Getting Started Clinical Research with InterSystems Automated OMOP Data Pipeline for Baystate Research Data Repository

Automated OMOP Data Pipeline for Baystate Research Data Repository

Overview

As a 5 hospital integrated delivery network in western Massachusetts, Baystate Health (Baystate) is also a teaching hospital for U Mass Medical School.  The research department has planned to enhance the research operation and collaboration with other regional medical centers including UMass Memorial Health and Tufts Medical Center.  Such collaboration will help the medical center research groups to conduct more sponsored studies and obtain associated fundings as additional revenue sources to the medical center.  The research team identified a critical technology requirement is the availability of a research data repository that is based on OMOP common data model (CDM).  OMOP is used by over 3,000 organizations around the world.  Many NIH fundings require OMOP data format.  OMOP community (OHDSI) is very active with many organizations contribute to the ecosystem.  Therefore, budgets are allocated accordingly to fund the development and maintenance of OMOP research data repository.

When the research team contacted the hospital IT, they found out that the main EHR system does not provide an easy way of creating research data repositories in OMOP format.  While the EHR vendor offers its own centralized research data lake with de-identified data, Baystate’s research partners do not use the same EHR solution.  On the other hand, the hospital IT team suggested that the research team leverage an innovation sandbox built on InterSystems HealthShare technology that already stores all the data from the main EHR system plus more.  This clinical data repository can be used to analyze data quality prior to feeding into OMOP repository.  In addition, data quality issues might be fixed upstream at ingestion.

Although there are open source OHDSI ETL tools available, Baystate research team would prefer a commercial data pipeline solution that can be supported by reliable vendors.  The research team would focus on using OHDSI analytics tools such as HADES instead of learning ETL tools.  With InterSystems recommendation, Baystate research team contracted Clinovera that has implemented OMOP solution at many sites globally.  Clinovera is a certified InterSystems technology solution partner and already worked with the other health systems that are similar to Baystate.

A key requirement for the OMOP data pipeline solution is the need to have incremental data refresh instead of the typical “purge and rebuild” approach used by many other sites using OMOP ETL tools.  Clinovera developed a fully automated data extraction, transformation, and insertion process using InterSystems IRIS for Health product.  The goal is to achieve weekly data refresh and evolve towards daily refresh so the system can be used to recruit patients for active clinical trials.

The solution went live in April 2023.  Baystate research team are able to use OHDSI tools to analyze hospital data patients to design studies and collaborate with other regional medical centers.