We are planning a study and database to develop deep learning algorithms based on mammograms, and relevant patient characteristics clinical data. I think it would be of value to be able to connect to other datasets. Anyone working with OMOP / CDM in the setting of breast cancer screening?
Not specifically but it’s certainly a very good use case for OMOP! 
A few caveats:
-
To explore breast cancer screening you would ideally want to understand the gap between orders and actual mammograms. If you are working with claims data, you will only have information on the latter. If you are working with EHR data, you may be in better shape.
-
Typically the results of the screening mammogram will be in free text so NLP may be valuable if you have these data. If your study is not dependent on the outcome of the screening, but rather just obtaining the screening, then you will be able to utilize the codes alone.
-
Remember that patients who appear to need mammograms by age may have legitimate reasons why these studies do not appear in their data-- e.g., prior mastectomy, obtained via mobile screening (off the grid), and so forth.
Best of luck
Jon