This paper presents a method for formally representing Computer-Interpretable Guidelines. It allows for combining them with knowledge from several sources to better detect potential interactions within multimorbidity cases, coping with possibly conflicting pieces of evidence coming from clinical studies. The originality of our approach is on the capacity to analyse combinations of more than two recommendations, which is useful, for instance, for polypharmacy interactions cases. We defined general models to express evidence as causation beliefs and designed general rules for detecting interactions (e.g., conflicts, alternatives, etc.) enriched with Linked Open Data (e.g. Drugbank, Sider). In particular we show that Linked Open Data sources enable us to detect (suspected) interactions among multiple drugs due to polypharmacy. We evaluate our approach in a scenario where three different clinical guidelines (Osteoarthritis, Diabetes, and Hypertension) are combined. We demonstrate the capability of this approach for detecting several potential conflicts between the recommendations and find alternatives.
CITATION STYLE
Zamborlini, V., Hoekstra, R., Silveira, M. D., Pruski, C., Teije, A. T., & Harmelen, F. V. (2017). Generalizing the detection of clinical guideline interactions enhanced with LOD. In Communications in Computer and Information Science (Vol. 690, pp. 360–386). Springer Verlag. https://doi.org/10.1007/978-3-319-54717-6_20
Mendeley helps you to discover research relevant for your work.