Tessa May - Technical Consultant

Implementing the Observational Medical Outcomes Partnership (OMOP) model presents an exciting opportunity to leverage real-world healthcare data for research and analysis. Recently, we have shared our thoughts on the benefits the OMOP model can introduce, and the challenges in implementing it. Successful implementation requires careful scoping, mapping definition creation, and subsequent deployment. In this post, we provide some insight on our process for effectively implementing the OMOP model.

Not sure what the OMOP CDM is? Please see our previous post!

Mapping source data to the OMOP model is complex and requires careful consideration. A collaborative, systematic approach, with robust documentation and governance can increase efficiency, accuracy and consistency. At a high-level, we have identified three key stages to be considered:

• Scope: What is your use case? What are your requirements? What data is available and how accurate is it? This stage is all about defining the scope of the source data that will be covered in the mapping. It is important to ensure mapping is comprehensive, while being achievable and manageable. It is unlikely to be feasible to map a data output from an entire EPR in one go.
• Define: The tricky, but interesting bit! Here is where you get into the weeds of the OMOP CDM. Defining, documenting and validating the mapping from the source data to the target model. It is important to consider aspects such as system and organisation context, clinical definitions, data best-practice and other complexities in the mapping. Mapping should be clearly documented, reviewed and validated, both to ensure it can be referenced in other mapping exercises and to provide traceability and transparency for end users of the data.
• Implement: Now to put it into action… Time to design, implement and test the mapping process, which forms the pipeline for converting source data into the live OMOP database. This stage should include consideration of the wider architecture and data security; the full technical solution beyond just the mapping definition.

The size of each stage may depend on your delivery approach. The process can be agile and iterative, for example keeping the scope small and delivering a full implementation in a sprint. Alternatively, the scope may be larger and the “Define” stage consists of multiple sprints before implementation begins.

Throughout the process, clear documentation and decision tracking is imperative for traceability. It will increase efficiency as key decisions can be referenced in subsequent mappings, and it will also enable researchers to produce more accurate results as they can consider the transformations of the data from its source.

Ready to take the first step towards effectively implementing the OMOP model for your healthcare data analysis? Contact us today to discuss your specific needs and embark on this transformative journey!