Publishers and mapping to the common model

Publishers are defined as any provider system, or provider/ system/version combination that sends data to the Voror Health Platform (VHP) or provides an API for the VHP to use to obtain data.

Some providers have systems that have a common mechanism of publishing data. Typical examples of this are GP system suppliers such as EMIS, TPP or Vision. In this case, the transformation and mapping can be standardised and used for any publisher.

Some providers have systems that do not have a mechanism of publishing data, or where they do, a particular provider may not have purchased it. In this case, a transformation and mapping exercise has to be undertaken specifically for that provider. In many cases that also means a variety of source systems are involved.

These articles summarise the necessary transformations and mappings needed for each provider system.

General approach

The initial transformation in the VHP is to an extended FHIR format i.e. a FHIR model with extensions, the extensions to accomodate the many dozens of properties and data types for particular specialist systems or specialties within providers. From the interim FHIR format, a further structural and semantic mapping occurs to a relational data model, stored in a form that can be efficiently queried and retrieved. This can be illustrated as follows:

High level (non-technical) overview

The data is received from the publisher (various different publishers, systems, and versions) and is sent in various different formats by various different mechanisms.
The gateway controls the IG and flow of data into the VHP file store and inbound/publisher queue.
The inbound queue reader (transform) is notified that there is new data and takes the new data, plus any existing FHIR mapping, transforms it, stores it in the FHIR store, and then informs the outbound/subscriber queue.
The outbound queue reader (subscriber) is notified that there is new FHIR data and takes that new data, plus any required data manipulation ( ontology, mapping, data modeling) and makes it available in a relational data store.