Guideline and Workflow Models
In: "Medical Decision-Making: Computational Approaches to Achieving Healthcare Quality and Safety", Robert A. Greenes (ed.), Elsevier/Academic Press, 2006
Clinical guidelines aim to improve quality of care, decrease unjustified practice variations, and save costs. In order for guidelines to affect clinicians' behavior, they should provide patient-specific decision support during patient encounters. Specifying guidelines in computer-interpretable guideline (CIG) formalisms that could provide automatic inference based on patient data may achieve this goal. The knowledge contained in guidelines is difficult to formalize due to the fact that despite efforts made to improve the quality of narrative guidelines, evidence-based recommendations are often incomplete and vague, and do not constitute a full care process. Several methodologies have been developed to support the transition from narrative guidelines into CIG implementations. They include (1) methodologies for marking-up narrative guideline elements in order to assess a guideline's quality and completeness and map it to CIG formalisms and (2) CIG formalisms. Many CIG formalisms exist, differing in their goals, computation model, the elements used to structure guideline knowledge, and the degree to which they support workflow integration. Specifying a narrative guideline as a CIG is a difficult task, yet the resulting application cannot be easily shared by different institutions and software systems. Therefore, sharing encoded knowledge is a challenging goal. The specification of standard methods to support such sharing is a major focus in the field. The road to achieving wide-spread use of guideline-based decision-support systems is long and difficult. This chapter reviews the current state-of-the-art in guideline-based decision support research and considers likely future directions that can be taken to reach the ultimate goal.