THIS FEATURE WILL BE AVAILABLE MARCH 2023
Objective
Utilize the concept of a reusable application programming interface (API) by using i2b2 API to adapt and support OMOP data model. Below steps involve creating datamodel-specific i2b2 ontologies and mapping these to query non-standard OMOP table structure
Steps to build ACT-OMOP Ontology
Step1: Create OMOP tables and load data using:
- OMOP 5.4 DDL SQL Scripts: https://github.com/OHDSI/CommonDataModel/tree/main/inst/ddl/5.4
- It is recommended to run the DDL, indices, and primary key scripts, but NOT the constraints scripts. It can cause problems during ETL.
- ETL OMOP data
- This is a site-specific process, but OHDSI's WhiteRabbit tool can help: https://github.com/OHDSI/WhiteRabbit
Step 2: Create ACT-OMOP CRC and i2b2 core tables (e.g., QT and breakdowns)
- Download i2b2. (Download and extract the newinstall zip package from "Download Binary Distribution" in the top section of https://www.i2b2.org/software)
- Follow the instructions in Chapter 3 of the i2b2 installation guide to create the additional i2b2 tables needed for the application server: Chapter 3. Data Installation, these sections specifically:
- 3.4 Crcdata Tables
- In 3.4.2, edit the CRC db properties file with ACT-OMOP project parameter which will be used to build the concept_dimension table from metadata tables later in the process.
- Skip section 3.4.5 for now.
Step 3: ACT-OMOP v4 ontology data load and Views creation process
- Follow the instructions in Chapter 3 of the i2b2 installation guide to load the ACT-OMOP metadata project.
- 3.7 Metadata Tables
- In 3.7.2, edit the Metadata db properties file with ACT-OMOP project parameter which will direct the ant script to load the ACT-OMOP ontology. In db.properties, set db.project=ACT-OMOP
- In 3.7.3, do they skip CREATE METADATA????
- 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_1-7\NewInstall\Metadata\act_omop\scripts\<db type> folder. This will:
- Create and load ACT-OMOP v4 Ontology metadata tables with indexes.
- Create the ACT-OMOP views by executing BUILD_ACT_OMOP_VIEWS_ALL_MSSQL.sql
The Lab metadata tables should have data in the C_METADATAXML column in order for the lab values to be queriable from UI
If the C_METADATAXML is missing from the source ontology Lab metadata file, you can load the values from the standard i2b2-ACT Lab metadata table.
Step 4: Build ACT-OMOP CONCEPT DIMENSION
Perform step 3.4.5. This must be done after the metadata are loaded. From the edu.harvard.i2b2.data\Release_1-7\NewInstall\CRC folder, run the ant command: ant -f data_build.xml db_demodata_load_data
This will execute the create_omop_act_concept_dimension SQL script from the edu.harvard.i2b2.data\Release_1-7\NewInstall\Crcdata\scripts\act-omop\<db type> folder and loads A concept dimension table
Step 5: Create Project
Finally 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
The concept_Dimension is populated from metadata tables.
Verify create_omop_act_concept_dimension.sql to ensure it has complete metadata tables information before running the script
Example: the ACT/OMOPVital_Signs information should be in the script to load the ACT/OMOPVital_Signs concept data
Notes
Both the ACT-OMOP Ontology tables, views and data is available under edu.harvard.i2b2.data\Release_1-7\NewInstall\Metadata\act-omop folder.
The scripts are in this folder:
These are the ACT-OMOP ontology tables.
ACT_ICD10CM_DX_V4_OMOP ACT_ICD10PCS_PX_V4_OMOP ACT_ICD9CM_DX_V4_OMOP ACT_ICD9CM_PX_V4_OMOP ACT_CPT4_PX_V4_OMOP ACT_HCPCS_PX_V4_OMOP ACT_MED_ALPHA_V4_OMOP ACT_MED_VA_V4_OMOP ACT_MED_VA_V4_OMOP ACT_LOINC_LAB_PROV_V4_OMOP ACT_SDOH_V4_OMOP ACT_VITAL_SIGNS_V4_OMOP ACT_VISIT_DETAILS_V4_OMOP ACT_COVID_V4_OMOP ACT_DEM_V4_OMOP ACT_ICD10_ICD9_DX_V4_OMOP |
These are the ACT-OMOP views:
CONDITION_VIEW; DRUG_VIEW; MEASUREMENT_VIEW; OBSERVATION_VIEW; PROCEDURE_VIEW; COVID_LAB_VIEW; ALL_SOURCE_CONCEPTS; PATIENT_DIMENSION; VISIT_DIMENSION; DEVICE_VIEW; |