You can also access the slides from both sessions: Session I | Session II
Chapter 4 of the Book of OHDSI focuses on the OMOP Common Data Model. Below is the introduction to this section, which was led by Clair Blacketer.
Observational data provides a view of what happens to a patient while receiving healthcare. Data are collected and stored for increasingly large numbers of patients all over the world creating what is often called Big Health Data. The purpose of these collections are threefold: (i) directly to facilitate research (often in the form of survey or registry data), or (ii) to support the conduct of healthcare (usually called EHR – Electronic Health Records) or (iii) to manage the payment for healthcare (usually called claims data). All three are routinely used for clinical research, the latter two as secondary use data, and all three typically have their unique formatting and encoding of the content.
Why do we need a Common Data Model for observational healthcare data?
Depending on their primary needs none of the observational databases capture all clinical events equally well. Therefore, research results must be drawn from many disparate data sources and compared and contrasted to understand the effect of potential capture bias. In addition, in order to draw conclusions with statistical power we need large numbers of observed patients. That explains the need for assessing and analyzing multiple data sources concurrently. In order to do that, data need to be harmonized into a common data standard. In addition, patient data require a high level of protection. To extract data for analysis purposes as it is done traditionally requires strict data use agreements and complex access control. A common data standard can alleviate this need by omitting the extraction step and allowing a standardized analytic to be executed on the data in it’s native environment – the analytic comes to the data instead of the data to the analytic.
This standard is provided by the Common Data Model (CDM). The CDM, combined with its standardized content (see Chapter 5), will ensure that research methods can be systematically applied to produce meaningfully comparable and reproducible results. In this chapter we provide an overview of the data model itself, design, conventions, and discussion of select tables.