Building a global network for better evidence.
[ ABOUT ]

What is OHDSI?

The Observational Health Data Sciences and Informatics (or OHDSI, pronounced "Odyssey") community strives to bring out the value of observational health data through large-scale analytics. We are a multi-stakeholder, interdisciplinary collaborative spanning many disciplines (e.g., clinical medicine, biostatistics, computer science, epidemiology, life sciences) and user groups (e.g., researchers, patients, providers, payers, regulators).

We aim to maximize the value of real-world data (RWD) by addressing challenges such as lack of data interoperability and reproducibility issues in health research by enabling robust multi-database studies through a federated approach that preserves patient privacy and respects local data governance.

How the OHDSI approach works (high level)

Old-school approach (study-specific datasets and code)
Traditional (non-standardized) approach
In traditional (non-standardized) observational research, analyses rely on study-specific minimal datasets that are manually prepared and shared by each data source. Because every study has a unique protocol, bespoke data transformations and code must be recreated for each study, often independently at each site. This leads to slow, expensive, non-transparent workflows that are difficult to reproduce and do not scale.
OHDSI / OMOP approach (standardized data, federated analytics)
OHDSI / OMOP approach
With the OMOP Common Data Model, data partners map their local data to a shared standard that stays on-site, enabling federated analytics. Once mapped, all data looks the same, so a single analysis script can run unchanged across every OMOP-mapped database. Shared code and local execution make the process transparent, scalable, and reproducible—enabling faster and more reliable evidence generation.

Why the OHDSI approach works:

Large-scale evidence generation

OHDSI enables federated analyses across observational health databases worldwide. Studies can be executed locally and results combined, without sharing patient-level data—supporting robust evidence across countries, populations, and care settings.

Rapid response to clinical questions

Standardized data and reusable analytical pipelines reduce the time and effort needed to answer new clinical and regulatory questions—so reliable evidence can be generated quickly when it matters most.

Scalable and reproducible research

Open-source tools built around a common data model allow studies to be replicated and run across multiple databases and diseases—improving transparency, consistency, and long-term scientific value.


The key building blocks of OHDSI:

Standardized data

The OMOP Common Data Model harmonizes diverse sources such as electronic health records, claims data, and registries into a single, consistent structure.

Standardized vocabularies enable both syntactic and semantic interoperability, making heterogeneous data comparable and analyzable with the same tools.

Learn more about the OMOP Common Data Model Explore the standardized vocabularies

Standardized analytics

OHDSI provides a comprehensive suite of open-source tools that support the full observational research lifecycle—from ETL and data quality assessment to cohort definition, characterization, prediction, and effect estimation.

ATLAS offers a user-friendly interface for common analyses on OMOP CDM data without programming skills. For advanced and customizable workflows, HADES provides a collection of R packages designed for reproducible and scalable analytics on OMOP CDM databases.

All OHDSI tools are freely available and continuously developed by the community.

Try ATLAS Explore HADES Browse OHDSI tools on GitHub

A global, collaborative community

OHDSI is powered by an active international community of researchers, clinicians, and data partners. Collaboration across institutions and countries strengthens study quality, accelerates learning, and avoids working in isolation.

Join a workgroup, attend our annual symposium, or connect with the community in your country via the Get Involved page.

Get involved