An Architectural Model for Extracting FHIR Resources from CDA Documents

Abstract

In the last decades, many standards, such as HL7 V2 and V3, have been proposed for the encoding of clinical documents with the purpose of ensuring syntactic and semantic interoperability among heterogeneous health information systems. Recently, HL7 has defined the new standard FHIR in response to limitations existing in HL7 V2 and V3. In Italy, CDA 2 is the reference standard for the storage of clinical documents. In order to encourage FHIR widespread getting its benefits, we propose an architectural model able to (i) receive a query about clinical data contained in a CDA document, (ii) identify required information, and (iii) present it as a FHIR resource, on the basis of several CDA to FHIR mapping schemas. We focus on the information extraction from the Italian Patient Summary, a collection of the patient's most significant clinical data. This choice emphasizes the advantages of using FHIR standard due to the possibility of extracting granular information of clinical interest from a full document.