The Clinical Data Model and FHIR
The Clinical data model is built to align with the Fast Healthcare Interoperability
Resources (FHIR) specification, which is an open API standard for the exchange of healthcare
information. The FHIR standard is defined by Health Level 7 (HL7), the organization that also
defined the HL7 messaging standard.
|
Available in: Lightning Experience Available in: Enterprise and Unlimited Editions with Health Cloud |
The Clinical data model is the successor to the older packaged EHR data model. The standard objects in the Clinical data model store all sorts of clinical data, including conditions, medications, encounters, immunizations, procedures, and so on. The Health Cloud implementation of FHIR aligns with FHIR v4.0.1.
The FHIR-Aligned Clinical Data Model
Health Cloud’s new clinical data model maps closely to FHIR R4 specification, but with a few minor differences.
- The new clinical data model supports most of the FHIR R4 attributes, except certain instances that aren’t relevant for the typical Health Cloud user.
- Entities like Identifier and Code Set have more attributes than their FHIR R4 counterparts. These extra fields increase data usability across the Salesforce platform and support specific needs of certain features in Health Cloud.
- FHIR R4 typically defines zero-to-one, zero-to-many, one-to-one, or one-to-many values for each attribute. However, the Salesforce platform doesn’t support one-to-many and zero-to-many values for a field, except in multi-select picklists. Instead, Salesforce uses child objects that can create multiple records that all reference the same parent record to achieve this type of logical zero/one-to-many values. For instance, an encounter can have multiple providers participating in it. To support this scenario, Salesforce has the Clinical Encounter Provider child object. This child object has records representing each provider in an encounter, and each of these records reference the record of the encounter that these providers were a part of.
Differences Between the Packaged EHR Data Model and the Clinical Data Model
The older EHR data model differs from the FHIR-aligned Clinical data model in six significant ways.
- The EHR data model is based on FHIR R1 standards, while the new clinical data model is based on the latest FHIR release, FHIR R4. The new clinical model is backwards compatible. Also, many resources in FHIR R4 have reached a normative stage of development, which means that the FHIR R4 standard is now stable and can be widely adopted.
- The EHR data model is built on the Health Cloud managed package, while the new clinical data model is built entirely on the core Salesforce platform with standard objects.
- In the packaged EHR data model, almost all EHR entities have a Lookup to Account. To share data related to a patient, the user needs separate access to EHR records belonging to that patient. In the FHIR-aligned clinical data model, entities have a Master-Detail relationship to the Account record, which allows multiple sharing settings for child entities without the use of extra permissions.
- The existing EHR data model doesn’t support Code Set resources as defined by FHIR standards. Data related to these resources was captured as plain text fields, making it difficult to upsert patient data. The FHIR-aligned Clinical data model uses the Code Set and Code Set Bundle objects that can store codes, which makes it easy to upsert patient data.
- The EHR data model has a highly denormalized structure, which can lead to data inconsistency because the same data is stored in multiple records. When data is inconsistent, updating and upserting data become more expensive. By contrast, there’s less data overlap with the new clinical data model because it uses references to store data.
- The records in the packaged EHR data model are view only, which means that only the external source system can be the source of truth. But the Clinical data model doesn’t have this restriction, so that Health Cloud can be a source of truth.