Date of Award


Document Type



Engineering & Computer Science

First Advisor

Yuanlin Zhang

Second Advisor

Michael Gelfond


Clinical practice guidelines (CPG) describe recommended actions for diagnosis and treatment of various patient conditions. These guidelines are most often presented in a narrative form, requiring time from a physician’s already busy schedule and careful study, considering the guidelines may contain poor organization and lack clear, descriptive evidence for recommendations. Too often, this means that the information provided by guideline authors is ignored in clinical practice. Over the past few decades, much effort has gone into translating clinical practice guidelines into clinical-decision support systems to make guideline information more accessible and improve physician-patient interactions.

To contribute to physicians’ accessibility of guideline information, we attempted to develop a methodology to represent clinical practice guidelines as computer-implementable guidelines (CIG) with declarative programming. There are many obstacles in this implementation, such as underspecified conditions for recommendations, lack of knowledge and consensus in several areas, and heavy use of ambiguous terms. We report the measures we took to counter each of these issues, which allowed us to ultimately produce several models that could serve as computer-implementable guidelines for use in clinical practice. Through close analysis of our guideline implementation process, we hope to recognize patterns of knowledge and issues in the medical domain that will ease future clinical practice guideline implementation.