Compilation of issues

[beau-lukecolton]

SCID QC adjustments

• Please see email to Caitlin

SCID QC errors

• chrscid_a27 : Fulfills main criteria but further value was wrong, a27, a28, a48_1
  --> This QC flag is incorrect. You might be thinking that >both< A1 and A2, or A27 and A27, require "3" - this is incorrect. From the SCID instructions: "at least one of the symptoms was either (1) depressed mood [A1/A27] or (2) loss of interest or pleasure [A2/A28]."
  Example: CG45823 | Baseline; ME00972 | baseline

[to be continued]

Update Nov/20: removal of forms from the 'within 21-days' rule, from Sophie's email Oct/30
RA Prediction
CBC with Differential
GCP_CHS (Med review)
GCP_CBC (Med review)
Digital – Actigraphy
Digital – EMA
All self-reports

Update Nov/20: include all site QC in site QC spreadsheets - this includes all Domain QC. However, as before, keep this true for Melbourne but not individual Melbourne-RA spreadsheets

[npenzel]

Hi Beau-Luke, the issue with this subject is I think that the rater did not continue filling out the form? There is the item a48_1 that is NaN. This item asks whether a27 or a28 are a 3 and in this case it should be 1 because a27 is 3.

What I do not understand, this is a calculated field, and in our dictionary I see the calculation being like this:
if([chrscid_a43]>0, if([chrscid_a27]=3,1,0)+if([chrscid_a28]=3,1,0),"NaN ")

if I now input the values from this subject:
chrscid_a43 = 1,
chrscid_a27 = 3
chrscid_a28 = 3

the final value should be "1". This is what Owen's program (in my opinion correctly) catches here.
Is there potentially a coding difference in PRESCIENT for this item? Can you/RPMS figure this out?

Please let me know if I am missing something.

[cridgew13]

Hi everyone, just reviewing this now.

My interpretation is that the calculation field is correctly formatted. For CG45823, chrscid_a48_1=1, which is correct since chrscid_a43=1, chrscid_a27=3, chrscid_a28=1. This is correct and reflects Beau-Luke's point that the instructions say either chrscid_a27 or chrscid_a28 must equal 3.

However, for chrscid_a48_1 doesn't look right for ME00972. In this case, chrscid_a48_1=NaN when chrscid_a43=1, chrscid_a27=3, and chrscid_a28=2. chrscid_a48_1 should be 1. And it was 1 at the original data upload (12/12/2024) but got overwritten during a more recent upload (10/17/2025). To my understanding, NaNs should only appear in this field if chrscid_a43 is blank. It is not blank for either case, so this field shouldn't be NaN. I think this is what Owen's QC is picking up.

What I've learned is that in the combined csv, chrscid_a27, chrscid_a28, and chrscid_a43 are all -3 (which explains why chrscid_a48_1 is NaN). Beau-Luke, are you aware of any reason why these would be changed to -3? Dheshan says these are from the RPMS csvs directly so does RPMS have -3s for these fields or values?

[beau-lukecolton]

Hello! Would you still like me to review "chrscid_a27 : Fulfills main criteria but further value was wrong, a27, a28, a48_1"? I see that all of our QC items with this flag have been resolved, along with other SCID items that I believe were errors

[cridgew13]

Nevermind! Thanks Beau-Luke.

[beau-lukecolton]

Hi again, all of these items have returned to our QC spreadsheets. What is happening please?

[owenborders]

I am not sure why they disappeared before, I did not intentionally disable them. I did now though. Do we want this flag still included? Let me know and I will re-enable it.

[beau-lukecolton]

From my perspective, no - the flag seems to be an error. Nora/Caitlin to respond

[cridgew13]

Sorry for losing track of this, @beau-lukecolton. After reviewing some other cases (for example, BM44608 baseline), it looks like the chrscid_a48_1 calculation logic is functioning properly but these items were still being incorrectly flagged in the QC tracker as errors. The QC flags are, however, now resolved again as of 12/2/25. @owenborders, can you confirm that these were resolved because you disabled the flag, or are these going to pop back up again like they did 3 weeks ago?

[owenborders]

Yes they are disabled now!