Stream: CARIN IG for Blue Button®
Topic: Test set
Shamil Nizamov (Jan 08 2021 at 01:01):
Is there any min/max data test set to test BB implementation? May there is one from connectathons.
Josh Lamb (Jan 08 2021 at 17:40):
Are you talking about boundary testing? There is a minimal set of example models included in the examples section of the IG. There are also initiatives to produce a CPCDS dataset using Synthia, but I doubt the tooling to convert from CPCDS to FHIR will be freely provided.
Saul Kravitz (Jan 11 2021 at 17:33):
There should be a test set developed using Synthea and other tools (as @Josh Lamb points out, some of the tooling is proprietary) available at the connectathon. At least, that is what we are working towards.
Richard Braman (FLY.HEALTH) (Mar 30 2021 at 12:24):
Saul Kravitz said:
There should be a test set developed using Synthea and other tools (as Josh Lamb points out, some of the tooling is proprietary) available at the connectathon. At least, that is what we are working towards.
Saul was the connectathon test data ever published? I worked very hard to make that data set valid, with the intent that it would be shared with the community. The tooling to produce FHIR from CPCDS is propritary for sure, but the outputs were meant to be shared.
Saul Kravitz (Mar 30 2021 at 19:59):
hi @Richard Braman (FHIRFLY.IO) -- the test set was not published, per se.
There were several remaining issues to iron out:
1) there were some validation issues
2) for loading into HAPI, instance ids need to start with an alpha character
Given the timeline for the last connectathon, Onyx decided simply to load the data into their SAFHir server. We have the data that was delivered, but have not tried to push that project forward. @Corey Spears @Dave Hill
Saul Kravitz (Mar 30 2021 at 20:52):
Here are the issues that I discovered the last I looked at this dataset:
I've spent some time reviewing the data, and have identified the following (probably not exhaustive) set of issues with the "CarrinBBExtract" data:
· EOB.supportingInfo.category -- display text is right, but categories should have system=http://hl7.org/fhir/us/carin-bb/CodeSystem/C4BBSupportingInfoType and an appropriate fixed code (e.g., billingnetworkcontractingstatus) from that system.
· mismatch of _history references with versions supplied ( Patient, Practitioner, Organization). Only one version of each resource is supplied. Sometimes the reference is to /2, and the data provided is /3.
· EOB item.adjudication.amount should have a $ value
· EOB.item.modifier includes only system
· EOB.item.locationCodeableConcept includes only system
· Profile compliance includes both abstract base profile,and concrete profile, which is unnecessary.
· EOB.procedures include references to devices (e.g., 261352009) not procedures.
· Some references to Patient are versioned (Coverage), and others aren't (EOB)
· Display values (e.g., Organization, benefit category) are just the resources ID. Not as useful in a demonstration as data with display values.
· ids are numerical, need to start with alpha to be loadable in HAPI(easily rectified by scriptology on our end)
Last updated: Apr 12 2022 at 19:14 UTC
