Ontology

Some awesome prose from Philip to come...

(Per Doug 5/13 - bring Philip to vetting meeting -

From Core Ontology:

Supported Medical Concept Systems

SHRINE currently makes patient demographics, diagnoses, medications and a sample of laboratory results data available to be searched. SHRINE Core Ontology is a collection of terms that describe these data and can be used to construct queries. The terms are arranged into a hierarchy for easy navigation; the ontology can be searched for a particular terms as well.

A standard terminology was selected (where possible) to represent each type of data accessible via SHRINE.
A terminology was modified as necessary to fit the requirements of SHRINE. The below standard terminologies provide the baseline for the Core ontology.

The Core ontology is a static snapshot of underlying standards. As they change (for example, regular updates of ICD-9-CM and RxNorm), we expect that the Core ontology will follow after a reasonable delay to accommodate the effort of introducing the changes, resolving issues and quality assurance processes.

Each participating institution's data is loaded with native coding schemes used by local information technology solutions. In order to understand SHRINE queries issued using terms from the Core ontology, a mapping is created and maintained for each institution which allows a translation from SHRINE Core ontology term to a corresponding local one in order to perform a query at the institution. While every effort was made during creation of mapping files and their testing to faithfully preserve the meaning of data, as with any mapping exercise, there may be instances where the translation between SHRINE Core ontology and local hospital's data is not perfect.

Demographics

• Age
• Gender - HL7 Administrative Gender
• Language - ISO 639-1
• Marital Status - HL7 Marital Status
• Race and Ethnicity - CDC Race & Ethnicity Code Sets

Diagnoses

•  ICD-9-CM terms and hierarchy

Medications

• RxNorm terms and NDF-RT hierarchy

Lab Tests

• LOINC common top 300 ranked common labs with LOINC hierarchy