So as we implement OMOP how have folks handled the ingestion of their data by partners? Ultimately that is what partial a driver for us in that we have institutions that want to ingest our data. Do you folks say here it is and let the institution figure out how to ingest things?
What do you mean by that? Whats ingestion? What Partners?
Basically the process of transferring our data to another researcher
Forgive me if I am misinterpreting the question or the context of the question, but you share study analysis results between researchers, not patient level data.
With a standardized data model and vocabulary, you can write a single analysis package (say, in R or raw sql) that expects the vocabulary and data model, and the result of that analysis could be a hazard ratio, incidence rate, characterization, etc. There are many examples of these types of studies in our StudyProtocol repository: https://github.com/ohdsi/studyProtocol. Most of these are implemented as an R package. The way collaborators ‘share data’ is: they each pull down the package, execute the code, and package up the results (usually summary statistics) for sharing. The patient-level data never leaves the boundaries of the individual collaborators. In addition to summary statistics, there may also be diagnostic results that inform the researchers if a given data-source has appropriate characteristics to answer the given research question, which is usually sufficient without needing to look at specific patient records.
Hope that helps, and answers your question.
-Chris
In our case we are sharing patient level data. Perhaps the way we want to do this via OMOP is not ideal but it is what we/I have to work with.
JD