Module 3 Skip Logic or Saving Issue: ALCOHOL9

[KELSEYDOWLING7]

There's a very large number of participants are breaking the QC rule for ALCOHOL9: "Only display ALCOHOL9 if (ALCOHOL2 = 3) AND (IF RESPONSE PROVIDED TO REGALCOHOL_AGE) AND ((ALCOHOL5 = 1) OR (ALCOHOL5 = 0 AND RESPONSE PROVIDED TO ALCOHOL6))"

There are currently 2,708 participants that are breaking that rule in one or multiple ways, and are able to see and answer ALCOHOL9. Rather then list all the Connect IDs, here are the ways participants are breaking this rule or comments about their responses

• None of these participants had a response to Alcohol 6
• None of these participants had a response to REGALCOHOL_AGE
• All but one of these participants answered ALCOHOL5 = 1
• One participant had no response for Alcohol2, REGALCOHOL_AGE, ALCOHOL5, or ALCOHOL6 but still answered ALCOHOL9

The issue spans from 2022-2024, I'm thinking this is a skip logic issue.

[joshid-ims]

This is what is in the Word document:

The last part you mentioned (OR (ALCOHOL5 = 0 AND RESPONSE PROVIDED TO ALCOHOL6))") is not there.

[KELSEYDOWLING7]

Ok. The amount of errors doesn't change if I remove the (OR (ALCOHOL5 = 0 AND RESPONSE PROVIDED TO ALCOHOL6))") part. We still have 2,708 errors with just "(ALCOHOL2 = 3) AND (IF RESPONSE PROVIDED TO REGALCOHOL_AGE) AND ((ALCOHOL5 = 1) "

Though we should probably update the Module 3 QC rules delivery regardless once the issue is resolved

[joshid-ims]

Up to 1/14/2024, the display logic was: [DISPLAY ALCOHOL9 – ALCLIFE4H IF (ALCOHOL2=3 OR 2) AND IF RESPONSE PROVIDED TO THE FOLLOWING VARIABLES: (1) ALCOHOL4; ALCOHOL5; AND (2) IF SHOWN, ALCOHOL6. OTHERWISE GO TO SUN EXPOSURE SECTION].
The 01/21/2024 version has display logic: [DISPLAY ALCOHOL9 IF (ALCOHOL2= 3), (REGALCOHOL_AGE= #), AND ([ALCOHOL5= 1] OR [ALCOHOL5= 0 AND ALCOHOL6= #]) ELSE, GO TO INTROSE]
Sometime between 1/22/2024 – 6/25/2024, the display logic changed to: [DISPLAY ALCOHOL9 IF (ALCOHOL2= 3)), (REGALCOHOL_AGE= #), AND (ALCOHOL5= 1) ELSE, GO TO ALCLIFE1]

Currently, the code is as per [DISPLAY ALCOHOL9 IF (ALCOHOL2= 3)), (REGALCOHOL_AGE= #), AND (ALCOHOL5= 1) ELSE, GO TO ALCLIFE1]

Were all the responses in error collected after the last logic change was made?

[KELSEYDOWLING7]

2,560 were actually before that first change on 1/14/24. Then 48 are on the 14th of January but before the change on the 21st and 100 on the 21st of after. Only 1 error after 6/25/2024

[joshid-ims]

Then should we make any change?

[huip-ims]

@KELSEYDOWLING7 Hi Kelsey, those dates that Deepti provided are just snapshots of when we downloaded the Word document from Box. Do you have dates when those different versions of skip logic got released into production? The release dates will give us a more accurate picture of how many ALCOHOL9 responses actually violated the skip pattern.

[KELSEYDOWLING7]

@cunnaneaq Do you happen to have those release dates for me to double check the time frame of these errors?

@joshid-ims @huip-ims Thank you for letting me know about the logic changes and rough time frames. It seems like these changes have prevented most errors from occurring. The one remaining error is only missing a value for REGALCOHOL_AGE so it's possible that that one value didn't save. I'll keep monitoring this one and will circle back to it if the number of errors jump up again.

[KELSEYDOWLING7]

@cunnaneaq separate issue but do you mind telling me if there was a major skip logic change to the 9B variables around August 2022? Or direct me where on Box to check? There's a ton of errors for those variables but almost all errors stop around that time

[cunnaneaq]

Module 3 was pushed to production on 5/1/2023 and most recently on 2/3/2024 which includes the ALCOHOL9 changes.

According to the change log, it seems like the 9B variables (tobacco and cannabis) had changes made around October of 2022. Does that makes sense?

[KELSEYDOWLING7]

Yes it does! Thanks, I'll add that to my logic.

I do still see about 100 people with errors for ALCOHOL9 after 2/3/2024. There's a mention above about a change made in the logic between January and Jun of 2024. That date you have does lie within that range, but it's odd that there's a ton of errors after this February prod push, but then nothing after June.

[cunnaneaq]

Quest update 1.1.6 was pushed on 6/7/2024 and included bug fixes to help resolve issues with the ALCOHOL questions. This date aligns with the errors. Kelsey is updating her QC rule to catch errors after this date. Resolved.

[KELSEYDOWLING7]

Not sure if its worth testing this issue, but one person after that date answered ALCOHOL2=3 and ALCOHOL5=1 (as they should have) but skipped REGALCOHOL_AGE and still got to ALOCHOL9

[cunnaneaq]

I tested this in stage for Connect ID: 6943183164
My response choices are below:

ALCOHOL2 = 3
ALCOHOL3 = 21
REGALCOHOL_AGE = NR
ALCOHOL 5= 1
ALCOHOL8 = 0
BEER1 = 6
BEER2 = 5

Then, I was brought to the Sun Exposure section (as expected) and did not see ALCOHOL9

[KELSEYDOWLING7]

Ok thanks! Then this is definitely yet another instance of a saving issue (as we see in these issues: #440 #307 episphere/connect#1110)