pharmaverse / sdtmchecks

Analysis facing checks for SDTM data
https://pharmaverse.github.io/sdtmchecks/
Apache License 2.0
20 stars 6 forks source link

check_dd_ae_aeout_aedthdtc - descriptive text to be updated in pkg dataset #310

Closed sarabodach closed 1 month ago

sarabodach commented 5 months ago

For check check_dd_ae_aeout_aedthdtc.R:

Report has: Check for AEs with outcome marked ‘FATAL’ but no death date in AE • This check looks for AEs with death date (AEDTHDTC) not populated when AE outcome (AEOUT) marked “FATAL” image

But in test run, record flagged has AEDTHDTC = [date] and AEOUT = 'NOT RECOVERED/NOT RESOLVED'

image

sdtmchecksmeta to be updated

sarabodach commented 5 months ago

Update return message of check in case only 1 record: image

sarabodach commented 5 months ago

There is overlap with check_ae_aeout.R image

sarabodach commented 5 months ago

Hi @harriscw -

I had started updating the documentation for check_dd_ae_aeout_aedthdtc.R to specify that in addition to AEOUT = FATAL and missing AEDTHDTC, non-missing AEDTHDTC but AEOUT = (not fatal) would be flagged.

But check_dd_ae_aeout_aedthdtc.R seems to be doing the same thing as check_ae_aeout.R.

Do you agree with the following:

1) update metadata to reflect what check_dd_ae_aeout_aedthdtc is doing, including a note in the header that 'dd' was used in earlier naming conventions because it was related to death - but this domain is not truly used 2) attach r lifecycle::badge("superseded") to check_dd_ae_aeout_aedthdtc since it's only using AE for the check and to avoid breaking changes

We otherwise are not using lifecycle badges, so I'm not sure if that's strange to implement so selectively. But, we can always start, particularly for some utility functions that can be tagged with an "experimental" badge?

Thoughts?

Thanks!