A Pattern-based Analysis of Clinical Computer-Interpretable Guideline Modeling Languages

Nataliya Mulyar, Wil M.P. van der Aalst, Mor Peleg

Journal of the American Medical Informatics Association

 

Abstract

Objective: A multitude of languages used for specification of Computer Interpretable Guidelines (CIG's) differ in their approaches to addressing particular modeling challenges. This paper addresses the lack of well defined semantics of such languages and absence of a single standard for their evaluation. We evaluated clinical guidelines modeling languages using the Workflow Patterns as an analysis framework. The analysis provided for clinical guidelines aims at an unbiased and vendor-independent survey of the expressive power of some modeling languages for representing the control-flow aspect of clinical guidelines.

Design: The pattern-based analysis has been applied to Asbru, EON, GLIF, and PROforma. We focused on control-flow and left other perspectives out of consideration.

Measurements: We evaluated the selected CIG modeling languages and identified their degree of support of the control-flow patterns.

Results: The results from the survey could be used for the selection of a language for representing a guideline, as well as for motivation and input to further development of any of the surveyed languages.

Conclusion: The evaluation shows that CIG modeling languages are remarkably close to traditional workflow languages. Moreover, as our paper shows, our analysis could help in the improvement and clarification of existing languages in this domain.

Keywords: Workflow Patterns, Evaluation, Clinical Guidelines, Modeling languages