Person
Epic equivalent: Patient header / registration demographics
The person table is the center of the OMOP universe — one row per patient. In Epic, demographic details can be split across registration modules, updated with each encounter, and duplicated across systems. In OMOP, all of that consolidates into a single record per individual.
Every other clinical table (condition_occurrence, drug_exposure, measurement, etc.) links back to person_id. This is always your starting point.
Epic-to-OMOP Field Mapping
Field reference (click to expand)
| OMOP Field | Epic Equivalent | What It Captures |
|---|---|---|
person_id |
MRN / Enterprise ID | Unique patient identifier (re-keyed for privacy) |
gender_concept_id |
Sex at birth / Gender | Standardized concept — "Male", "Female", etc. May differ from gender_source_value |
year_of_birth, month_of_birth, day_of_birth |
Date of birth | Split fields to support date shifting for de-identification |
race_concept_id |
Race (registration) | Standardized race value mapped from source |
ethnicity_concept_id |
Ethnicity (registration) | Standardized ethnicity value mapped from source |
location_id |
Home address / ZIP | Foreign key to location table (often de-identified) |
provider_id |
PCP / Managing physician | Primary provider if available; may be null |
care_site_id |
Primary facility / department | Main care site attribution |
*_source_value fields |
Raw EHR text | Original values from the source system (e.g., "M", "Hispanic or Latino") |
What to Watch For
Common pitfalls
- Snapshot, not history
- This table captures the current state. If a patient's address, race, or gender was updated over time, only the latest value is here. For longitudinal demographics, check the
observationtable. - Birth dates may be shifted
- De-identified datasets may truncate or shift dates of birth. Don't assume full precision.
- Use
concept_id, notsource_value - Always use
gender_concept_id,race_concept_id, etc. for analysis. The*_source_valuefields contain raw EHR text that varies across source systems.
Research Patterns
| Question | Tables Involved |
|---|---|
| Proportion of female African American patients in the database | person.race_concept_id + person.gender_concept_id |
| Average age at death | person.year_of_birth + death |
| Statin prescribing disparities across racial groups | person.race_concept_id + drug_exposure |
| Pediatric patients seen for asthma in the past year | person.year_of_birth + condition_occurrence |
| Patient characteristics at a specific care site | person.care_site_id + visit_occurrence |