Relationships between people.: Difference between revisions

From Endeavour Knowledge Base
No edit summary
No edit summary
 
(14 intermediate revisions by the same user not shown)
Line 1: Line 1:
Family relationships and none professional relationships between people, are included in the common information model in two main ways:
Family relationships, and none professional relationships between people, are included in the common information model in two main ways:


# An ontology of relationships, using relationship types and person as a 'relation' subclasses, the latter being aligned with the Snomed-CT 'Person' subclasses with additions for in-laws.
# An ontology of relationships, using relationship types and a 'relation' as subclass of related person, the latter being aligned with the Snomed-CT 'Person' subclasses with additions for in-laws.The ontology properly defines relationships and relations.
# The data model, whereby a [[https://wiki.discoverydataservice.org/index.php?title=Health_data_model_content#Related_Person related person]] is linked to another person transitively via a relationship type and relation entity, the latter being aligned with the above.
# The data model, whereby a [[https://wiki.discoverydataservice.org/index.php?title=Health_data_model_content#Related_Person related person]] is linked to another person via a relationship type, the latter being aligned with the above.
 
As with the common information model in general, the two are related to each other so an enquirer can build sophisticated relationship queries and advanced family trees.
 
[[Image:Family relationships.jpg|right|thumb|Example of a relationship. The related person {Mary} is both the mother of {Fred}, and the Sister of {Helen}. 'is mother of' is modelled as an inverse property of 'has mother'..]]


As with the common information model in general, the two are related to each other so an enquirer can build sophisticated relationship queries and advanced family trees. Most people would struggle with understanding who their third cousins were, yet they are genetically quite close.


Relationships appear complex to model because of the ambiguous way the terms are used in plain language.  
Relationships appear complex to model because of the ambiguous way the terms are used in plain language.  


For example a "mother" is a person who is a mother of somebody. A son "has a mother" who is a mother. However, the same person who is a mother may also be a "sister" (whether or not the relationship to her sister has been authored). Consequently, there is a need not only to model the relationship between people (using a sub property of "related to"), but also to model the relation as a role independently of the relationship itself. The introduction of the role entity along the path of the relationship resolves the difficulty.
For example a "mother" is a person who is a mother of somebody. A son "has a mother" who is a mother. However, the same person who is a mother may also be a "sister" (whether or not the relationship to her sister has been authored). Consequently, there is a need not only to model the relationship between people (using a sub property of "related to"), but also to model the relation as a role independently of the relationship itself. The ontology resolves this difficulty as follows:


An additional complication is that a related person may or may not be a patient or person recorded in the record store. It may be someone unknown or abroad and there may be little information on them other than a contact detail. Therefore the related person also has to be modelled as an entity separately from patients.  These classes and properties can be illustrated as follows:[[File:Related person.jpg|center|thumb|600x400px]]
By building up simple family tree relations (parent, child etc), it is possible both to deduce relations, and relationships, including the relations that people struggle to remember.  Most people would struggle with understanding who their third cousins were, yet they are genetically quite close. A third cousin is defined as a child of one's parent's second cousin, or perhaps more precisely, a great grandchild of a great grandparent's sibling! Anyone capable of remembering this, is probably a genealogist.
Thus there are 4 potential entity types in a relationship


a) The source patient (the person who has the relationship with someone else)
Within the data model, the two entities involved are the patient and the related person, and the relationship between them "has related person/ is relation of " subtype. Within the ontology, relations are defined as subclasses of related person by dint of the relationships they have to other people.


b) The role of person who is the relation e.g. Mother, Carer, Sister.  
Thus it is possible to deduce that Mary is a mother and a sister simply by noting the relationship as part of a family tree.


c) Information about that person, which would be the same, whatever relation they were to other people.
[[Image:Family property chains.jpg|right|thumb|Example of a property chain. (is sister of -> is mother of) is a subproperty of> is aunt of.]]


d) Another patient in the record store i.e. where the person is also a patient.
Relations are more complex when chained together. Common parlance is to use an abbreviated term to name a once removed relationship. In the above example, Helen is an Aunt to Fred. Once again the ontology defines this by the use of property chains.


It is possible to model all links between people as relationships of this kind. However, in practice most relationships can be more direct e.g. the subject of an observation being a patient, or an encounter being performed by a doctor. In the common model these more obvious ones are modelled as direct relationships. Relationships of the kind modelled above are reserved for family relationships or non professional relationships such as carers.
A third cousin can equally be defined as a sub property of ''" has Great Grand Parent - > is Sibling Of - >  is Great Grand Parent Of"''  and this is indeed modelled in the ontology in case anyone might need it.<br />

Latest revision as of 09:15, 20 May 2020

Family relationships, and none professional relationships between people, are included in the common information model in two main ways:

  1. An ontology of relationships, using relationship types and a 'relation' as subclass of related person, the latter being aligned with the Snomed-CT 'Person' subclasses with additions for in-laws.The ontology properly defines relationships and relations.
  2. The data model, whereby a [related person] is linked to another person via a relationship type, the latter being aligned with the above.

As with the common information model in general, the two are related to each other so an enquirer can build sophisticated relationship queries and advanced family trees.

Example of a relationship. The related person {Mary} is both the mother of {Fred}, and the Sister of {Helen}. 'is mother of' is modelled as an inverse property of 'has mother'..


Relationships appear complex to model because of the ambiguous way the terms are used in plain language.

For example a "mother" is a person who is a mother of somebody. A son "has a mother" who is a mother. However, the same person who is a mother may also be a "sister" (whether or not the relationship to her sister has been authored). Consequently, there is a need not only to model the relationship between people (using a sub property of "related to"), but also to model the relation as a role independently of the relationship itself. The ontology resolves this difficulty as follows:

By building up simple family tree relations (parent, child etc), it is possible both to deduce relations, and relationships, including the relations that people struggle to remember.  Most people would struggle with understanding who their third cousins were, yet they are genetically quite close. A third cousin is defined as a child of one's parent's second cousin, or perhaps more precisely, a great grandchild of a great grandparent's sibling! Anyone capable of remembering this, is probably a genealogist.

Within the data model, the two entities involved are the patient and the related person, and the relationship between them "has related person/ is relation of " subtype. Within the ontology, relations are defined as subclasses of related person by dint of the relationships they have to other people.

Thus it is possible to deduce that Mary is a mother and a sister simply by noting the relationship as part of a family tree.

Example of a property chain. (is sister of -> is mother of) is a subproperty of> is aunt of.

Relations are more complex when chained together. Common parlance is to use an abbreviated term to name a once removed relationship. In the above example, Helen is an Aunt to Fred. Once again the ontology defines this by the use of property chains.

A third cousin can equally be defined as a sub property of " has Great Grand Parent - > is Sibling Of - > is Great Grand Parent Of" and this is indeed modelled in the ontology in case anyone might need it.