gecco-questionnaire

GECCO Questionnaire

This project converts the internal metadata model (Logical Model) into a FHIR Questionnaire with specific extensions. These extensions can be used to map the data back to the logical model and create a different set of resources of the FHIR standard.

Logical Model

The LogicalModel ist defined as a set of Kotlin classes:

data class LogicalModel(
    @Text("Anamnese / Risikofaktoren")
    var anamnesis: Anamnesis = Anamnesis(),
    @Text("Bildgebung") @Ignore
    var imaging: Imaging = Imaging(),
    //[...]
)

data class Anamnesis(
    @Text("Leidet der/die Patient*in unter einer chronischen Lungenerkrankung?")
    var hasChronicLungDiseases: YesNoUnknown? = null,
    @Text("An welcher chronischen Lungenerkrankung leidet der/die Patient*in?") @EnableWhenYes("anamnesis.hasChronicLungDiseases")
    var chronicLungDiseases: AnamnesisChronicLungDiseases = AnamnesisChronicLungDiseases()
    //[...]
)

data class Demographics(
    @FhirProfile("https://www.netzwerk-universitaetsmedizin.de/fhir/StructureDefinition/sex-assigned-at-birth")
    @Text("Biologisches Geschlecht")
    var biologicalSex: BirthSex? = null,
    @FhirProfile("https://www.netzwerk-universitaetsmedizin.de/fhir/StructureDefinition/pregnancy-status")
    @Text("Liegt eine Schwangerschaft vor?") @EnableWhen(
        "demographics.biologicalSex",
        "http://hl7.org/fhir/administrative-gender",
        "female"
    )
    var pregnancyStatus: PregnancyStatus? = null,
)

enum class YesNoUnknown(override val coding: Coding) : CodeableEnum<YesNoUnknown> {
    YES(Coding("http://terminology.hl7.org/CodeSystem/v2-0136", "Y", "Ja")),
    NO(Coding("http://terminology.hl7.org/CodeSystem/v2-0136", "N", "Nein")),
    UNKNOWN(Coding("http://terminology.hl7.org/CodeSystem/data-absent-reason", "asked-unknown", "Unbekannt"))
}

enum class PregnancyStatus(override val codeableConcept: CodeableConcept) : ConceptEnum<PregnancyStatus> {
    NOT_PREGNANT(CodeableConcept().apply {
        coding = listOf(
            loinc("LA26683-5", "Not pregnant"),
            snomed("60001007", "Not pregnant (finding)")
        )
    }),
    // [...]
}

The model will be converted recursively to Questionnaire.item resources, complex items will be converted to item.type = "group" elements.

• The @Text() annotation controls the Questionnaire.item.text attribute.
• The @Ignore attribute omits the resulting sub-part from the resulting Questionnaire
• The @EnableWhenYes annotation emits a Questionnaire.item.enableWhen component with answerCoding = {system: http://terminology.hl7.org/CodeSystem/v2-0136, code: Y}
• The YesNoUnknown enum entries will be rendered as Questionnaire.item.answerOption
• The PregnancyStatus enum will be rendered Questionnaire.item.answerOption with answerOption.valueCoding as the first Coding of the CodeableConcept
• The @ComboBox annotation will emit a questionnaire-itemControl=drop-down extension
• The @FhirProfile annotation will control the GeccoTargetProfile extension, which is also inherited to child elmenents.
• The recursive behaviour can be overwritten for specific datatypes by overwriting the ClassToItemRenderer<T> interface

interface ClassToItemRenderer<T> {
    fun render(compassId: String, item: QItem)
    fun parse(geccoId: String, item: Map<String, QRAnswer>): T
}

Usage

val renderers = listOf(/* [...] */)

val questionnaire = LogicalModel.toQuestionnaire(renderers)

val questionnaireResponse = generateResponse(questionnaire)

val logicalModel = toLogicalModel(questionnaire, questionnaireResponse, listOf(renderers))

val geccoBundle = logicalModelToGeccoProfile(logicalModel, patiendId, recordedDate, bundleBuilder)