OMOP CDM and Advancing Thailand’s Health Research to Global Standards

When Patient Data from Multiple Hospitals Cannot Communicate with Each Other

Every hospital has a vast amount of patient data, ranging from medical histories, diagnoses, medications, and laboratory test results to information about healthcare visits over different periods. However, when this data is used on a larger scale—for example, to analyze data from multiple hospitals, conduct clinical research, or generate Real-World Evidence (RWE)—a major limitation becomes apparent: the data does not speak the same language. For example, a research team may want to study patients with type 2 diabetes from multiple hospitals to analyze treatment trends or health outcomes among this patient population. When data from each hospital is collected, the same diagnosis may be recorded differently, such as:

• Hospital A uses DM Type 2
• Hospital B uses the code E11 based on the ICD-10 standard
• Hospital C uses the code 250.00 from ICD-9

For healthcare professionals, these data points have the same meaning. However, for computer systems, all three may be interpreted as different values. Without a common standard to connect the meaning of the data, combining data from multiple sources requires extensive Data Cleansing and Data Mapping. IT teams need to write new Transformation Logic for each project, while researchers spend more time preparing data than analyzing it. This has become a limitation faced by healthcare systems around the world and is one of the reasons behind the development of the OMOP Common Data Model (OMOP CDM).

OMOP CDM: A Standard That Enables Health Data to Speak the Same Language

OMOP CDM (Observational Medical Outcomes Partnership Common Data Model) is a healthcare data standard that enables data from different systems to be stored and analyzed in a common format. It is developed and used within the international research community OHDSI (Observational Health Data Sciences and Informatics), pronounced “Odyssey,” like the Greek legend or Christopher Nolan’s film. The goal of OMOP is not to replace the HIS or EMR systems used by hospitals. Instead, it provides a common standard layer for data, allowing data from different source systems to be used together.

From Raw Data to OMOP CDM: The Standardization Process

Bringing healthcare data into OMOP CDM consists of three main steps.

1. Extract: Extract Data from Source Systems

Data is extracted from various systems within an organization, such as Hospital Information Systems (HIS), Electronic Medical Records (EMR), Laboratory Information Systems, and other specialized systems. At this stage, the data remains in the original format of each system, which may differ in both structure and how information is recorded.

2. Transform and Vocabulary Mapping: Converting Data into a Common Standard

Diagnosis data, medications, laboratory test results, and other clinical data must be transformed into the OMOP standard by linking them to a Concept ID.

A Concept ID serves as a representation of a common meaning, allowing differently written data to be interpreted as the same concept. OMOP supports important Standard Vocabularies, including:

• SNOMED CT for Clinical Concepts
• RxNorm for medication data
• LOINC for laboratory test results

3. Load: Storing Data in the OMOP CDM Standard Structure

After Mapping, the data is stored in standardized structures such as Person, Visit Occurrence, Condition Occurrence, Drug Exposure, Measurement, and Procedure Occurrence. Once data from multiple organizations is stored in the same structure, analytical tools and research processes can work with the data without having to redesign everything from scratch each time.

OMOP Is Not Just a Data Model, but an Ecosystem for Health Research

What makes OMOP CDM different from general data standards is that OMOP does not stop at defining the data structure.

Once data has been transformed into OMOP CDM, organizations can access the OHDSI Ecosystem, which includes tools and research methodologies collaboratively developed by researchers around the world. Examples include:

• ATLAS: A tool for creating Cohorts and designing research studies
• Achilles: A tool for assessing data quality and providing an overview of the data
• HADES (Health Analytics Data-to-Evidence Suite): A suite of tools for statistical analysis and Real-World Evidence

When Researchers No Longer Have to Wait in Line for Data Engineers

The most significant change is who can take action in the research process.

Previously, when a researcher had a question—for example, wanting to know the incidence rate of stroke among patients with type 2 diabetes who received a particular drug—the process would typically begin with a data request. The researcher would then wait for the IT team to write a Query, wait for the results to be reviewed, and, if the inclusion or exclusion criteria were found to be inappropriate, go through the process again. This process could take weeks or even months for a single research question.

Once the data is on OMOP CDM, researchers can use ATLAS to create Cohorts themselves, define the Index Event, specify Inclusion and Exclusion Criteria, and define Time-at-risk without writing SQL. Users can define Cohort Definitions with varying levels of complexity and run them against the database to immediately see the number of patients meeting the criteria at each stage. They can then extend the analysis to other modules, such as Cohort Characterization, Incidence Rate, or Treatment Pathways, to gain an overall view of the population of interest.

As a result, asking a question and obtaining preliminary results can take anywhere from minutes to several hours, depending on the size of the data and available system resources. Researchers can therefore experiment with different criteria, test hypotheses, and validate the reasonableness of the results themselves before deciding to invest in a full-scale study. The role of Data and IT teams consequently shifts from responding to project-by-project Query requests to maintaining Pipelines and Governance so that data remains continuously ready for use.

The More Organizations Use the Same Standard, the Greater the Value

The benefits of OMOP CDM become most apparent when multiple organizations use the same standard. A Cohort Definition created in ATLAS can be exported as a standard file, such as JSON or SQL, and processed again by another organization using OMOP CDM without having to redesign the criteria or write new Transformation Logic.

This approach enables research models that previously involved very high costs to become practical and achievable in a shorter amount of time, such as:

Multi-center Study: Hospitals within the same network or across multiple institutions can run the same analytical package and compare results based on consistent definitions.

Federated Study: Each organization analyzes data within its own system and sends only aggregated results without transferring raw patient data across organizations. This helps reduce privacy risks and supports requirements such as PDPA.

National and Regional Study: When the network expands to the national or regional level, policy questions or drug safety questions can be addressed using large-scale population data, while significantly reducing the time required for preparation.

In other words, investing in OMOP CDM does not generate value only for an individual organization. It also enables that organization to become part of a research network that is ready to collaborate as soon as important questions arise.

Why OMOP CDM Matters to Thailand Now

Thailand’s healthcare system is facing several major transitions at the same time, including the 30-Baht Healthcare Scheme, the expansion of Health Data Exchange between organizations, and the growing demand for Real-World Evidence (RWE) research from both Regulators and Pharmaceutical Companies.

Using data at this scale requires data from multiple sources, including hospitals, universities, and health-related organizations, in order to provide broader coverage and better reflect real-world situations. The challenge is that healthcare data from each organization remains different in terms of data structures, codes, and storage methods based on the HIS or EMR systems in use. Without a common standard, connecting data requires new ETL development and Mapping for each project. This increases the time required to initiate research or data analysis and means that the large amount of existing data cannot yet be fully utilized.

What Happens When Data Enters the OMOP CDM Standard

The difference between data before and after Standardization is not limited to the data format. It changes how organizations work. OMOP CDM does not require every organization to use the same system. Instead, it enables data from different systems to connect through a common standard.

Before Using OMOP CDM	After Standardization with OMOP CDM
Preparing data for RWE takes a long time	Data has undergone Standardization and is ready for analysis
Researchers have to wait a long time for data preparation from each organization	Standardized Queries can be used on the same data structure
Cross-hospital research requires significant time to standardize data (Data Cleansing)	Data from multiple organizations can work together under the same standard
IT teams need to develop new Transformation Scripts for each project	Standard approaches and tools from OHDSI can be used
Data is not ready for international research	Organizations can participate in research networks using OMOP CDM

Challenges of Building OMOP CDM at the Organizational Level

Although the concept of OMOP CDM may appear straightforward, the actual implementation can be highly complex. Converting data into OMOP CDM is not simply a matter of renaming Fields or restructuring tables. It requires an understanding of the source data structure, the meaning of the data, and medical standards. The key challenges include:

Continuously Increasing Data Volumes

Medium-sized and large hospitals have vast amounts of patient data accumulated over many years. Processing this data requires systems that can handle high data volumes and Scale according to data growth.

Complex Vocabulary Mapping

Data from each hospital may use internal codes or different recording formats. Mapping these data to Standard Vocabularies such as SNOMED CT, RxNorm, and LOINC requires systematic management to ensure that the meaning of the data is accurate.

Data Quality and Data Governance

Healthcare data is highly sensitive. Using this data for research requires data quality controls, access control, and operations that comply with privacy requirements such as PDPA.

Long-term Pipeline Management

Healthcare data is constantly changing. As new patient data enters the system, the system must be able to continuously Update the data into OMOP CDM without having to create a new process each time. Therefore, building OMOP CDM at the organizational level requires a Data Platform capable of supporting data processing, Pipeline management, and Governance simultaneously.

The Role of Blendata Enterprise in Building OMOP CDM

Blendata Enterprise is not an OMOP CDM standard and does not replace the OHDSI Ecosystem. Instead, it serves as an Enterprise Data Platform that helps organizations efficiently build and manage Data Transformation processes for OMOP CDM. Within the OMOP ETL process, the platform needs to support everything from extracting data from multiple systems and processing large volumes of data to performing Transformation and managing Workflows so that processes can be systematically repeated. Blendata Enterprise can support these requirements through capabilities such as:

Distributed Data Processing

Provides Scale and supports the processing of large volumes of data from multiple sources without requiring organizations to build the Infrastructure themselves.

ETL Pipeline Management

Helps create and manage ETL processes for transforming data into OMOP CDM, allowing processes to be systematically monitored, reviewed, and rerun.

Visual Workflow Monitoring

Helps teams such as Healthcare IT and Clinical Informaticists visualize the status of data processing and easily review different steps in the process.

Data Governance

Supports systematic data management, including Data Catalog, access control, and data governance in accordance with organizational requirements.

This approach reduces the burden on Data and IT teams of developing ad hoc Pipelines for each project and transforms the process of building OMOP CDM into Data Infrastructure that can be continuously used over the long term.

When data management processes are standardized and can be continuously maintained, organizations can fully leverage the potential of healthcare data while supporting future research, collaboration between institutions, and data analysis.

If you would like more information, please contact the Blendata team at:

Email: hello@blendata.com 

Website: www.blendata.com