Step 1: Acquire downloads

First, download the necessary components onto a computer that can connect to your target database and from which you can run SQL queries.

• Download i2b2.
  • Download and extract the newinstall newi nstall zip package from "Download Binary Distribution" in the top section of  https://www.i2b2.org/software
  • Determine if you need to install any additional required packages, such as Java JDK or Ant: Chapter 2. Requirements
  • If running less than i2b2 1.8, separately download the latest data package from GitHub https://github.com/i2b2/i2b2-data to get the ACT-OMOP ontologies. 
• OMOP 5.4 DDL SQL Scripts: https://github.com/OHDSI/CommonDataModel/tree/main/inst/ddl/5.4
  • Either clone the repository or download the ETL scripts for your database platform.

Step 2:

 Install i2b2, with ACT-OMOP core tables

, views, and ontologies.

• OMOP: If you will be creating a new OMOP database for this installation, then:
  • Run the OMOP 5.4 DDL SQL Scripts you downloaded in the last step, using a SQL client.
    
    • It is recommended to run the DDL, indices, and primary key scripts, but NOT the constraints scripts. This can cause problems during ETL.
  • This step can be skipped if you already have OMOP on your target database.
• If you have not installed i2b2 previously (i.e. you are not setting this up as an i2b2 Project in a larger installation):
  • 3.5 Hivedata Tables
  • 3.8 PM Table

• i2b2: You will be installing i2b2 as explained in the installation guide. Follow the chapters in the installation guide, but see the notes in the next bullet on Chapter 3.  
• Chapter 3. Data Installation requires special handling. All the steps in the installation guide should be followed except for the following changes in the CRC and Metadata sections::
  • You need to select the act-omop project for Metadata and CRC. Specifically, in 3.4.2 Set Database Properties (CRC) and 3.7.2 Set Database Properties (Metadata), edit the CRC db properties file to set db.project parameter  to act-omop.

• •      
  • Image Added
  • Instead of CRC install step 3.4.

• • 4, drop the patient_dimension and visit_dimension tables from the database. For convenient reference, the correct steps 3.4.3-3.4.5 are summarized here:
    

     ant -f data_build.xml create_crcdata_tables_release_

• • <version no>
    FROM THE DATABASE: DROP PATIENT_DIMENSION; DROP VISIT_DIMENSION;
    ant -f data_build.xml create_procedures_release_

• • <version no>
    ant -f data_build.xml db_demodata_load_data
    

• • 
    The final ant script will create act-omop Views, which point to the OMOP tables, and loads the concept_dimension table with act-omop concepts data

• • . The first two steps create the standard CRC tables and procedures.

Image Removed

Step 3: ACT-OMOP v4 ontology data load process

• • In 3.7.

• • 4, only load the regular data, not the identified data. Run the ant command:

  •                                      ant -f data_build.xml db_metadata_load_data  

    • • This will execute the  SQL scripts from the edu.harvard.i2b2.data\Release_

• • • • <version no>\NewInstall\Metadata\act-omop\scripts\<db type> folder and:
        • Creates ACT-OMOP v4.1 Ontology metadata tables with indexes. 
        • Loads the act-omop metadata
  • Make sure i2b2 is configured for multiple fact tables by changing the appropriate entry in the database.
    • 3.11 Multiple Fact Tables

Step 4: Create Project

Step 3: Advanced Users - Create Project

If an OMOP project will coexist with an i2b2 project, set up an i2b2 OMOP project pointing to the i2b2-OMOP database that you just configured.  

6.6.0. [DRAFT] Guide to creating a new project with the admin tool.

Verify the Ontology is displayed and queriable from the webclient

Notes

Both the ACT-OMOP Ontology tables, views and data is available under edu.harvard.i2b2.data\Release_1-7<version no>\NewInstall\Metadata\act-omop folder.

Image Modified

The scripts are in this folder:

These are the ACT-OMOP ontology tables.

These are the ACT-OMOP views:

• ETL OMOP data
• This is a site-specific process, but OHDSI's WhiteRabbit tool can help: https://github.com/OHDSI/WhiteRabbit