Observational Health Data Sciences and Informatics

The availability of very large-scale healthcare databases in electronic form has opened the possibility to generate systematic and large-scale evidence and insights about the application of healthcare to patients. This discipline is called Observational Outcome Research, and it uses the longitudinal patient level clinical data in order to describe and understand the pathogenesis of disease and the effect of other clinical events as well as treatment interventions on the progression of the disease. This research constitutes secondary use of the data, which is being collected usually for purposes other than research: administrative data such as insurance reimbursement claims and Electronic Health or Medical Record (EHR, EMR).

Because of the collection purpose for primary use, the format and representation of the data follows that primary use. It also introduces artifacts and bias into the data. In addition, all source datasets differ from each other in format and content representation. Since healthcare systems differ between countries, the problem becomes even harder for research carried out internationally. All this makes robust, reproducible and automated research a significant challenge.

The solution is the standardization of the data and a standardization of the representation. This allows methods and tools to operate on data of disparate origin, freeing the analyst from having to dissect the idiosyncrasies of a particular dataset and manipulating the data to make it fit for research. It also allows to develop analytical methods on one dataset, and applying it an any other dataset in CDM format.

The OMOP CDM and Standardized Vocabularies provide such a framework for systematic research. It consists of the following components and mechanisms:

• Multiple Vocabularies used in observational data consolidated into a common format. This relieves the researchers from having to understand and handle multiple different formats and life cycle conventions of the Vocabularies.
• Assignment of a clinical Domain for each Concept. This also defines in which CDM table a clinical entity should be placed into or looked up in at query time.
• Relationships and mappings between Concepts within the Vocabularies and across Vocabularies.
• Hierarchical structure within Concepts of a Domain. This allows to researcher to query for all Concepts (e.g. drug products) that are hierarchically subsumed under a higher level Concept (e.g. a drug class).

It is important to note that these components are constructed strictly for the purpose of supporting observational research. In that regard the Standardized Vocabularies differ from large collections with equivalence mapping of concepts such as the UMLS. UMLS resources have been used heavily as a basis for constructing many of the Standardized Vocabulary components, but significant additional efforts have been made to the make the framework for for purpose:

• Additional Vocabularies were created, mostly for metadata purposes.
• Mappings and relationships were added to achieve comprehensive coverage. If equivalence couldn't be achieved, “uphill” relationships from more granular non-standard to higher level Standard Concepts were created.
• A comprehensive domain structure was established and each Concepts was assigned a Domain (or combination of Domains).
• A hierarchical tree within Domains was built representing classifications used in medical science and clinical practice.