Pointers to Navigate Between a Graph Database and a Clinical Data Warehouse

Use of graph databases for structuring and leveraging complex healthcare data has been growing. The approach taken at the Geneva University Hospitals is a graph database which does not directly contain patient data, but rather a representation of data using semantic standards. This allows for a full view of patient level information for answering clinical questions and discovery, without containing sensitive information. However, a link must be kept to the clinical data warehouse to retrieve the information discovered using the graph. For this, pointers created using the JMESPath query language have been created to link the expressions in the graph database to the source data in the clinical data warehouse.