By 2026, more than 800 million patients worldwide have been standardized to the OMOP Common Data Model. FDA Sentinel, EMA DARWIN EU, EHDEN (200+ European data partners), and N3C (the U.S. COVID-19 effort) all run on OMOP. This article explains why OMOP matters and why it fits Vietnam as the country rolls out the Electronic Health Record (HSDT) program under Decision 3516/QĐ-BYT (Ministry of Health digital transformation 2025-2030).
1. RWD vs. RWE — the core distinction
- RWD (Real-World Data): raw healthcare data from EHRs, social-health-insurance (BHYT) claims, registries, wearables, and consumer health apps
- RWE (Real-World Evidence): contextualized, scientifically rigorous insight that can drive clinical or regulatory decisions
For example:
- An RCT (clinical trial) answers "Does drug A work in 1,000 carefully selected patients?"
- RWE answers "How does drug A behave for 1 million Vietnamese patients in real life? How do age, sex, and comorbidity change the picture?"
RWE matters because RCTs cover less than 5% of the clinically important questions.
2. Why a Common Data Model is needed
Five benefits:
- Multi-source analytics: one study runs across 200 organizations in parallel (federated)
- Reproducibility: the same R/SQL code runs identically against any CDM
- Unified vocabulary: ICD/SNOMED/RxNorm/LOINC are pre-mapped
- Open-source tooling: ATLAS, HADES, DQD, ACHILLES are free
- Network: join EHDEN, OHDSI workgroups, and have a community to support you
3. History of OMOP and OHDSI
OHDSI = Observational Health Data Sciences and Informatics. It is not a company — it is an open community (Apache 2.0) with working groups, network studies, and an annual symposium.
4. The 2026 OHDSI stack
5. CDM 5.4 — 37 tables grouped by purpose
| Group | Representative tables |
|---|---|
| Clinical Data | PERSON, VISIT_OCCURRENCE, CONDITION_OCCURRENCE, DRUG_EXPOSURE, PROCEDURE_OCCURRENCE, MEASUREMENT, OBSERVATION, DEVICE_EXPOSURE, NOTE, NOTE_NLP, SPECIMEN, DEATH, EPISODE |
| Health System | LOCATION, CARE_SITE, PROVIDER |
| Health Economics | PAYER_PLAN_PERIOD, COST |
| Standardized Vocabularies | CONCEPT, VOCABULARY, DOMAIN, CONCEPT_RELATIONSHIP, CONCEPT_ANCESTOR, CONCEPT_SYNONYM, CONCEPT_CLASS, RELATIONSHIP, DRUG_STRENGTH |
| Derived Elements | DRUG_ERA, DOSE_ERA, CONDITION_ERA, COHORT, COHORT_DEFINITION |
| Metadata | CDM_SOURCE, METADATA |
PERSON sits at the center — every clinical event references it through the person_id foreign key.
6. Vocabulary — the heart of the CDM
OMOP does not invent its own vocabulary. It reuses international standards and picks one Standard Concept for each clinical idea:
| Domain | Standard Vocabulary | Common source |
|---|---|---|
| Condition | SNOMED CT | ICD-10, ICD-9 |
| Drug | RxNorm (US) / RxNorm Extension | NDC, ATC, Vietnam MoH drug list |
| Procedure | SNOMED CT, CPT4, ICD-10-PCS | Vietnam DVKT (technical service catalogue) |
| Measurement | LOINC, SNOMED CT | local lab codes |
| Observation | SNOMED CT, LOINC | local |
| Unit | UCUM | local |
| Visit | SNOMED CT (visit subset) | local |
ETL must map each source code to a standard_concept_id. The original code is preserved in the matching *_source_value column for traceability.
7. OMOP use cases
| Use case | Real-world example |
|---|---|
| Drug safety | Sentinel monitors post-market side effects of every drug |
| Comparative effectiveness | Metformin + SGLT2 vs. Metformin + DPP4 for diabetes |
| Patient-Level Prediction | Predicting 30-day hospital readmission |
| Disease characterization | Describing the epidemiology of a rare disease |
| Health economics | Analyzing care costs across the BHYT-insured population |
| AI/ML training | Using a standardized cohort as a clinical-LLM training dataset |
8. Comparison with other CDMs
| CDM | Community | Strengths | Weaker than OMOP |
|---|---|---|---|
| OMOP | OHDSI, open | Vocabulary, federated, tooling | — |
| i2b2 | Harvard / community | Built-in UI | Less standardized vocabulary |
| PCORnet | PCORI (US) | Simple, claims-friendly | Limited deep analytics |
| Sentinel | FDA (US) | Drug-safety pharmacovigilance | Closed, FDA-only |
| CDISC | Trial data | Clinical-trial submission | Not built for RWE |
OMOP is the most comprehensive option, used by both the FDA and EMA — that is why it is worth investing in.
9. Vietnam context
Decision 3516/QĐ-BYT (November 2025) — Healthcare digital transformation 2025-2030:
- A national healthcare data system
- HSDT (Electronic Health Record) on VNeID (34M+ records as of January 2026)
- Full-coverage electronic BHYT (social health insurance)
- Encouragement of evidence-based research
OMOP fits a national research data lake very well:
- It can standardize many sources (public hospitals, private hospitals, BHYT, registries) at once
- It enables federated analytics — sensitive data does not need to leave its location
- It is open source — no vendor lock-in
- The OHDSI community is ready to help
A few research groups (VNU-HCM, Hanoi Medical University Hospital, public-health institutes) have started experimenting. There is a huge first-mover opportunity.
10. Does OMOP replace FHIR?
NO — they complement each other. See the HL7 FHIR Practitioner roadmap:
| Criterion | FHIR | OMOP |
|---|---|---|
| Goal | Operational exchange | Analytics / RWE |
| Schema | Independent resources (50+ resource types) | 37 normalized relational tables |
| Vocabulary | Flexible CodeableConcept | Mandatory Standard Concepts |
| Transport | REST API | SQL on a database / file extracts |
| Real-time | Yes (Subscription, CDS Hooks) | No (batch ETL) |
| Use cases | EHRs, mobile, telemedicine | Network studies, ML, BI |
A mature organization should run both — FHIR for the operational layer, OMOP for the analytics layer. See the FHIR ↔ OMOP bridge for details.
11. Where to start?
- Read the Book of OHDSI (free, open source)
- Install Eunomia (a sample CDM R package) and practice SQL
- Look up concepts on Athena (athena.ohdsi.org)
- Spin up Broadsea (Docker Compose with ATLAS + WebAPI + Postgres)
- Join the OHDSI forum
Conclusion
OMOP CDM is the data standard for the next generation of healthcare research. The OHDSI community is open, the tooling is free, and Vietnam has a big opportunity to build a national research data lake. Investing in OMOP today is investing at the right time.
Next article: Comparing OMOP, FHIR, i2b2, PCORnet, Sentinel — which CDM should you pick? · Or browse the OMOP CDM Practitioner roadmap.