Because of discussion here: https://github.com/OHDSI/Achilles/issues/102
I would like to introduce the idea of “multiple levels of compliance” to CDM.
ETL is a lot of work and we should not force adopters to populating every possible table.
Some past historical studies often require only some tables. For example, exposure tables (not era tables).
Some datasets don’t have observation_period table or cost tables.
When is a dataset minimally compliant with CDM?
To at least start with some proposal:
I would propose that the dataset must have (for level 1) (for v5) at least the following tables: person, death, visit_o,condition_o,procedure_o,drug_exposure,measurement
(in level 2) I would add derived tables: drug_era, condition_era
(in level 3): cost tables, some other aspects.
HIMSS has 7 levels of adoption of EHR.
Why not have the same for “OHDSI data warehouses”?