We encountered an exception when executing a C2 test for CMS22v7 "Preventive Care and Screening: Screening for High Blood Pressure and Follow-Up Documented"

Error

• After uploading our QRDA I deck, Cypress provided the exception: "Calculated value (0) for DENEX (...) does not match expected value (1)" for patient "10_Carolyn_Barker.xml" 
• *** We reviewed the individual patient-level results in Cypress and compared them with our generated results. All matched 100% for each patient. *** We could not see any issue with our logic and the aggregate results matched on the C3 test.

Testing

We then dug into the Cypress test a bit.

• We observed that in Cypress's patient list there were two separate patients with identical names - "Carolyn Barker" - but with different DENEX results (see screen shots below)

• That in itself was interesting. It implies that Cypress sees these as two separate patients.
• Inspecting the two patients showed that there were completely different clinical data. These are NOT duplicate patients according to the specifications. We reviewed the de-duplication guidelines and JIRA tickets and do not believe these are duplicates. They are, unfortunately, named identically.
• We believe this is a key point - Cypress had two identically named patients, but with two different sets of clinical data and two different outcomes. (entire "known good" QRDA I attached)

Additional Validation

• We uploaded our test deck, but with only one of the "Carolyn Barker" patients. Cypress fails, observing that we are one patient short (the second Carolyn Barker).
• We then took the "known good result" QRDA I deck and uploaded that into Cypress. It, too, failed with the same "Calculated value (0) for DENEX (...) does not match expected value (1)" message. Cypress is failing on its own test deck for this test.

Possible Cause

At this point our belief is that Cypress is somehow incorrectly matching these patients during upload. We have to assume it's picking one of the "Carolyn Barker" patients from the QRDA I and mismatching it to one of the two that exist in the patient collection.

Thank you,

Eric