OHDSI Symposium 2022: There are no "good" mappings.
When: 15th October 2022
Speaker: Nicolas Matentzoglu, Melissa Haendel, Tiffany Callahan
Abstract: The current state of observation data transformation management has resulted in a proliferation of approaches to data normalization and alignment that have accordingly created an explosion of maps. When publicly available, data transformation maps are of varying quality and are often configured to meet single clinical domain or project-specific objectives. Sharing maps in a community requires metadata to help map comparison and evaluation, and release management over different versions. This is typically lacking. In this workshop we will present rationale for translational research data transformation management, including implementation examples mapping Real World Data to OMOP, at-scale. We will discuss the way in which source vocabularies are managed in OMOP and how to improve that process in the future. Data transformations that accommodate divergent underlying ontologies supporting a variety of use cases will be solicited from the participants both in advance and during the workshop.