Observational Health Data Sciences and Informatics

This is an old revision of the document!

---

What do I have to have to do an OMOP conversion?

Pre-analysis

1. Data from source (or at least range for each field).
2. Data dictionary including:
   1. taxonomies
   2. OMOP vocabulary (reference codes)
   3. are any fields masked or modified for privacy and if so, what is the algorithm?
3. Identify subject matter expert for source.
4. Identify OMOP subject matter expert.
5. Use cases for destination.
6. List of standard questions:
   1. Number of records
   2. How often is data refreshed?
   3. Is the conversion feasible?
7. Draft ETL Specs.

Meeting with subject matter experts

Post-analysis

1. Initial ETL Spec including:
   1. Business rules for mapping in a detailed specification, preferably in a computer readable format, like White Rabbit.
   2. View of mapping, preferably in a computer readable format, like rabbit in a hat.
2. Identify any additional mapping needed:
   1. custom or local mapping of organizational codes
   2. country codes not already in OHDSI vocabulary like OHIP codes
3. Preferred development methodology

ETL Conversion

1. CDM Builder
2. SQL Development
3. Sample size for development
4. Major pieces of work (Scrum Epics)
   1. Each table
   2. Dimension table first
   3. GRA table always last because derived