Need help with implementing OMOP CDM on local system

Implementers
1

Hi,

I am Ashwini , working as a data analyst with an healthcare organization and new to this forum. I am planning to install and implement the OMOP CDM on my local machine. Could you please guide me on this.
What are minimum requirements of the machine (RAM, Memory) to have OMOP implemented on the system.
what are the basic application i need to download from ODHSI.
I can work with SQL Server or PostGre SQL for CDM . My source data is relational DB.

Thanks,
Ashwini

2

I would start with The Book of OHDSI. That will answer a lot of questions.

1 Like
3

Thanks Roger . I have started going thru the book. My question is it possible to install the OMOP CDM on local system. If yes what are the applications i need .

4

The Broadsea docker implementation is often recommended as a good place to start.

5

It depends on what you mean by “OMOP CDM”. As Chris said, there’s a whole stack of technologies/applications that makes up OMOP.

However, if you’re just looking to create an OMOP compliant database, that’s a lot easier.

  1. Go to the OMOP GitHub Wiki for the defintions: https://github.com/OHDSI/CommonDataModel/wiki

  2. Find the release you want: https://github.com/OHDSI/CommonDataModel/releases

  3. For each release, there is a file called Source Code (zip) which has the SQL DDL to create the tables, constraints, and indexes. (for v 6.0 it’s on the page above. for other versions, click on the version title)

  4. Download vocabularies from Athena here: http://athena.ohdsi.org/vocabulary/list

  5. Load the vocabulary files (they are quite large) into your OMOP database. There are DDL scripts for this too also in the Source Code zip file in the VocabImport folder

  6. You also might want to check out the ETL tutorial on this page: https://www.ohdsi.org/2019-us-symposium-page/

  7. For mapping proprietary codes from your source, USAGI is a useful tool: https://www.ohdsi.org/web/wiki/doku.php?id=documentation:software:usagi

  8. After that, it’s a matter of querying your source data, mapping it to appropriate concepts, and writing it into the standard data tables. But definitely check out the tutorial.

6

Thanks Roger. your reply cleared my confusion and helped to go with correct option.

7

Look at Eunomia package.

Or surgicaly remove just the sql lite file here https://github.com/OHDSI/Eunomia/blob/master/inst/zip/cdm.zip

That is CDM (with data) as a single file. It is super cute. I am a big fan of Martijn for developing it!

8

Since I had started down the path of setting up a CDM outside of Broadsea, when I wanted to run more of the tools, I had to work through that. I ended up putting together a script that installs a number of the pieces from github repositories. This is where you start addressing questions about the DDL for the various applications. Happy to share and discuss with anyone interested.

1 Like
9

Sure Chris. Thanks for letting me know about this. I am happy to know your experience