Cypher query:
Show:
Show:
Key
Patient
Event
Diagnosis (ICD10)
Procedure (OPCS4)
This graph demonstrates how knowledge graphs - heterogeneous graphs of entities linked by semantically meaningful relations - can provide a natural and intuitive representation for a patient's pathway.
KGs also allow more complex contextual data to be integrated - for example, a medical procedure is not just a single code, but has its place in a hierarchy of codes that forms a natural ontology. So patient similarity could be identified where two patients might undergo similar, but not identically coded procedures (In the future, we will try to further refine this and add horizontal disease-disease links, for example where two very different diseases commonly co-occur). The diagnosis codes could easily be expanded into a hierarchy in much the same way, and in the future we will aim to integrate further information such as prescriptions, imaging, and testing, with contextual information from their specific domain ontologies.
KGs also allow more complex contextual data to be integrated - for example, a medical procedure is not just a single code, but has its place in a hierarchy of codes that forms a natural ontology. So patient similarity could be identified where two patients might undergo similar, but not identically coded procedures (In the future, we will try to further refine this and add horizontal disease-disease links, for example where two very different diseases commonly co-occur). The diagnosis codes could easily be expanded into a hierarchy in much the same way, and in the future we will aim to integrate further information such as prescriptions, imaging, and testing, with contextual information from their specific domain ontologies.