In the modern world, it is impossible to push forward the progress on your own. Leading researchers and research labs join dedicated communities to share their knowledge and best practices and to find answers to long-standing problems. In healthcare, one of such communities of researchers is OHDSI — and in this blog post we’ll talk about it and the role the Sciforce teams plays in it.
What is OHDSI?
If we look at how large-scale analytics in healthcare works, we’ll see that it is ruled not only by corporations, but, rather, is moved forward by communities of researchers who share similar interests. OHDSI (pronounced as “Odyssey”), or Observational Health Data Sciences and Informatics, is an umbrella initiative for such researchers and the successor of the Observational Medical Outcomes Partnership (OMOP) defined as “multi-stakeholder, interdisciplinary collaborative that is striving to bring out the value of observational health data through large-scale analytics.”
Its objective is to enable analysis and sharing of health or observational data between different institutes and companies. OHDSI has created and supports the Common Data Model (CDM) the goal of which is to standardize seemingly disparate databases that might be used in a variety of research areas, such as comparing alternate treatment paths, personalized medicine, product safety, and overall quality improvement in healthcare.
How does it work?
Observational outcome studies leverage secondary electronic data that might come from diverse sources such as Electronic Health Records, bills, claims, public health sources, biobanks inventories, or pharmaceutical sources, in order to perform large-scale evidence-based longitudinal patient-level clinical studies. However, in most cases, the available data is so versatile that cannot be used directly, but should be first transformed into a standardized warehouse of information. For this, OHDSI focused OMOP’s design on three areas of harmonization: a common format (i.e., the data model), a common representation of vocabularies, and systematic analyses using a library of standard analytic routines that have been written based on the common format.
Similar to other data warehouses, OHDSI transforms data to fit a general data model: the OMOP common data model (CDM). Its central entity is the patient and it contains tables for the data commonly needed in clinical trials and observational studies such as drug use, procedures performed etc. These tables are grouped together in the “clinical data” part of the data warehouse. Besides, the CDM includes the major commonly used ontologies: SNOMED, Loinc, RxNorm etc.
For the actual transformation of the data any preferred language and application can be used, such as SQL or Python). Once the data is transformed and loaded into the OMOP CDM database, it is easy to get it from the database, share analyses with other groups using OHDSI or run them on their OMOP data sets, without necessarily sharing the actual data.
In this way, the CDM model enables clinicians and researchers to systematically identify specific cohorts, compare results, reproduce protocols using data, and investigate combinations of interventions and outcomes.
OHDSI Community
One of the cornerstones of OHDSI is its openness both in their findings, including sharing methods, tools and evidence, and in their networking. It is not an organization, but a real community, trying to improve health outcomes for patients around the world together. Whether you’re a software developer, physician or clinical researcher, there is a place for everyone in the OHDSI community.
Its central coordinating center is at Columbia University, but it is fueled by data from almost 100 organizations across dozens of countries, and we are happy to be among them.
Where does Sciforce fit?
As an IT company with a broad expertise in the development of solutions for healthcare, we joined OHDSI in 2014 to develop and improve the CDM model so that it can cover more cases, understand and analyze data according to the customer’s needs. To achieve this, we map data to a single format, for instance, with the same variable name, attributes, and other metadata. We also are engaged in the development of the general platform that enables much more rapid responses to research-related questions.
Besides, with the CDM, we also perform predictive analysis and observational studies, proving its viability and usefulness in clinical research.
In these years, our team members have prepared 8 posters for OHDSI symposiums, authored or co-authored 3 scientific articles and ran dedicated training sessions. In 2018, our team members even received the Titan Award for Data Standards for their contribution to development and evaluation of community data standards, including OMOP common data model and standardized vocabularies.
These are remarkable achievements, but it is also a big step for our company — to be recognized in the community of health IT researchers. It’s a big adventure and a long journey and it feels exciting to be on board and be praised as valuable members of a project of such a scale and importance.
