I’m on the IQVIA Omop CDM team and I’m trying to get some insight on how the OHDSI community typically does validation. I’m familiar with Achilles but in that case I would like to get some idea of where those rules came from. Thanks!
Achilles pre-computations are based on needs for data characterization and data quality assesement (DQA)
Achilles Heel DQA rules are based on community input. They are not study-specific (done on general level).
Relevent article may be https://www.ncbi.nlm.nih.gov/pubmed/28154833
One approach to validation is defining what we expect to see (specifying that somehow) and checking if we observe the expected.
E.g., we expect life expectancy of women be greater than that of men. (currently not yet implemented as DQA rule in Achilles (exists only as draft SQL query in my scratch DQA code)