Business modeling tools for managing decision support systems.

Abstract

We sought to investigate the role of business modeling tools in facilitating the implementation of interactive decision support systems (DSS) for laboratory test ordering. We have recently shown that Ripple Down Rules (RDR), a new strategy for building expert systems (ES), enables institutions to build DSS maintained by local experts without knowledge engineering support. While DSS hold great promise for improving health outcomes by providing doctors with timely access to care guidelines, implementation will require careful planning and change management. The Situated, Strategic, and AI-Enhanced method was used to define and assess a DSS-driven process change. DECmodelTM, a graphical business modeling application, was used to build representations of current (paper-based) and future (ES-supported) test ordering environments. Models described both the interactions between new processes and the impact on healthcare quality of ES-supported DSS. Models were able to represent the key processes involved in test ordering and the impact of DSS. Users could interact with the model to examine the time and cost implications of the transition from current to future states. The level of detail in which processes could be examined was readily controlled by the user. We have shown that graphical business modeling applications are useful tools for enabling institutions to view the impact of DSS. Accuracy of models should be maximized in workshops involving key management personnel. Whilst the impact of modeling on the cost and quality of DSS implementation remains to be established, we believe these tools are valuable for institutions planning to align DSS technologies to their service needs.