FHIR Shorthand (FSH)

Define FHIR profiles and extensions using a concise, human-readable syntax. FSH allows authors to specify constraints, cardinality, and terminology bindings directly on elements without manually editing complex XML or JSON StructureDefinitions.

• Differential approach: Only specify the changes from a base resource
• Element path syntax: Use dot-notation like name.family to target nested elements
• Compiled output: FSH compiles to valid FHIR Shorthand via SUSHI, the reference compiler

Embed formal constraints and validation logic directly within profile definitions. FSH supports FHIRPath expressions to define co-occurrence rules, conditional required fields, and cross-element validation.

• Invariant keyword: Define named rules with severity levels (error/warning)
• Context specification: Target specific elements or the entire resource
• Natural language descriptions: Pair machine-readable logic with human-readable explanations for implementation guidance

Create sample FHIR resources for testing and documentation using the Instance keyword. These instances serve as realistic test data that conform to defined profiles, accelerating implementation validation.

• Assignment syntax: Set element values using a clean key = value format
• Reference resolution: Define inter-resource links that compile into valid FHIR references
• Usage markers: Tag instances as #example or #definition to control their role in the Implementation Guide

Author terminology artifacts natively in FSH without switching to external tooling. Define custom code systems and compose value sets using inclusion and exclusion rules.

• Hierarchical codes: Define codes with parent-child relationships using indentation
• Intensional and extensional: Compose value sets by enumerating codes or by filtering code system properties
• External imports: Reference codes from established terminologies like SNOMED CT and LOINC

Orchestrate a complete FHIR Implementation Guide from a single configuration file. FSH specifies the canonical URL, dependencies, pages, and artifact groupings that SUSHI compiles into a publishable IG.

• Dependency management: Declare external IGs like US Core as dependencies
• Page generation: Author narrative content in Markdown that integrates with generated artifacts
• Parameterized builds: Define IG-wide parameters for consistent artifact naming and versioning

Reduce repetition and improve maintainability through aliases and mixins. Aliases assign short names to long URLs, while mixins define reusable sets of rules that can be applied across multiple profiles.

• Alias syntax: Alias: SCT = http://snomed.info/sct for cleaner terminology references
• Mixin composition: Define a common constraint pattern once and insert it into multiple profiles
• RuleSet insertion: Apply predefined rule blocks to ensure consistent profiling patterns across an IG

A constrained subset of a base FHIR resource tailored to specific requirements. In FSH, profiles are declared with the Profile keyword and refined using rules. Common rule types:

• Cardinality rules: * status 1..1 to make an element required
• Binding rules: * valueCode from http://example.org/ValueSet (required)
• Must Support flags: * subject MS
• Slicing: * component ^slicing.discriminator.type = #pattern FSH profiles compile into StructureDefinition resources with both differential and snapshot views.

A mechanism for adding custom data elements to standard resources. In FSH, extensions are authored using the Extension keyword and can be simple (single value) or complex (nested sub-extensions). Key concepts:

• Extensions are defined on a specific context (resource type or element path)
• Use * value[x] only CodeableConcept to constrain the value type
• Modifier extensions use * ^modifierExtension = true
• Extensions compile into standalone StructureDefinitions with type: Extension FSH makes extension authoring declarative rather than hand-editing complex JSON differentials.

A domain-specific language for defining transformation rules that convert legacy healthcare data formats (HL7 v2, CDA) into FHIR resources. While distinct from FSH, both are text-based authoring tools in the FHIR ecosystem. The Mapping Language focuses on:

• Element-by-element mapping from source to target
• Nested structures for complex transformations
• Conditional logic for context-dependent mapping
• Integration with StructureMap resources for executable transforms Together with FSH, these languages form a complete text-based toolchain for FHIR interoperability engineering.