FAQ

• Why is demographics count so different from enrollment in i2preport?

Demographics has one entry per patient. The go.sql script selects all patients with a fact since 1/1/2010. The enrollment script inserts one entry per patient for the subset of patients in demographics who also have visits. So patients with facts but no visits will have no enrollment.

• I get primary key errors in the demographics inserts.

This is an unknown bug that causes errors on a handful of demographic entries. It will not affect your overall patient count.

• I get primary key errors in the encounter inserts.

This will cause the procedure to rollback and all other procedures to fail - this must be corrected. Most likely, you do not have globally unique encounter numbers - they are perhaps unique only within a patient_num. This is technically ok in i2b2, but is not recommended and is not supported by these scripts. It is not ok in the PCORnet CDM, which expects each encounter number to be unique. It is possible to remove the encounter constraints to make this work but you do so at your own risk.

• I get data overflow errors in the lab result inserts.

Most likely your tval_char does not match the expected CDM value set. We plan to support value translation at some point in the future. At the moment, either make the database field longer or modify the script to insert UN (unknown).

• No data is being copied to CDM, but I mapped it and it works in i2b2!

There are many reasons for this. Check that you did your mapping in the ontologies and not in the concept_dimension. While the latter works fine in i2b2, it is not supported by these scripts.

• More coming!