Step 3: Build Your Mapping
Here’s where the magic happens—Wave will create intelligent field mappings between your source and target data. This takes about 4-5 minutes.
What You’ll Do
- Ask Wave to create a mapping between your profiles
- Review the auto-generated field connections
- Refine the mapping using the mapping interface
Generate the Initial Mapping
Wave’s AI will analyze both profiles and create smart field connections for you.
-
In project chat, request the mapping:
Create a mapping from FHIR Patient to OMOP PERSON. Map demographics fields and handle birthsex and race appropriately for US Core. -
Wave creates an initial mapping with intelligent connections:
- Patient.id → person_id
- Patient.gender → gender_concept_id (with value transformation)
- Patient.birthDate → birth_datetime
- Patient extensions → race_concept_id and ethnicity_concept_id
-
Review the mapping in the mapping interface—Wave shows you exactly how each field transforms
Understanding Your Mapping
Wave created more than simple field connections—it built intelligent transformations:
| Source Field (FHIR) | Target Field (OMOP) | Transformation Logic |
|---|---|---|
Patient.id | person_id | Convert string to integer |
Patient.gender | gender_concept_id | Map ‘male’→8507, ‘female’→8532 |
Patient.birthDate | birth_datetime | Parse date string to datetime |
Patient.extension[us-core-race] | race_concept_id | Map race categories to OMOP concepts |
Refine Your Mapping
The auto-generated mapping is a great start, but you can refine it:
- Click on any field connection to see or edit the transformation logic
- Add business rules by chatting with Wave:
For missing gender values, default to concept_id 0 (No matching concept) - Handle edge cases Wave might have missed
Pro tip: Wave understands healthcare semantics, so it can suggest appropriate OMOP concept mappings and handle common data quality issues automatically.
What Just Happened?
Wave didn’t just connect fields—it created a structured transformation specification:
- Healthcare-aware mappings - Understands FHIR → OMOP concept translations
- Data type conversions - Handles strings, dates, integers, and concept codes properly
- Business logic capture - Documents transformation rules for your team
- Foundation for code generation - This mapping will drive your ETL code
Check Your Progress
You should now have:
- ✅ A complete FHIR Patient → OMOP PERSON mapping
- ✅ Intelligent field transformations with healthcare context
- ✅ Business rules for handling edge cases
Ready to turn this mapping into working code?