Processing OMOP CDM exports¶

Image extracted from www.ohdsi.org

OMOP CDM stands for Observational Medical Outcomes Partnership Common Data Model.

You can find a link here to OMOP CDM documentation.

Step 1: Install `Convert-Pheno`¶

We recommend downloading from CPAN.

First, install system-level dependencies:

sudo apt-get install cpanminus libbz2-dev zlib1g-dev libperl-dev libssl-dev

We will install Convert-Pheno and its dependencies at ~/perl5:

cpanm --local-lib=~/perl5 local::lib && eval $(perl -I ~/perl5/lib/perl5/ -Mlocal::lib)
cpanm --notest Convert::Pheno
convert-pheno --help

To ensure Perl recognizes your local modules every time you start a new terminal, you should type:

echo 'eval $(perl -I ~/perl5/lib/perl5/ -Mlocal::lib)' >> ~/.bashrc

Optional: Installing Athena-OHDSI database

If your OMOP CDM data is not self-contained, you might need to download the Athena-OHDSI database.

The database file is available at this link (~2.2GB). The database may be needed when using -iomop.

See latest download installations here.

Once downloaded, use the options --ohdsi-db and --path-to-ohdsi-db

We'll be using the convert-pheno command-line interface to consolidate all rows from a given individual into a single object.

Usage:

convert-pheno -iomop omop_dump.sql -obff individuals.json

If you have to use the Athena-OHDSI database:

convert-pheno -iomop omop_dump.sql -obff individuals.json --path-to-ohdsi-db ./

Additional options

Convert-Pheno has more options. Find a more detailed explanation at the documentation.

Now, you can use the standard nomenclature for running jobs.

Cohort modePatient mode

pheno-ranker -r individuals.json

pheno-ranker -r individuals.json -t target.json