search

num-codex / odm2fhir

This tool maps study/patient data in CDISC ODM based on the GECCO data dictionary to HL7 FHIR which adheres to the GECCO profiles, value sets and code systems.
MIT License
4 stars 1 forks source link

Conditions for "unanswered" questions should not be generated #13

Open noemide opened 3 years ago

noemide commented 3 years ago

Describe the bug Currently, if questions about the presence of a condition (e.g. chronic lung diseases) are not answered at all, it is still exported without any verification status. Since those conditions are typically interpreted as being present, this leads to false information being exported.

To Reproduce Export empty for unaswered questions about preexisting Conditions.

Expected behavior For unanswered questions, no Conditions should be exported.

@makampf

cerbelding commented 3 years ago

Thank you for your report. I will have a closer look and come back to you.

holger-stenzhorn commented 3 years ago

@noemide Actually, I had a quite similar question this very morning which I posted on Zulip (https://chat.fhir.org/#narrow/stream/179183-german-%28d-a-ch%29/topic/GECCO-Datensatz/near/224294828) but have not received an answer yet...

makampf commented 3 years ago

I do not get the differentiation between unknown and not answered.

Why not assuming any unanswered field as unknown, and why not exclusively exporting only known answers like yes/no?

holger-stenzhorn commented 3 years ago

@makampf Well, technically, if you do not answer a question at all then you do not provide any information but if you answer that you do not know then you provide at least that information... 🙂 How relevant this is in the end and how this should be "modeled" in FHIR is another question. 🤷‍♂️

cerbelding commented 3 years ago

The CodeSystem DataAbsentReason describes the differences. (and also possible different codings)
As @holger-stenzhorn already said, the relevance is indeed questionable.

makampf commented 3 years ago

I wondered about the relevance in cohort identification an thats actually different level of questionable relevance :D