SNOMED implementers must develop a clear strategy for context representation, considering the requirements of existing data models, cultural preferences for data entry, and the need to align with national standards. This strategy will address the overlap between the information model and the terminology model, defining which data elements will be represented in each component of the health information system. As discussed in earlier chapters, these decisions will also influence the design of terminology bindings and user interfaces.

Context Representation Options

1. Representing Context in the Information Model

   In this approach, terminology bindings are designed to include only context-free concepts from SNOMED CT. For example:

   • Diagnosis/Problems: Bind to << 404684003 |Clinical finding (finding)|
   • Procedures: Bind to << 71388002 |Procedure (procedure)|

   Contextual information is captured in specific fields of the information model, using SNOMED concepts that represent context attributes such as:

   • Finding context: Describes the nature of the finding (e.g., suspected, confirmed)
   • Subject relationship context: Indicates the subject's relationship to the patient (e.g., family member)
   • Temporal context: Specifies the time frame (e.g., past, present)

2. Representing Context in the Terminology

   In this approach, the terminology bindings are broader, including both context-free and context-specific concepts:

   • Diagnosis/Problems: Bind to << 404684003 |Clinical finding (finding)| or << 413350009 |Finding with explicit context (situation)|
   • Procedures: Bind to << 71388002 |Procedure (procedure)| or << 129125009 |Procedure with explicit context (situation)|

   This strategy minimizes the use of ad-hoc fields for context values, focusing on utilizing the available SNOMED CT concepts to represent both the primary concept and its context. Ad-hoc fields may still be used for context information not covered by the SNOMED CT Concept Model.

Pros and Cons of Each Strategy

Representing Context in the Information Model:

• Pros:
  • Greater coverage: Not all possible combinations of findings/procedures and context values are represented in SNOMED CT terminology.
  • Flexibility: Allows for more dynamic representation and combination of concepts and context.
• Cons:
  • Increased development effort: Requires additional UI development to manage context fields.
  • Risk of higher data entry effort: Without careful UI design, clinical users may encounter more clicks and decisions, leading to potential workflow inefficiencies.

Representing Context in the Terminology:

• Pros:

  • Simpler and more intuitive text search: Pre-coordinated concepts (concepts combined with context) can be easily searched and understood.
  • Efficient data entry: Reduces the need for additional context fields, streamlining the user's data-entry process.
  • Expandability: Terminology can be extended with new concepts to represent emerging combinations of clinical findings/procedures and contexts. Post-coordination can be implemented to store complex concept-context combinations as expressions.

• Cons:

  • Limited coverage: This may not capture all possible context variations due to the finite set of pre-coordinated concepts available.
  • Potential for complexity: Requires careful maintenance and updating of the terminology to accommodate new contexts and concepts.
  • Harder to adhere to standardized information models like FHIR.

Both strategies have their advantages and challenges. The choice between them should be guided by the specific requirements of the implementation context, the capabilities of the software tools, and the needs of the end users. Balancing technical feasibility with user experience will ensure effective context representation in SNOMED CT implementations.