✨ Add family relationship conformance resources

[liberaliscomputing]

This PR adds:

• a family relationship StructureDefinition resource;
• corresponding example resources

It also closes #21.

Re @RobertJCarroll @allisonheath

[bwalsh]

https://terminology.hl7.org/1.0.0/CodeSystem-v3-RoleCode.html seems like a reasonable choice for ontology - given our desire to stay close to "stock" FHIR.

Things to think about:

• RoleCode seems to be a catch-all of misc roles e.g [ IncidentalServiceDeliveryLocationRoleType, OutpatientFacilityPracticeSetting, ..] that happens to hold some family relationships.
• Given we are scoping this to family relationship, is there a more specific ontology that might provide richer choices ?
• AnVIL family relationships in the raw data include :

{
    '#N/A',
    'Affected',
    'Affected1',
    'Affected2',
    'Aunt or Uncle',
    'Brother',
    'Brother or Sister',
    'Brother or Sister in Law',
    'Brother or Sister-in-Law',
    'Brother or Sister-in-law',
    'Cousin',
    'DZ twin',
    'Father',
    'Father1',
    'First Cousin',
    'Husband or Wife',
    'Identical Twin of Proband',
    'Maternal Aunt or Uncle',
    'Maternal Grandparent',
    'Mother',
    'Mother1',
    'Nephew or Niece',
    'Not Related to the Proband',
    'Not related to the Proband',
    'Other 2nd Degree Relative',
    'Other 2nd degree relative',
    'Other 3rd Degree Relative',
    'Other 3rd degree relative',
    'Other 4th Degree Relative',
    'Other 4th degree relative',
    'Other Not Related to Proband',
    'Other fifth degree relative',
    'Other forth degree relative',
    'Other not related to proband relative',
    'Other second degree relative',
    'Other third degree relative',
    'Parent',
    'Paternal Aunt or Uncle',
    'Paternal Grandparent',
    'Proband',
    'ProbandBrother or Sister',
    'Sibling',
    'Sister',
    'Son or Daughter',
    'Son or Daughter-in-law',
    'Twin brother',
    'Twin sister',
    'Unaffected Sibling',
    'affected half-sib',
    'affected sibling',
    'affected twin',
    'brother',
    'brother2',
    'father',
    'father2',
    'mother',
    'mother2/sister1',
    'proband',
    'sibling',
    'sister2'
}

[bwalsh]

It would be useful to document if we have considered a more specific model i.e:

• http://hl7.org/fhir/relatedperson.html
• http://hl7.org/fhir/valueset-relatedperson-relationshiptype.html

[liberaliscomputing]

Re @bwalsh, thanks for pointing this out! Yes, the role code ValueSet may not be enough inclusive. GA4GH's pedigree WG also proposes a new ValueSet (https://docs.google.com/spreadsheets/d/1AxRufaSICADO_f9CAhbkVh5dChT4FjHZgvGFFdC-2WI/edit?pli=1#gid=140786158). We may want to have a discussion about using a more authoritative, inclusive ValueSet for pedigree?

[liberaliscomputing]

Re @bwalsh,

It would be useful to document if we have considered a more specific model i.e:

* http://hl7.org/fhir/relatedperson.html

* http://hl7.org/fhir/valueset-relatedperson-relationshiptype.html

• Re RelatedPerson, I've once considered this profile for modeling pedigree as well. By definition, however,

Information about a person that is involved in the care for a patient, but who is not the target of healthcare, nor has a formal responsibility in the care process.

• Re the ValueSet, this ValueSet is the combination of contactRole2 and RoleCode. And contactRole2 seems less relevant to our effort in modeling pedigree.