Disease specific ontology-guided rule engine and machine learning for enhanced critical care decision support

A method for creation, enhancement and use of a disease-specific ontology for computer assisted clinical decision support in critical care settings, the method comprising:

• capturing clinical knowledge of experts to create a core Vmap using a graphical user interface and an ontology editor;
• creating a disease specific ontological model Vmap comprising of complex expert rule-based reasoning;
• extending the core Vmap with disease-specific reasoning concepts from the disease specific ontological model Vmap to create a disease specific extended core map;
• extending the disease specific extended core map with existing medical ontologies and mappings that associate disease specific concepts with EMR data to create a disease specific extended core map;
• translating the disease specific extended core map and existing medical ontologies into OWL and rule artifacts to create disease-specific executable ontologies;
• ingesting raw EMR data as historic population level batch files and patient-specific real-time data streams;
• normalizing and mapping the raw EMR data to the disease-specific executable ontologies to create normalized data using semantic mapping and proprietary and disease-specific medical natural language processing engine;
• using the disease-specific executable ontologies to enrich the normalized coded and free-text data to create semantically enriched EMR data;
• using the disease-specific ontology associated rules to semantically characterize the enriched data in the context of a specific disease;
• outputting disease-specific and semantically characterized clinical data and labels to be used in a semantically enhanced CDS rule engine using real-time data in real-time mode to trigger alerts:
• using semantically enhanced retrospective data as training data for a predictive machine learning process;
• and
• using both semantically enhanced rule-based and machine learning based models in a CDS that support dual reasoning modes of clinicians.