What is an ontology?
The idea of "Ontology" is an idea borrowed from philosophy, and it is used to discuss "existence."
In the realm of i2b2, an ontology is actually a "taxonomy," which is a hierarchical group of terms or concepts. If you recall (from your high-school biology class) Linnaeus's "binomial nomenclature" model of how to name living organisms, then you already know what a taxonomy is. For instance, the common house fly Musca domestica and other flies in the family Muscidae are members of the order Diptera, which are organisms in the class Insecta, phylum Arthropoda, kingdom Animalia. A taxonomy defines the hierarchical "is-a" relationship between two concepts. Likewise, a Camry is a Toyota, which is an automobile, which is a type of motorized, wheeled conveyance.
Why does i2b2 have an ontology?
An i2b2 instance requires an ontology so that researchers can locate "concepts" of interest and query data.
The medical concepts contained in an ontology are essentially the building blocks of an i2b2 query:
This is what “drives” the upper-left corner of the i2b2 webclient. In order to make facts or observations you wish to load into your i2b2 query-able, you must have corresponding entries in your ontology (that account for the codes used to represent these facts or observations).
These concepts are rendered as a collection of nested folders and are assumed to represent child-parent (or is-a) relationships. Researchers may locate these concepts by using a hierarchical tree, or taxonomy. When you log in to the demo i2b2 site (username "demo", password "demouser" are already filled in), you will see the i2b2 ontology tree on the upper left side of the user interface. We also refer to that listing as a group of ontologies, where each domain or top-level taxonomy is called an ontology in its own right.
In this screenshot, we would say that i2b2 is displaying an ontology tree with 13 top-level ontologies or domains, namely: Clinical Trials, Custom Metadata, Demographics, etc.
For instance:
- diabetes mellitus is an endocrine disorder, which is a type of diagnosis; it will be found in both the "Diagnoses" tree and in the "Diagnoses (ICD10)" tree;
- aspirin is a non-steroidal anti-inflammatory drug, which is a type of medication; it will be found in the "Medications" tree;
- tonsillectomy is a type of surgery, which is a procedure; it will be found in the "Procedures" tree;
- hemoglobin A1c is a blood test, which is a type of laboratory test; it will be found in the "Laboratory Tests" tree; and
- gender or sex is a patient-demographic descriptor; it will be found in the "Demographics" tree.
How does it work? How does the i2b2 ontology make the patient data queryable?
A) The ontology is made up of medical concepts. The ontology is visible to researchers through the webclient, and each medical concept in the ontology includes specific code(s) that match medical observation facts (such as diagnoses, medications, procedures, etc). A researcher can select concepts they are interested in to build a query, and medical data that matches those concepts' codes will be counted.
B) The i2b2 ontology in the i2b2 metadata database works together with the patient data in the i2b2 CRC database. Each patient "observation" in the CRC database must have a code associated with it, and that code must match a healthcare concept — diagnosis, procedure, medication, lab test, demographic descriptor, etc. — that exists in the metadata database. For instance, if the patient's electronic health record indicates that the patient had the procedure "tonsillectomy with adenoidectomy," then that fact needs to be recorded in the CRC database using the standard code for that particular procedure. In the case of using the default i2b2 Procedures metadata tree, that code would be "ICD9:28.3". So the CRC patient data will be queryable in i2b2 only if the patient facts in the CRC database are recorded utilizing standard codes that are referenced in the i2b2 metadata (ontology trees).
What are the sources of the concepts and domains? (Where do they come from?)
Typically, the concepts that make up domains are based on medical terminology standards. Some commonly used standards to represent basic structured clinical patient data collected in EHRs include:
| Data domains | Typical Standards |
| Demographics | HL7 Administrative |
| Diagnoses | ICD |
| Procedures | ICD, CPT, HCPCS |
| Medications | RxNorm + VA Classes hierarchy |
| Labs | LOINC |
| Vital Signs | LOINC |
- Medical terminology releases are not directly usable by i2b2
- Standardized I2b2 tables must be created for each desired medical terminology to make it usable within i2b2.
Do all i2b2 instances always have the same ontology?
No. Several i2b2 ontologies have been developed and are openly available for use. Any organization may also modify an existing ontology for its own use, or develop a new ontology. (See Ontologies 201.)
What ontologies can I use right away ('out of the box')?
The i2b2 database loading modules come with at least 3 sets of "metadata" or ontology trees. These are the demo ontology, the ACT ontology, and the ACT-on-OMOP ontology.
| Name | Description | Included Domains (I don't think all domains can be listed, or it will be hard to understand. I think this merges with description and tries to convey the level of detail and comprehensiveness). | Target Data Model |
|---|---|---|---|
| i2b2 demo ontology | default metadata from i2b2 authors | fixme | i2b2 Common Data Model (star-schema); default CRC database has matching concepts |
| ACT ontology | ENACT project | fixme | i2b2 Common Data Model (star-schema); ACT CRC demo database has matching concepts |
| ACT-on-OMOP ontology | ENACT project | fixme | i2b2 Common Data Model (star-schema), but modified with views into the OMOP Common Data Model; CRC database loaded with SYNPUF demo data has matching concepts |
That may be the "demo" ontology, but it doesn't resemble the "ACT" ontology that our PI was showing us. How many different ontology trees does i2b2 have?
The i2b2 database loading modules come with at least 3 sets of "metadata" or ontology trees. These are the demo ontology, the ACT ontology, and the ACT-on-OMOP ontology.
How about a thumbnail description of these ontologies? How do they differ from each other?
Sure. Here is a table describing the major differences among them.
How should I choose which ontology to use?
The ontology you choose to use depends on your local goals and local data. As long as your primary purpose is to understand and show how the i2b2 webclient works, the i2b2 demo ontology is sufficient. If your goal is to set up i2b2 for research use, one of the ACT ontologies will be more useful. In general, to set up an i2b2 instance to support research, start with the ACT Ontology. This Ontology includes a wide range of terminologies that should cover most code-sets you'll find in your EMR data. OMOP refers to a data architecture, so if you are using this format locally, then ACT on OMOP will be the most relevant.
Are the ontology and data stored separately? How to they connect and where are the data?
You just mentioned Domains. What do you mean by that?
sfsf saf
There are many terms related to i2b2 ontologies that are new to me. Is there a Glossary?
Yes! Here is a Glossary that provides explainers for many of the key terms related to i2b2 ontologies and the i2b2 architecture.
