the Marburg IT concept. Such architectures promise to enable adaptable and flexible IT support for interorganizational processes in health care networks.