This document contains early notes on an analytic approach to content modelling, which is to say, a structured approach to modelling health information content using formal tools and resources such as archetypes and terminology.
The starting point is to identify the following potentially competing needs:
- In an HIS, e.g. the EHR, clinical ‘documents’ etc: data that can be reliably processed and queried, so that clinical safely is always ensured.
- In content modelling of those data: flexibility and re-use, while ensuring semantic coherence in the data. Note that we take ‘content modelling’ to mean the creation of models that not only formally describe data content, but constitute a formal basis for building queries on the same data.
1) requires a theory of how to model EHR and other health data. We had a pretty solid theory on this in openEHR (described here in Architecture Overview (sections 4 & 6) the EHR specification (sections 5 & 8). This is now a decade old. We have learned many lessons from openEHR, and also other EHR and related modelling frameworks such as HL7v3, CDA, ISO 13606, ASTM CCR, Intermountain Healthcare’s health computing environment, UK GP systems, the discussions in CIMI, and many others. Some outlines for an improved theory of health data, for the moment still limited to EHR- and clinical document-like environments are here.
2) is the subject of this paper, and is aimed at establishing rules, guidelines and formal approaches to building models of content which ensure / support 1), and which additionally enable re-use, model verifiability, lifecycle management and other useful functions.