summarized from this document, which has more details....
Synopsis
Wikidata is a community-maintained database, and we have been loading data on biomedical topics (primarily genes, drugs, and diseases). One data source is the Disease Ontology (DO), which defines diseases and their relationships to other diseases. DO also maintains links to other disease terminologies like the International Classification of Disease (ICD) (which has two commonly-used versions, ICD-9 and ICD-10). The links in DO to ICD are typically maintained as non-specific cross-references (also known as “dbxrefs”), but in some cases, the ICD entry is more general or specific than the linked DO entry. We would like to create a web interface that allows Wikidata members to make these mappings more precise by specifying one of three mapping types -- “exact match”, “broad match” and “narrow match”.
Straw man workflow
This workflow is roughly based on what’s available from the Wikidata Mix-n-match tool.
Extract all DO mappings to ICD (from Wikidata or from the DO OWL file; use latest releases here)
For cases where DO name exactly matches ICD name, infer an “exact match” link (ICD data here at ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Publications/ICD10CM/2018/ICD-10-CM-Codes-Tables-and-Index-2018.zip)
Use remaining cases to populate a web portal
a) User logs in using Wikidata oauth
b) User is shown information about one DO entry and its mapping to ICD (show name, description, aliases, parents, childrens, and “exclusions”)
c) User selects “exact match”, “broad match”, or “narrow match” (with an optional comment on why)
d) Web app writes new statement to Wikidata
e) User is shown next
An existing Wikidata monitoring application would then detect that addition and create a monthly report for review by DO curators for possible inclusion in the official DO release.
summarized from this document, which has more details....
Synopsis
Wikidata is a community-maintained database, and we have been loading data on biomedical topics (primarily genes, drugs, and diseases). One data source is the Disease Ontology (DO), which defines diseases and their relationships to other diseases. DO also maintains links to other disease terminologies like the International Classification of Disease (ICD) (which has two commonly-used versions, ICD-9 and ICD-10). The links in DO to ICD are typically maintained as non-specific cross-references (also known as “dbxrefs”), but in some cases, the ICD entry is more general or specific than the linked DO entry. We would like to create a web interface that allows Wikidata members to make these mappings more precise by specifying one of three mapping types -- “exact match”, “broad match” and “narrow match”.
Straw man workflow
This workflow is roughly based on what’s available from the Wikidata Mix-n-match tool.
An existing Wikidata monitoring application would then detect that addition and create a monthly report for review by DO curators for possible inclusion in the official DO release.