I heard that there may have been an effort to create an ETL script for CDM to i2b2 (not necessarily an effort performed by the OHDSI community). Curious if anyone is familiar with such an effort and could pass along any information… Feel free to re-categorize this message. Thanks.
Al:
Nobody seems to be answering. Let me try waking up the people I know have been thinking about this.
I have been waiting to see what other folks post. Vojtech wrote a paper comparing the systems for AMIA - see http://omop.org/sites/default/files/06_Huser_NIH_IDR%20comparison%20and%20desiderata.pdf and http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3900207/
Yes. My AMIA paper is a bit relevant (it talks about schema differences between i2b2 and OMOP CDM v4.
What is an interesting piece (tha can be separated out) is converting the CDM vocabulary into i2b2 “tree” of concepts.
i2b2 has one main fact table so the ETL is would be sequence of i2b2 views of the Dx, Proc, Rx and lab CDM tables.
I think many sites may have i2b2 shape of data and may want to also have CDM (the other way round)
Thanks all. I appreciate the feedback… We have implemented independent ETL jobs, from our own schema to both i2b2 & CDM and I was hoping the transform from i2b2<->CDM was more straight forward and that we could ultimately scrap one of our, less-than straight forward, processes. Thanks again for the references.