FHIR Profiling Guidelines for FHIR R4

Uit informatiestandaarden
Ga naar: navigatie, zoeken


Inhoud

1 Introduction

This page lays out the conventions for creating FHIR profiles and associated conformance resources within Nictiz. These guidelines are specifically aimed at FHIR R4; for the STU3 profiling work, a separate document exists.

This page is titled "profiling guidelines", but actually addresses all conformance resources (profiles, extensions, value sets, code systems, CapabilityStatements) and associated examples. We use these terms somewhat interchangeable throughout this document; 'profile' can usually be read as 'the whole set of conformance resources'.

1.1 Language

Unless stated otherwise, the FHIR conformance materials will be created in English in order to encourage adoption and re-use.

2 Open vs. closed world modeling

When profiling, an "open world" or a "closed world" model can be chosen. The former means that the profile only allows the elements to be used specified by the functional model, with all the restrictions from the functional model. The latter means that the profile can accommodate the elements specified by the functional model, but doesn't impose further restrictions.

We adopt the "open world" modeling approach to aid re-usability beyond the known use cases. When restrictions are deemed necessary for a specific use case, it will be added to the information standard specific profiles.

× Open Closed
Pros
  • Forward compatibility
  • Modelers don't have to think about what you shouldn't support, only what must be supported
  • Implementers can fit more data, even if it's not in specified explicitly by the profile
  • Implementers, don't have to support all elements that maybe, someday could be used, according to the model
  • Model becomes more specific
  • Model becomes smaller and more straightforward
  • More implementer feedback, about elements they want to support, but currently can't
Cons
  • Implementers, have to support all optional elements that maybe, someday could be used, according to the model
  • Model becomes more vague,
  • Model becomes larger and less straightforward about what should actually be supported, and what can optionally be supported
  • Less implementer feedback: elements they want to send can be easier 'hacked' in a not yet explicitly specified element. Model won't be improved.
  • More versions of models, after more elements have to be supported
  • No forward compatibility, only backwards
  • Implementers have to wait for a new version of the model, if they want to support elements, that are currently not in scope.

2.1 General guidelines for open world modeling

We only profile elements, cardinalities and bindings that require profiling. We leave other elements, cardinalities and bindings as-is.

2.2 Cardinality

The functional description will specify the cardinality for each concept as a minimum required and maximum allowed number of times it may occur, which is the same mechanism as in FHIR. However, one needs to be careful as the cardinality can only be restricted in derived profiles, and never widened. Being too strict could thus hinder the re-use of these profiles. This is especially true for cardinalities in zibs, which should be interpreted as 'purely conceptual'; a use case might allow for data that conceptually always should be there to be absent in practice.

For zib profiles:

    • A minimum of 0 or 1 will be profiled as 0.
    • A maximum (1, n, *) will be profiled as-is.

For nl-core profiles:

  • No further restrictions are added, as these profiles only cover the currently known use cases. Adding restrictions here would hinder as yet unknown of undefined use cases.

For information standard specific profiles:

  • If the corresponding zib has a minimum of 1 and the use case doesn't contradict this, the minimum will be profiled here as 1.
  • Cardinalities may be further restricted if the use case defines this.

2.3 mustSupport

The basic notion for any element or extension in a profile, most specifically when it has a mapping to a functional definition, is: provide what you have, and support what you can. At the generic profiling levels - zib and nl-core - we can not know what the exact context of use will be. We can therefore never be sure that a formal FHIR mustSupport flag by any definition would hold in every imaginable use case.

For zib and nl-core profiles:

  • Do not use mustSupport. Implementers should understand the presence of a mapping as a hint that a semantic understanding exists and should interpret this as a generic encouragement to support this element.

For information standard specific profiles:

  • Both constrain cardinality and apply mustSupport as applicable for the information standard. Implementers have generally pushed back on too many profiles for the same building blocks so we generally do not go into that level of profiling.

The way to know the exact requirements for the use case at hand is thus not immediately clear from just looking at the generic profiles, but from reading the transaction that should be part of any information standard. A transaction is the dataset constrained with cardinalities and conformance applied, comparable to what FHIR calls a logical model.

2.4 Constraining references

Concepts in a functional description oftentimes refer to each other; e.g. most clinical concepts will refer to the concept of a Patient. In FHIR these connections are realized using the reference datatype, which allow to specify the target (base resource of profile) as well. For example, a profile representing the zib Problem could set the target of the .subject element to the profile representing the zib Patient.

However, setting the target explicitly means a restriction of the allowed targets, which runs counter to the principles of open world modeling. On the other hand, specifying a dedicated profile provides useful guidance on how to handle this profile. To address both concerns, the target profile will be added next to the base resource. For example, the Condition resource allows Patient and Group as target in the .subject element. When profiling the zib Problem, the profile representing the zib Patient will be added as a third option.

3 Functional model as base

Most, if not all, conformance resources are based on an underlying functional model. The functional model is the specification to which the conformance resources should adhere.

The basis for most other functional models is formed by the 'zibs' ('Zorginformatiebouwstenen'), in English also known as are Clinical Information Models (CIMs), Health and Care Information Models (HCIMs) or Clinical Building Blocks (CBB) – we will use the Dutch term 'zib' for all profiling work as it has become a recognizable term over the past years. The zibs are defined by the program ‘Registratie aan de bron’ (Data capture at the point of care) and provide a foundation of use case neutral building blocks from which use cases can be built. The formal definition of the zibs can be found on the zibs wiki.

Use cases or information standards use and refine those zibs that are relevant to the situation. The formal specification for these information standards can usually be found in Nictiz' instance of ART-DECOR.

3.1 Layering: zibs, nl-core profiles and information standard specific profiles

The profiles and other conformance resources align to this layering of information standards. We recognize three levels of profiles:

zib profiles
profiles that represent the zib as faithfully as possible
nl-core profiles
profiles derived from the zib profiles that that might be enriched by concepts from the different use cases that need the zib. See #The nl-core layer for more information.
information standard specific profiles
optional derived profiles from the nl-core profiles that further restrict or enhance these profiles for a specific use case.

3.2 The nl-core layer

The goal of the nl-core layer is to ensure that similar concepts defined at the use case level are implemented in a uniform way at the FHIR level. Whether the concept is used by a single or multiple use cases is irrelevant – adopting it at the intermediate nl-core level provides clear guidance to future use.

On the other hand, concepts defined at the intermediate level are only of value if they can be reused across situations. Therefore, it should be investigated how the concept can be created in a reusable way before it is implemented at the nl-core level. If there is a vanishingly small opportunity for reuse, then it is best to define it at the use case level to avoid clutter at the nl-core level. In practice, this means:

  • Extensions from use cases should usually be generalized and added to the nl-core layer.
  • Adding reference types to specific profiles (see #Constraining references) should usually be done at the nl-core layer.
  • Cardinalities are usually determined by the zibs and further restricted by the use cases, so it makes little sense to implement them at the nl-core layer. These should normally be implemented at the use case layer. Also see #Cardinality.
  • Mappings to the functional specifications for use cases (see #Associating the functional definition to StructureDefinitions) are by definition use case specific and should be added at the use case layer.
  • Bindings to ValueSets (see #ValueSet binding) are usually use case specific, but might be added at the nl-core layer.
  • SearchParameters should usually be generalized and added to the nl-core layer.

3.3 Associating the functional definition to StructureDefinitions

Any StructureDefinition that profiles a Resource does so because there is some kind of logical definition dictating how. Profiles SHALL have a traceable relationship with their functional counterpart(s).

The FHIR specification contains several solutions for this problem, ConceptMap (>=STU1), StructureMap (>=STU3), ElementDefinition.mapping. Neither of these solutions are mature and cover our use case.

  • ConceptMap: works best for value sets and codes.
  • StructureMap: overcomplicated and unclear how to apply and whether or not this has a future.
  • ElementDefinition.mapping: is a free text mapping inside the profile. This means we cannot add mappings to profiles from third parties without updating their resource.

3.3.1 Current implementation

We use the StructureDefinition.mapping and ElementDefinition.mapping elements in profiles to map functional concepts to resource elements. The following rules apply:

  • The mapping .mapping.uri SHALL resolve to a description of the functional model:
    • For zibs, this SHOULD be the English page on zibs.nl.
    • For functional models in ART-DECOR, this SHOULD be the link to the dataset on ART-DECOR v3 in the format of [host]/ad/#/[prefix]/datasets/dataset/[dataset id]/[dataset effectiveDate]
  • The .mapping.identity should be a constructed in the following way:
    • For zibs, this SHOULD be the zib-[English zib name, lowercased]-v[version]-[publication]EN
    • For functional models in ART-DECOR, this SHOULD be, all in lowercase, [project prefix][dataset name]-[dataset effectiveDate as 'yyyymmdd']. The dataset name SHOULD be the English name if available, otherwise it SHOULD be the Dutch name. All spaces are replaced by dashes. If the project prefix and dataset name are similar, it's better to not repeat it and simply use 'dataset'.
  • The .mapping.name should be constructed in the following way:
    • For zibs, this SHOULD be zib [English zib name]-v[version]([publication]EN)
    • For functional models in ART-DECOR, this SHOULD be ART-DECOR Dataset [dataset name] [dataset effectiveDate]. The dataset name SHOULD be the English name if available, otherwise it SHOULD be the Dutch name. If the dataset name contains the word 'dataset', it SHOULD not be repeated.
  • Functional concepts are referenced based on their id in the functional model, which should be used to populate .mapping.map.
  • The description of the functional concept is normally placed in .mapping.comment, unless other rules apply (see #References that are reversed in FHIR).
  • When slicing, the mapping is made on the content of the slice, not the slice itself (see also #Placing-Info).

So on the root of the StructureDefinition, the mapping for a zib should be defined as:

<mapping>
    <identity value="zib-medicationagreement-v1.2-2020EN" />
    <uri value="https://zibs.nl/wiki/MedicationAgreement-v1.2(2020EN)" />
    <name value="zib MedicationAgreement-v1.2(2020EN)" />
</mapping>

A specific element can then be mapped using:

<element id="MedicationRequest.extension:usageDuration">
    <path value="MedicationRequest.extension" />
    <mapping>
        <identity value="zib-medicationagreement-v1.2-2020EN" />
        <map value="NL-CM:9.6.19936" />
        <comment value="PeriodOfUse" />
    </mapping>
</element>

Note that implicitly mapped elements and reversed references use a slightly different comment.

3.4 Handling errors, proposed changes and omissions in the functional model

Conformance resources which are based on a functional model SHOULD faithfully represent the functional model. This implies that errors in the functional model cannot be fixed in the FHIR implementation unless it's really trivial; doing so would lead to a difference between the FHIR implementation and the functional model, resulting in unclear and hard to use specifications.

Practically, this results in the following guidelines:

  • If the error is fixed on the functional level, it is adopted in the conformance resources.
  • If an error is trivial AND recognized as something that will be fixed on the functional level, it can be adopted (e.g. spelling mistakes).
  • In any other case, the error will be implemented as-is in the conformance resources. This includes the following cases:
    • The error is not recognized as something that needs to be fixed.
    • The error is recognized as something that needs to be fixed, but the fix has to wait until the next version.

    The error is documented using a hyperlink to the ticket in the suitable element, usually the .comment field of the relevant profile element.

    Note: if the error prevents the creation of the conformance resources, then the functional model is deemed unimplementable and no conformance resources are created.

Omissions in the functional model are handled like any other error, with one important exception: if extra information is needed in order to create an implementation (like a code for a functional concept), and this information will be added to a future version of the functional model, it makes sense to already adopt it for the current implementation. This addition is documented with a reference to the relevant ticket.

3.4.1 For zibs specifically

For zibs specifically, these guidelines are practically fulfilled in the following way.

  • For any error found in the functional model, a BITS ticket is created in the ZIB project.
    • If the ticket results in an erratum, the fix is adopted without notice.
    • If the error is not recognized, or won't be fixed until the next (pre-)release:
      • If the conformance resources can be made, the error is documented with a hyperlink to the BITS ticket in the suitable element, usually the .comment field of the relevant profile element.
      • If the conformance resources cannot be made, this is documented in the release notes using a hyperlink to the BITS ticket.
      • (Only) if the error pertains something trivial like a spelling mistake AND will be fixed in a future release, it is adopted without any notice.
  • For missing definition codes, a BITS ticket is created in the ZIB project.
    • If the tickets results in a definition code, it is implemented in the conformance resource. The .comment field of the element containing the code will document the code using the following remark: This code aligns with the future version of this zib. See [BITS ticket] for details., where [BITS ticket] should be a Markdown formatted hyperlink to the relevant ticket with the ticket key as title text.

4 Versioning

In general terms, FHIR conformance resources could be affected at several different layers:

  1. The version of the package that the conformance resources reside in: versioned according to SemVer 2.0.
  2. The version of the conformance resource themselves (StructureDefinition.version): used to indicate the business version to the user, without strict specifications.
  3. The FHIR version (StructureDefinition.fhirVersion): this document is specifically aimed at FHIR R4, meaning this element will be fixed on 4.x.
  4. The version of the underlying data model.

Regarding point 1 and 2: Nictiz uses the package level as the main versioning mechanism. As a result, the conformance resources within the package are not individually versioned; they should be regarded as a consistent set. To identify the package version a conformance resource, its version number will be set to the package version.

Regarding point 4: the life cycle of the underlying data model is not reflected directly in the version number of the conformance resources, but a change in de the underlying data could result in a change in one or more of the conformance resources. In this case, the normal SemVer rules will determine what happens; if some of the conformance resources need to be changed in a backwards compatible way, a new patch release of the package should be made, if major functionality is added, a new minor version of the package should be released, etc. When a new version of the underlying data model reflects a fundamental change, the choice can be made to create a new package under a different name rather than a new version (eg. each zib releases will have their own package).

Version updates of conformance resources normally do not affect their canonical URI. Any resource that references another resource normally does so without a version indicator (uri|version). Instead, this is handled at the package level; reference targets either reside within the same package or in a versioned package that has been added as a dependency.

5 Mapping semantic codes to profiles

Oftentimes a functional concept has an equivalent FHIR element, e.g. a 'comment' in the functional description maps in FHIR to the Resource.comment element. When this is not the case, a code (usually SNOMED or LOINC) is needed to add the proper definition to the FHIR resource, usually using the Resource.code element. For example, the root concept of an Observation is unknown unless it is defined by the .code element, and when components are used in an Observation, each Observation.component needs to be defined using an individual .code element. These codes should be provided by the functional description (for zibs this is the DefinitionCode).

These codes are profiled as required elements using a ElementDefinition.pattern as described in #Values that should be present verbatim.

6 Slicing

Elements are sliced for three reasons:

  1. To specify different requirements for repetitions of an element.
  2. To add the mapping to the functional specification to a specific instance of the element (even if all requirements are the same as the default slice).
  3. To allow other uses of a base element than what is required by the functional specification, e.g. the reuse of an .identifier next to the profiled .identifier.

Slicing can often occur on different levels within an element, e.g. when the element is of type CodeableConcept, a slice can be made on the element itself or on the child element of type Coding (.element.coding). In general, slicing will occur on the highest level possible, thus on the element itself in the example above. Note that this is only possible for repeating elements (max cardinality > 1). When an element of type CodeableConcept does not repeat, the Coding element can be sliced, but this is discouraged as this is often not what is meant; when .coding repeats within a single element of type CodeableConcept, it must have the same semantic meaning although some granularity difference is accepted (eg. when communicating the same concept in different code systems).

In general, the slice discriminator is set to discriminator.type = value and discriminator.path to the path where the discriminator value is located (if it is the root element, it should be noted as $this). The sliced element gets a pattern or a required ValueSet binding on the location of the path. No further cardinality constraints are added in the sliced element that correspond with the pattern[x] as they are made mandatory by the pattern. These general slicing guidelines apply to the following situations. See this page for examples in XML:

  • Mapping semantic codes to profiles requires the presence of the semantic code while leaving the option to use the element for other concepts, e.g the zib LaboratoryTestResult profile uses Observation.category for a semantic code that represents the zib and to map the zib concept ResultType.
  • Identifier systems are always defined on a sliced .identifier, even if there is only one defined.

Other forms of slicing that are common (Forge sets these slicing details by default):

  • A concept is mapped to one or multiple types of a polymorphic element, e.g. Observation.value[x]. The mapping and functional description are placed on the slice with the matching data type(s).
  • Adding an extension slices the .extension element that is discriminated by the value on extension.url.

6.1 Slice names

The slice name should describe the concept it represents. It should align with the name from the functional description where possible, but a more suitable name may be chosen if deemed more applicable:

  • E.g. the zib concept 'PatientIdentificationNumber' defines a patient id for Dutch patients in the form of a bsn, hence the slice name 'bsn' for a slice on Patient.identifier is more informative than the zib concept name.
  • For CodeSpecification extension slices, a custom rule applies.
  • When slicing by .targetProfile, a workaround may be needed resulting in extra additions to the slice names.

Slices follow the convention of FHIR elements in general, using a camelCased name starting with a lowercase letter. If the slice name represents or starts with an abbreviation, the entire abbreviation should be lowercased (e.g. 'bsn' instead of 'BSN').

6.2 Slicing by .targetProfile

A common pattern is to slice elements of datatype Reference by the .targetProfile that the reference acts on, using discriminator.type = profile and discriminator.path = resolve() (this is predominantly done to define the mapping to the functional model). There is a limitation in different FHIR tool stacks that prevent the correct handling of multiple .targetProfiles on the same slice. As a workaround, slices with multiple .targetProfiles should be split up. If this happens and the slice names would be the same according to the normal naming rules, the slice names will be appended with an appropriate suffix like the .id (thus the leaf part of the canonical URL) of the target profile.

7 Restricting the range of allowed values

7.1 ValueSet binding

7.1.1 General approach

As a general guideline, the ValueSet defined by the functional model should get a binding on the FHIR element representing the functional concept. This means:

7.1.2 Binding strength

The FHIR profiles should faithfully reproduce the binding strength from the functional description. It may not always be possible to adopt the binding strength verbatim. For example, when creating slices for ValueSets, the binding strength must be set to required in order to make the discriminator work, even though the binding strength in the functional description is extensible. However, by making the slicing open, conceptually the same result is achieved.

Also see the examples.

7.1.3 Binding multiple ValueSets

The functional model might require multiple ValueSets to be used for some data element, e.g. when offering the choice between multiple code systems. However, FHIR allows to bind just one ValueSet to an element. There are two strategies to handle this situation:

  1. Create and bind a combined ValueSet that encompasses all the ValueSets defined in the functional model.
  2. Create slices for each ValueSet binding.

The use of combined ValueSets is the preferred approach. Although the use of slices would make the different concepts visible in the profiles, this strategy doesn’t allow for overlapping ValueSets (which often doesn’t make much sense semantically but happens for example when the different ValueSets include the same NullFlavor codes). It is also less straightforward for derived profiles to expand or restrict the base ValueSet when using slices. Using the first approach, a new ValueSet can be created and bound (if the binding strength permits it and it makes sense semantically).

However, option 2 should be used if there are different requirements for the ValueSet bindings.

Also see the examples.

7.1.4 Handling conflicts with base bindings

FHIR base resource might specify a required binding for some of its elements, whereas the corresponding concept in the functional specification uses alternative or more specialized codes. To conform both to FHIR and to the functional description, a ConceptMap will be created to map the codes from the functional specification to the codes from the FHIR ValueSet. This ConceptMap will be linked to the existing ValueSet using the extension permitted-value-conceptmap extension.

If all mappings in the ConceptMap are equal or equivalent, no further action is needed.

7.1.4.1 Datatype code

When not all codes from the functional specification map cleanly to a required ValueSet from a FHIR base resource with datatype code:

  • The ValueSet from the functional description will be bound with the binding strength of the functional concept using the CodeSpecification extension. The extension slice is named after the full English name of the value set according to the functional description, in camelCased notation.
  • Mapping to the concepts of the functional description is applied both to the FHIR element and the extension.
  • The ConceptMap will be added using the extension permitted-value-conceptmap extension on the FHIR ValueSet.

7.1.4.2 Datatype CodeableConcept

When not all codes from the functional specification map cleanly to a required ValueSet from a FHIR base resource with datatype CodeableConcept:

  • CodeableConcept.coding will be sliced with a discriminator of value/$this, based on a required binding on the CodeableConcept.coding level.
  • A slice will be created that binds the ValueSet of the functional description:
    • The slice is named after the full English name of the value set according to the functional description, in camelCased notation.
    • The binding strength is set to required. This is needed in order to make a valid slice definition.
    • The minimum cardinality of the slice is set to 1; if the concept is included, both the FHIR and function model codes are required.
    • The ConceptMap is documented using the permitted-value-conceptmap extension on the ValueSet bound in this slice.
    • binding.description is set to: In addition to a coding from this ValueSet, the corresponding coding from the FHIR base ValueSet SHALL be communicated. The ConceptMap <[canonical of ConceptMap]> can be used to relate these two ValueSets.
  • Mapping to the concepts of the functional description is applied only to the FHIR element, not to the slice.

7.2 Values that should be present verbatim

On occasion the profile needs to ensure that an element in the resource being profiled has an exactly defined value. Common use cases for this:

  • A particular code must be present at the element (mostly to map a semantic code to a profile).
  • Only codes from a particular code system are allowed, so its .system can only have one value.
  • Only Quantity units in UCUM are allowed, so Quantity.system must be set to http://unitsofmeasure.org.
  • A boolean, string or other primitive type must have a particular value.

FHIR offers two competing ways to do this: fixed values (ElementDefinition.fixed) and patterns (ElementDefinition.pattern) (there are also less obvious ways to achieve this, like constraints or ValueSet bindings, which could be used when the solution calls for it). Fixed values have the advantage that they are more readable and are better at communicating to the implementer what the expectations are. However, it is a rather rigid mechanism as it states that the element must be exactly that value and nothing else. Oftentimes, the profiled value is a minimum expectation and other additional values are still allowed – for example, the datatype CodeableConcept allows for more than one .coding, as long as they are semantically compatible.

Therefore, in FHIR R4, the guidance is to use fixed values for primitive types and patterns for complex types. However, from FHIR R5 onward, the guidance is to always use patterns over fixed values, as even on primitive datatypes a fixed value would prevent the use of extensions or element ids.

Therefore, when there is a need for fixing values in the profile, the following guidance applies:

  • Use a pattern on the element representing the datatype (thus not on one of the elements within the datatype).
  • For coded elements where a fixed .system is required, including elements of datatype CodeableConcept/Coding and Quantity or one of its derived datatypes:
    • If the functional model specifies the fixed value as a value set (e.g. "all codes from SNOMED"), use a #ValueSet binding.
    • Otherwise, use a pattern with .system set to the required value.
  • Fixing values should only be done for data that needs to be present verbatim in the resource being profiled. Concepts like Coding.display and Quantity.unit should never be fixed, as these are descriptions which might change depending on the context.

8 Extensions

Sometimes a concept from the functional model cannot be implemented using the building blocks FHIR offers by default. In this case, an extension might be used to implement such concept. Keep in mind that extensions are often seen as a burden for implementers:

  1. If it possible to model the concept (cleanly) without an extension, this is usually the preferred way.
  2. If that's not possible, check if HL7 provides an extension to implements the concept.
  3. If that's not possible, try to create an extension in a reusable way (or reuse a previously defined extension).
  4. If that's not possible, create an extension specific for the profile.

Usually the mappings to the functional model, bindings to specific ValueSets and any functional descriptions will be added when the extension is used within a profile. When the extension pertains to a particular profile, this information SHALL be added to the StructureDefinition of the ValueSet.

8.1 Extensions containing a reference to a zib profile

Extensions on the zib level that contain a reference to a zib profile, have that zib profile placed as type.targetProfile on value[x] (valueReference). Since it is problematic to make derived extensions, no nl-core counterpart of the extension is created, but the extension is made generic for both the zib and nl-core layer (see also #Profiles, extensions and datatypes). The nl-core profile that uses this extension should treat the zib profile reference in the extension like any other zib profile references and restrict the targetProfile to the nl-core counterpart.

9 Common patterns

Sometimes different profiles use a common pattern, for example when references to a particular zib requires special guidance. Special care should be taken to ensure that this pattern is implemented in a consistent way. There are several approaches for this:

  • When all these profiles share a common base, a base profile can be defined that includes this pattern.
  • When the pattern is part of an extension, then it is applied in a common way by definition.
  • A datatype profile may be created that can be re-used across resource profiles (using type.profile). This datatype profile may include special guidance aimed at profilers. An example of this approach is described in #Referencing zib HealthProfessional.

10 Mandatory FHIR elements without a functional counterpart

On occasion, FHIR requires an element to be populated while the functional model doesn't specify it. For example, the Procedure resource has the required element Procedure.status, but the zib Procedure, which is mapped to this FHIR resource, doesn't specify such a field.

In these circumstances, an effort should be made to map the zib concepts implicitly to these elements. For example, FHIR Procedure.status can be inferred from the zib concept ProcedureEndDate; if it is in the past, then the status should be assumed to be completed. If it is not possible to make such an implicit mapping, the FHIR element is left as-is; the implementation guide should describe how the implementer should deal with it, probably using the data-absent-reason extension.

Implicit mappings are documented in the following way:

  • The comment of the implicitly mapped element should provide guidance on how to map the functional concept.
  • The mapping element is populated with the mapped concept, but the description is altered to: [concept name] (implicit, main mapping is on [element name]).
  • Other metadata (.definition, .short, .alias) isn't populated according to the normal guidelines.

11 References that are reversed in FHIR

Functional building blocks often refer each other, but on occasion, the logical direction for the reference in FHIR is in the other direction. For example, zib TextResult defines a reference to zib Procedure, but the association between FHIR resources DiagnosticReport (for zib TextResult) and Procedure (for zib Procedure) is defined on Procedure.report.

In these situations, the FHIR approach is followed rather than some custom extension or other mechanism. Although this approach cannot enforce the cardinality from the functional model, the use of the default implementation requires the least effort from implementers and gives the best results for interoperability.

These reversed mappings are profiled in the following way:

  • In the profile that contains the reference:
    • If the element that represents the association between the two resources is a repeating element, it is sliced with a discriminator of profile/resolve() and a slice is added with the target(s) set to the target profile(s). Otherwise, the target profile is just added to the list of targets.
    • The .mapping element is populated with the mapped concept, but the description is altered to: Reversed reference for [functional building block name].[concept name]"
    • The .short and slice name, if used, use the English name of the target rather than the concept name.
    • An .alias containing the Dutch name of the target is added.
    • The .comment contains an (additional) note in the following form: Please note that on a functional level, [functional building block A] references [functional building block B], but in FHIR this direction is reversed."
    • .definition can be populated with the (modified) definition from the functional building block, but only if it provides useful information to the implementer.
    • The cardinality of the element is unaffected.
  • In the target profile (where the reference would normally be made if the functional definition is followed):
    • The .comment on the root contains an (additional) note in the following form: Please note that on a functional level, [functional building block A] references [functional building block B], but in FHIR this direction is reversed. Therefore, the concept [concept name ]([concept identifier]) is mapped on [path to reference] in profile [profile B] instead of in this profile.

Example from profile zib-Procedure-event:

    <element id="Procedure.report">
      <path value="Procedure.report" />
      <slicing>
        <discriminator>
          <type value="profile" />
          <path value="resolve()" />
        </discriminator>
        <rules value="open" />
      </slicing>
    </element>
    <element id="Procedure.report:textResult">
      <path value="Procedure.report" />
      <sliceName value="textResult" />
      <short value="TextResult" />
      <comment value="Please note that on a functional level, zib TextResult references zib Procedure, but in FHIR this direction is reversed." />
      <alias value="TekstUitslag" />
      <type>
        <code value="Reference" />
        <targetProfile value="http://nictiz.nl/fhir/StructureDefinition/zib-TextResult" />
      </type>
      <mapping>
        <identity value="zib-textresult-v4.4-2020EN" />
        <map value="NL-CM:13.2.5" />
        <comment value="Reversed reference for zib TextResult.Procedure" />
      </mapping>
    </element>

And in zib-TextResult:

    <element id="DiagnosticReport">
      <path value="DiagnosticReport" />
      ...
      <comment value="Please note that on a functional level, zib TextResult references zib Procedure, but in FHIR this direction is reversed. Therefore the concept Procedure (NL-CM:13.2.5) is mapped on `Procedure.report:textResult` in profile zib-Procedure-event instead of in this profile." />
      ...

12 Practical guidelines

12.1 Identity of artifacts

12.1.1 Canonical URL, id, name and title

Conformance resources can have multiple types of identifying information, which are related at some level:

.url
The canonical URL, which is the external identifier for conformance resources. All conformance resources SHALL have a canonical URL. This URL is preferably resolvable but does not have to be processable. Canonical URL's are about the identity of artifacts, not necessarily about retrieval location. Canonical URLs aren't meant to be human recognizable.
.id
This should align with the latter part of the canonical URL.
.name
A recognizable name that is still computer processable.
.title
A recognizable title purely for human consumption.

URIs for the latter (for artifacts living in ART-DECOR) are described in ART-DECOR FHIR_URIs.

12.1.1.1 Profiles, extensions and datatypes

  • The .id will be constructed in the following way:
    • For profiles
      • representing a functional building block:
        • zib profiles:zib-[English zib name]
        • nl-core profiles:nl-core-[English zib name]
        • standard specific:[project prefix]-[concept name]
        • with optionally the following suffix:
          • If the functional building block is spread out over multiple resources, -[ResourceName] is appended, e.g. zib-HealthProfessional-Practitioner and zib-HealthProfessional-PractitionerRole. This suffix may be absent from the "main" resource, if there's one.
          • If there are multiple implementations of the same building block for different uses, a use specific suffix is appended. A common pattern for this is when zibs are used both for registered and planned events, resulting in an event and request resource from the FHIR workflow perspective (e.g. zib Procedure which is mapped to both a Procedure and ServiceRequest resource). The following suffixes may be used:
            • Past use: -event
            • Future use: -request
      • directly representing a particular concept of a functional building block: [profile id].[English concept name]. Underscores are ignored.
    • For extensions:
      • directly representing a specific concept in a single profile:ext-[English root concept name].[English concept name]. Underscores are ignored.
      • pertaining to multiple profiles, or not pertaining to specific profiles and generally applicable:
        • if the use context is a single resource:ext-[resource]-[purpose]
        • otherwise:ext-[purpose]
    • For datatype profiles:
      • representing a functional building block: this is regarded a normal profile as described above
      • representing a pattern: pattern-[purpose]
  • The canonical URL will then be created as: http://nictiz.nl/fhir/StructureDefinition/[id]
  • The name will be the .id capitalized, with hyphens removed.
  • The title will generally be the .id with hyphens replaced by spaces.

Where:

  • [project prefix] is preferably the same as its matching ART-DECOR project prefix
  • [purpose] and [English concept name] are generally a PascalCased name joining words together, with the first letter of every word capitalized.

12.1.1.2 ValueSets

  • For value sets as specified on ART-DECOR, the id will be:[value set OID]--[effective date as yyyymmddhhmmss]
  • The canonical URL will then be created as: http://decor.nictiz.nl/fhir/ValueSet/[id]
  • Both the name and title will be the name of the value set (usually in Dutch).

12.1.1.3 CodeSystems

  • For code systems as specified on ART-DECOR, the id will be:
    • If a canonical URL is available, then the latter part of the canonical URL.
    • Otherwise the code system OID.
  • The canonical URL will then be created as:
    • When a URI is explicitly stated, then this is used.
    • If the code system can be resolved with a URI in the OID register, then this is used.
    • Otherwise: urn:oid:[code system OID]
  • Both the name and title will be name of the code system (usually in Dutch). If the code system can be resolved in the OID register it will get that name as registered.

12.1.1.4 NamingSystems

NamingSystems are rarely needed. No strict guidance is given, but in general, the following conventions apply:

  • The id will be namingsystem-[identifier], where identifier is usually the latter part of the canonical URI this NamingSystem defines. However, a more informative identifier may be chosen if deemed more appropriate.
  • The name will be [identifier] capitalized and with dashes and such removed, unless a better scheme is deemed more appropriate.
  • NamingSystems don't have a title or canonical URI.

12.1.1.5 ConceptMaps

  • The id will be constructed as: [source ValueSet.name]-to-[target ValueSet.name]
  • The canonical URL will then be: http://nictiz.nl/fhir/ConceptMap/[id]
  • The name will be constructed as: [source ValueSet.name]_to_[target ValueSet.name]
  • The title will be constructed as: [source ValueSet.name] to [target ValueSet.name]

12.1.1.6 SearchParameters

SearchParameters are generally added from the nl-core layer onwards because the zib layer is without an use case definition.

  • The code should consist of one or multiple words in lowercase ASCII characters, with multiple words separated using a dash ("-"):
    • If the search parameter doesn't pertain to a specific functional model or use case, the FHIR element that the search parameter acts on should be used.
    • If the search parameter does pertain to a specific functional model or use case, the English name of the functional concept should be used.
    • If these guidelines aren't feasible, another descriptive name may be used at the profilers discretion.
  • The id will be: [target]-[code], where [target] is PascalCased and should be:
    • The resource name if the search parameter applies to a single resource type.
    • A descriptive name for the target if the search parameter applies to more than one resource type, e.g. "Medication" if the search parameter applies to medication related resources.
  • The canonical URL will then be: http://nictiz.nl/fhir/SearchParameter/[id]
  • The name will be: [target]_[code] with dashes and such removed in [target] and [code], unless a better scheme is deemed more appropriate.
  • The description will describe the working of the search parameter.
  • The purpose will be populated to indicate why the SearchParameter is created.
  • The search parameter is expressed using .expression.

12.1.1.7 Examples and fixtures

Examples and fixtures (example instances used for test/qualification purposes) are not conformance resources and lack the .url, .name and .title elements. However, to ensure consistency, the .id is standardized in the following way:

  • For fixtures: [profile id]-[3 character project prefix]-[unique string], capped to 64 characters. The [project prefix] is necessary to prevent the same id from being used by different projects in the same test environment. It is preferably the same as its matching ART-DECOR project prefix.
  • For examples: [profile id]-{[3 character project prefix]-}[unique string], capped to 64 characters (the [3 character project prefix] is not needed for examples, but may be included to facilitate re-use of tools for both fixtures and examples).

12.1.2 Folder structure and file name

profiles
resources/[id].xml
extensions
resources/[id].xml
valuesets
resources/terminology/[Dutch name]-[id].xml
conceptmaps
resources/terminology/conceptmap-[id].xml
codesystems
resources/terminology/codesystem-[name].xml
namingsystems
resources/terminology/[id].xml
searchparameters
SearchParameters/SearchParameter-[id].xml
examples
examples/[profile id]-[serial number, two digits].xml

12.2 Metadata

12.2.1 StructureDefinitions

12.2.1.1 Metadata on the StructureDefinition root

  • version: see #Versioning
  • status: as applicable (normally draft or active)
  • publisher: Nictiz
  • contact:
    • name: Nictiz
    • telecom:
      • system: url
      • value: https://www.nictiz.nl
      • use: work
  • description:
    • For profiles:
      • For zib profiles: the 'Concept' section from the zib. When the profile represents just a part of a zib and the 'Concept' section does not cover the actual use of the profile, a relevant excerpt or an adaptation should be used.
      • For nl-core profiles: copy the description from the zib profile.
      • For standard specific profiles: as applicable.
    • For extensions:
      • A description of what the extension is for.
  • purpose:
    • For profiles:
      • For zib profiles: This [resource type] resource represents the Dutch [zib ('Zorginformatiebouwsteen', i.e. Health and Care Information Model) [English zib name] [version]([release])]([link to the English zib page on zibs.nl]).
        N.B.: This template includes a markdown link: [text](url). A complete example would be: This Patient resource represents the Dutch [zib ('Zorginformatiebouwsteen', i.e. Health and Care Information Model) Patient v3.2 (2020)](https://zibs.nl/wiki/Patient-v3.2(2020EN)).
      • For nl-core profiles: A derived profile from [[id of zib profile]]([canonical of zib profile]) to provide a version better suited for implementation purposes. This profile augments the base profile with elements found in the various use cases that have adopted the zib.
      • For standard specific profiles: a description with a reference to the base profile, with an explanation of why it has been added.
    • For extensions:
      • For extension representing a specific concept: This extension represents the [concept name] of [name of the building block]], followed by a link to the functional description.
      • For other extensions this will usually be absent.
    • Note: for extensions and datatype profiles, guidance for profilers may be placed here as well.
  • copyright: Copyright and related rights waived via CC0, https://creativecommons.org/publicdomain/zero/1.0/. This does not apply to information from third parties, for example a medical terminology system. The implementer alone is responsible for identifying and obtaining any necessary licenses or authorizations to utilize third party IP in connection with the specification or otherwise.
  • mapping: see #Associating_the_functional_definition_to_StructureDefinitions
  • abstract: usually false, but may be set to true when this profile is to be used only as a base profile from which other profiles should be derived.
  • comment: when special guidance on the use of this profile is needed, it will be placed here. For example: when this profile has a (non-trivial) relationship to other profiles, when a concept of the functional description is mapped over multiple elements, etc.
    Note: this should be guidance aimed at implementers. When the guidance is aimed at profilers, for example in extensions or datatype profiles, it should be added to StructureDefinition.purpose.
  • alias:
    • for zib profiles: the zib concept name as dictated by #Associating the functional definition to StructureDefinitions.
    • for nl-core profiles: the id of the profile (this is actually just a placeholder for when there is no actual difference with the base profile, as FHIR requires StructureDefinition.differential or StructureDefinition.snapshot to be populated).

12.2.1.2 Metadata for elements

  • For elements that map directly (not implicitly) to zib concepts:
    • The English name from the zib concept is placed on element.short (in elements that refer to other zibs, the name is built up like ElementName::ReferenceZibName, but we only use the part before the double colons). If there are multiple zib elements mapped to a single FHIR element, it will be formatted like ZibConcept1 / ZibConcept2 (unless the names are the same).
    • The zib concept description is placed on element.definition, unless it's the root concept of the zib (because this will always be Root concept of the [...] information model ..., which doesn't provide much information). If there are multiple zib elements mapped to a single FHIR element, the descriptions will be placed in a bullet list.
    • The Dutch zib concept names are placed on element.alias
  • For elements that bind ValueSets:
    • If there is a ConceptMap, .binding.description will be: Use ConceptMap [ConceptMap.name] to translate terminology from the functional model to profile terminology in ValueSet [ValueSet.name]
    • Otherwise, .binding.description is not populated.
  • For elements that are (directly or indirectly) included in a custom constraint defined elsewhere:
    • The keys of all related constraints are added to .condition (note: this is thus not necessary when the constraint is defined on the element itself). Because .condition is added to notify the implementer that a constraint applies to the usage of this element, this rule applies to both the element directly mentioned in the constraint as well as its child element(s) that contain the mapping to the functional model.

12.2.2 ConceptMaps

12.2.2.1 Metadata on the ConceptMap root

  • status: as applicable (normally draft or active)
  • publisher: Nictiz
  • contact:
    • name: Nictiz
    • telecom:
      • system: url
      • value: https://www.nictiz.nl
      • use: work
  • description: Maps [functional ValueSet] codes as found in [linked description to the functional model] to [FHIR ValueSet] codes as found in FHIR R4.

    For zibs, the linked description will be the Dutch [zib ('Zorginformatiebouwsteen', i.e. Health and Care Information Model) [English zib name] [version]([release])]([link to the English zib page on zibs.nl]).

  • copyright: Copyright and related rights waived via CC0, https://creativecommons.org/publicdomain/zero/1.0/. This does not apply to information from third parties, for example a medical terminology system. The implementer alone is responsible for identifying and obtaining any necessary licenses or authorizations to utilize third party IP in connection with the specification or otherwise.
  • In addition, the [bidirectional extension (http://hl7.org/fhir/StructureDefinition/concept-bidirectional)](http://hl7.org/fhir/StructureDefinition/concept-bidirectional) is added with a value to indicate whether this ConceptMap can be used in both ways.

12.3 Miscellaneous

  • As a rule of thumb, info is placed 'as close as possible' to the elements that actually represent the data. For example: when slicing, the texts are placed on the slice content, not the slice itself. Similarly, in concept containers like Observation.component, this info is placed on the .valueXXX elements. This is also true for the zib mappings.
  • Slices get a lowerCamelCased name based on the functional concept name.
  • New elements defined through extensions get a lowerCamelCased name based on the functional concept name.
  • The key for invariants should be:
    • for zibs: zib-[English zib name]-[number], eg. zib-Alert-1.
    • for nl-core: nl-core-[English zib name]-[number].
    • for use case specific profiles: [dataset id]-[number], where [dataset id] is preferably the same as its matching ART-DECOR project prefix.
  • Inline examples in profiles are added (only) when the usage of an element in the profile is not trivial.

13 Special cases

13.1 Special use of FHIR resources

13.1.1 Measurement qualifiers in Observation.component

Functional models for measurements will usually be mapped onto the FHIR Observation resource. The type of measurement is conveyed using Observation.code and the result of the measurement is conveyed using Observation.value[x], or Observation.component.value[x] if the measurement consists of multiple components.

However, sometimes the functional model requires that the type of measurement needs to be further specified using a qualifier. For this purpose, the Observation.method and Observation.bodySite elements can be used. If it is not possible to map these concepts to these elements or to a core extension of Observation, they may be mapped to Observation.component, according to a remark in the specification ("or may provide qualifying information to Observation.code") and a discussion on Zulip, which concludes that the use of Observation.component for qualifiers is appropriate and preferred above (custom) extensions.

13.1.2 HL7 Vital Signs

The FHIR R4 specification for the Observation resource defines a profile for vital signs measurements. Furthermore, it states that:

If implementations use this Resource when expressing the profile-specific concepts as structured data, they SHALL conform to the following profiles

This means that the following zib profiles would need to be based on the HL7 Vital Signs profile:

  • BloodPressure
  • BodyHeight
  • BodyWeight
  • BodyTemperature
  • HeadCircumference
  • HeartRate
  • O2Saturation

The FHIR spec defines derived profiles for each of these vital signs, plus an IG with further recommendations.

However, the HL7 Vital Signs profile imposes two issues that make it impossible to implement zib profiles for the relevant vital signs measurements directly onto it:

  • The required terminology binding on Observation.component to the Vital Signs Units dataset makes it impossible to use .component's for representing anything else than a Quantity (like a CodeableConcept or a boolean), which is required by some of these zibs.
  • The mustSupport = true flag found on numerous elements isn't further specified in the HL7 Vital Signs IG, as is required by FHIR.

To circumvent this problem, it has been chosen not to directly base the zib profiles onto the HL7 Vital Signs profiles, but instead create profiles that aim to be (as) functional equivalent (as possible). This means that constraints, terminology bindings, etc. are adopted (except for the terminology binding mentioned above) and that the zib profiles are not in conflict with the equivalent HL7 profiles, except for the requirement that Observation.hasMember and Observation.derivedFrom should have a HL7 Vital Signs target – this requirement cannot be satisfied when cloning a profile. Where possible, the current profiling guidelines are followed rather than the approach by HL7 (e.g. where the HL7 profiles use slices with a fixed code and system to enforce the inclusion of a coding, the zib profiles use a pattern).

On occasion, the zib takes another approach to communicate a certain concept than HL7 Vital Signs. For example, zib BodyHeight specifies a ValueSet for the body position during measurement (lying or standing), while the HL7 Vital Signs suggests to include an additional Observation.code.coding for "body height measurement while lying down" (only). When the zib approach is not incompatible with the HL7 approach, the zib approach will be followed and the difference will be documented in the .comment of the relevant element.

To enforce a consistent approach for creating the vital signs zib profiles, a pattern profile has been created. Please note that this should not be interpreted as an equivalent to the base HL7 Vital Signs profile; the zib profiles deliberately skip the common base profile and are directly matched directly to their equivalent HL7 profiles like BloodPressure, BodyHeight, etc. This base profile is not needed for the current use case and omitting it prevents a layer of complexity; for example, the HL7 Vital Signs base profile binds certain ValueSets which are then restricted again in the derived profiles.

13.1.3 Person characteristics: extension or Observation?

Some type of information can be described as a characteristic or property of a person, like life stance, marital status, language proficiency, nationality, etc. There is a tendency in FHIR to inline this kind of data in the Patient resource, either using a core element or a core extension (and to a lesser degree to the RelatedPerson and Practitioner resources).

However, this choice seems somewhat arbitrary. Oftentimes this kind of information could equally well be regarded as an Observation. For example: whereas the name can be seen as something that is truly part of the patient, and a heart rate truly as a measurement made at some point in time, one could argue that both religion and body height fall somewhere in between – both of them usually apply to longer periods of time, neither of them require additional data apart from the main characteristic, and both of them can have an independent existence. So when modeling these kind of characteristics, the question is whether an Observation should be used or an extension.

Given that:

  • In international IGs, there doesn't seem to be a strong preference for one of the two approaches;
  • FHIR core should be used where possible; and
  • custom extensions are discouraged,

the following guidance applies:

  1. If possible, use a core element to model the characteristic.
  2. If this is not possible, use a core extension to model the characteristic.
  3. If this is not possible, create a profile on Observation.

13.2 Zib 2020 specific considerations

13.2.1 Implicit subject reference

Starting from release 2020, zibs oftentimes do not explicitly define their subject, i.e. who or what the information is about. However, the general guideline for zibs is that they capture healthcare information about a patient. So implicitly there is a subject, and this subject is usually a patient – although in some cases it might be inferred that the subject is something or someone else, or can be more than just a patient.

The following guidelines apply for profiling a zib without an explicit subject:

  • In the zib profiles, the subject will not be modeled explicitly.
  • In the nl-core profiles, the subject will be modeled as a reference to the relevant nl-core profile.

The reason for this is that even though the subject is assumed, when it is not explicitly defined, no formal mapping can be made to a profile. On the other hand, it can be assumed that the subject will be modeled explicitly once the zib is used in a use case scenario,

13.2.2 Referencing zib HealthProfessional

Zib HealthProfessional is mapped onto two profiles: one on FHIR resource PractitionerRole (zib-HealthProfessional-PractitionerRole) and one on FHIR resource Practitioner (zib-HealthProfessional-Practitioner). The PractitionerRole resource covers the recording of the location and types of services that HealthProfessionals are able to provide for a HealthcareProvider, whereas the Practitioner resource captures the personal information, including the identifiers of the health professional.

In FHIR, it is usually possible to define a reference to either resource. However, the reference to this zib from other zibs will be modelled in the following way:

  • Only zib-HealthProfessional-PractitionerRole is added to the list of target resources.
  • As described in #Constraining references, no restrictions on the base Practitioner or PractitionerRole are made.

This approach has been adopted for two reasons. First, the relation between these two resources points from PractitionerRole to Practitioner; having a PractitionerRole instance, the Practitioner instance can be resolved, but not the other way around. Second, information about a health professional will rarely be communicated without the details captured using PractitionerRole.

This mechanism guides implementers to the use of the zib-HealthProfessional-PractitionerRole profile as the main entry point for zib HealthProfessional. Meanwhile, when there is a need to communicate the zib-HealthProfessional-Practitioner profile instead, this is still possible as the base resource Practitioner is still available.

To ensure consistent documentation of this mechanism, the profile http://nictiz.nl/fhir/StructureDefinition/pattern-ZibHealthProfessionalReference has been created. The .title of this profile is simply 'Reference' instead of 'zib HealthProfessional Reference', as is expected following the profiling guidelines. This deviation from the guidelines is the result of the .title being used to render the profile in various tooling (mainly Simplifier), and the latter title would result in a rendering that is deemed confusing in some cases (see MM-3854).

14 Data type mapping

In most cases, the mapping between ART-DECOR, zib and FHIR datatypes is straightforward, though there are a number of cases which are not as clear-cut. The table below shows how datatypes from the functional models should be mapped to FHIR.

ART-DECOR zib FHIR Remarks
count INT Count or integer Count in FHIR is derived from Quantity. This is appropriate for observations, such as a count of red blood cells in a specimen. Count in FHIR is a Quantity with system set to http://unitsofmeasure.org and code set to 1 (for units). If this is not appropriate, use integer.
boolean BL boolean
blob ED base64Binary or Attachment
code CD CodeableConcept or coding or code In general, CodeableConcept will be the right choice. See Using Codes in Resources for details.
complex ANY BackboneElement complex is data which is not further defined in DECOR, and thus cannot be translated.
currency - Money Does not have a zib counterpart currently. Theoretically this could be MO.
date - date Does not have a zib counterpart currently. Closest counterpart could be TS.
datetime TS dateTime
decimal - decimal Does not have a zib counterpart currently. Theoretically this could be REAL.
duration - Duration Does not have a zib counterpart currently. Closest counterpart could be PQ.
identifier II Identifier
ordinal CO CodeableConcept or coding or code See Using Codes in Resources for details.
string ST string or markdown
ratio - Ratio Does not have a zib counterpart currently. Closest counterpart could be RTO.
text - string Does not have a zib counterpart currently.
quantity PQ Quantity

14.1 Notes on binary data (ED or blob)

Binary data requires special considerations when profiling. One aspect of binary data is that it tends to represent large multimedia files. Healthcare systems might want to have explicit control over the exchange of such files, rather than being confronted with unexpected large data transfers when performing a search or read operation. In addition, these files could have a need for additional metadata next to the structured content present in the clinical data model.

In FHIR R4, the exchange of binary data is mediated using a combination of resources and data types:

  • Data type Attachment; this data type provides the means to transfer the actual bytes and annotate it with the proper content type, so that the receiver knows how it is interpreted. This data type offers two ways of transporting the binary content:
    • inline (Attachment.data).
    • as reference to an external URL (Attachment.url).
  • Resource type DocumentReference; this resource type provides a way to specify metadata about a multimedia file specified using the Attachment data type.
  • Resource type Media; this resource type is similar and has been deprecated in FHIR R5 in favor of DocumentReference.
  • Resource type Binary; this resource type is only meant to wrap binary data in a FHIR resource. It could be the target of Attachment.url if there's a strict need to keep all exchanges within the FHIR RESTful framework.

It should be noted that the Observation resource does not offer Attachment as a valid data type for .value[x] (or .component.value[x]), although FHIR R5 does support this.

This translates to the following guidelines:

  • Binary data should be mapped onto data type Attachment.
    • For the generic zib and nl-core layers, no restrictions should be imposed on the question of inlining (Attachment.data) versus referring (Attachment.url) the actual binary data.
    • For use case specific profiles, it should be carefully considered which of these options should be supported to prevent unexpected transfers of large amounts of data.
  • If the resource type supports an element of Attachment natively that matches the intent of the functional model, it should be used, unless additional metadata needs to be provided which cannot be accommodated by the resource type.
  • If the resource type does not support an element of Attachment natively or if there's a need to provide additional metadata, a profile on DocumentReference resource should be defined. The Media resource should not be used as it has been deprecated in FHIR R5.
    • If there's no obvious way to refer between this DocumentReference and the focal resource of the data model, DocumentReference.context.related can be used.
  • The use of the Binary resource is dependent on the use case. Normally, it should not be needed to explicitly specify the use of this resource.