Transforming child elements in a list

To map two lists against each other that contain a set of child items we need to iterate both of these lists. When iterating each of their appearances each of the corresponding paths need to be mapped.

Example

  - name: "collection"
    with:
      fhir: "$resource.collection" # 0..n
      openehr: "$archetype/data[at0001]" # 0..n
      type: "NONE"
    followedBy:
      mappings:
        - name: "collected"
          with:
            fhir: "collected"
            openehr: "items[at0015]"

        - name: "collector"
          with:
            fhir: "collector"
            openehr: "items[at0070]"

In the pseudocode example above, both collection and $archetype/data[at0001] are recurring elements. Therefore, they are specified on the same level in the mapping. All mappings of a specific occurrence of collection are within followedBy, each of which maps in the followedBy the elements of the cluster. Therefore, these lists would be transformed into each other.

  "collection": [
    {
      "collected": "12:00",
      "collector": "13"
    },
    {
      "collected": "13:00",
      "collector": "14"
    }
  ]

to

  "at0001": [
    {
      "at0015": "12:00",
      "at0070": "13"
    },
    {
      "at0015": "13:00",
      "at0070": "14"
    }
  ]

Wrong mapping

In this example, the data[0001], which has a cardinality of 0..n, is contained as part of the parent mapping paths. FHIRconnect always appends to a 0..n path if it’s not iterated. The below example would append a new entry in data[0001] for each occurrence of collection.

  - name: "collection"
    with:
      fhir: "$resource.collection" # 0..n
      openehr: "$archetype"
      type: "NONE"
    followedBy:
      mappings:
        - name: "collected"
          with:
            fhir: "collected"
            openehr: "data[at0001]/items[at0015]"

        - name: "collector"
          with:
            fhir: "collector"
            openehr: "data[at0001]/items[at0070]"
  "collection": [
    {
      "collected": "12:00",
      "collector": "13"
    },
    {
      "collected": "13:00",
      "collector": "14"
    }
  ]

would result in this:

{
  "at0001": [
    {
      "at0015": "12:00"
    },
    {
      "at0070": "13"
    },
    {
      "at0015": "13:00"
    },
    {
      "at0070": "14"
    }
  ]
}

Example with Slot

In this example, both ServiceRequest.specimen as well as $archetype/activities[at0001]/description[at0009]/items[openEHR-EHR-CLUSTER.specimen.v1] are recurring elements, therefore they are specified on the same level in the mapping. All mappings of a specific occurrence of specimen are within followedBy. For each specimen, one cluster is created. Each of which maps in the followedBy the identifier.

  - name: "specimen"
    with:
      fhir: "$resource.specimen" #0..n
      openehr: "$reference"
      type: "NONE"
    reference:
      resourceType: "Specimen"
      mappings:
        - name: "specimenRecurring"
          with:
            fhir: "$fhirRoot"
            openehr: "$archetype/activities[at0001]
/description[at0009]/items[openEHR-EHR-CLUSTER.specimen.v1]"   #0..n
          slotArchetype: "CLUSTER.specimen.v1"
          followedBy:
            mappings:
              - name: "specimenIdentifier"
                with:
                  fhir: "identifier"
                  openehr: "items[at0001]"

Incorrect Mapping

If, on the contrary, this would create two specimen clusters—one with the slot and one with only an identifier.

  - name: "specimen"
    with:
      fhir: "$resource.specimen"
      openehr: "$reference"
      type: "NONE"
    reference:
      resourceType: "Specimen"
      mappings:
        - name: "specimenRecurring"
          with:
            fhir: "$fhirRoot"
            openehr: "$archetype/activities[at0001]/description[at0009]/items[openEHR-EHR-CLUSTER.specimen.v1]"
          slotArchetype: "CLUSTER.specimen.v1"
        - name: "specimenIdentifier"
          with:
            fhir: "identifier"
            openehr: "$archetype/activities[at0001]/description[at0009]/items[openEHR-EHR-CLUSTER.specimen.v1]/items[at0001]"

Example of double nesting

Imagine the following example. We have a category in FHIR with 0..n elements, and this category has 0..n text elements. The same goes in openEHR, where you have one at0094 in the protocol with 0..n at0063 codes. Mapping that correctly would mean matching the corresponding paths like this:

mappings:
  - name: "category"
    with:
      fhir: "$resource.category" # 0..n
      openehr: "$archetype/protocol[at0004]/items[at0094]" # 0..n
    followedBy:
      - name: "category"
        with:
          fhir: "text" # 0..n
          openehr: "/items[at0063]" # 0..n
  "category": [
    {
      "text": [
        "1",
        "2"
      ]
    },
    {
      "text": [
        "3",
        "4"
      ]
    }
  ]

would result in this:

{
  "at0094": [
    {
      "at0063": ["1", "2"]
    },
    {
      "at0063": ["3", "4"]
    }
  ]
}

Incorrect Mapping

This mapping would iterate one time through all the text contained, resulting in four entries of text being mapped into openEHR. These entries are now mapped to a path with two 0..n, since we always append to 0..n, this would resolve in 4 single entries inside at0094. This would be the same case when mapping from openEHR to FHIR.

mappings:
  - name: "category"
    with:
      fhir: "$resource.category.text" # 0..n
      openehr: "$archetype/protocol[at0004]/items[at0094]/items[at0063]"
  "category": [
    {
      "text": [
        "1",
        "2"
      ]
    },
    {
      "text": [
        "3",
        "4"
      ]
    }
  ]

would result in this:

{
  "at0094": [
    {
      "at0063": ["1"]
    },
    {
      "at0063": ["2"]
    },
    {
      "at0063": ["3"]
    },
    {
      "at0063": ["4"]
    }
  ]
}