Investigate backend generation and data validity of D_597869478_D_597869478_num in Module 2

[gloria-trivitt]

Overview
The Question “In total, how many months or years have you used other progestin-only medication?”, CID 597869478, in Module 2 all versions, has a response category not indicated in the data dictionary or the Quest questionnaire mark-up (Nicole checked multiple versions in GitHub). We expect answers under D_597869478_D_970604592 which has the unit of years or D_597869478_D_434243220 which has the unit of months. However, in Module 2 version 1 we see an additional answer category D_597869478_D_597869478_num for which the units are unknown and in Module 2 version 2, D_597869478_D_970604592, the years category disappears altogether, but D_597869478_D_597869478_num is still present.

We would like DevOps assistance in determining how this D_597869478_D_597869478_num variable is being generated on the back end and testing to determine if D_597869478_D_597869478_num contains any valid year or month data, especially in Module 2 version 2 where D_597869478_D_597869478_num contains about 50% of all replies.

Concept definitions

Concept ID	Definition
597869478	In total, how many months or years have you used other progestin-only medication?
434243220	Months used
970604592	Years used

Query results

SELECT
    COUNT(*) AS count,
    (D_597869478_D_970604592 IS NOT NULL) AS has_years,
    (D_597869478_D_597869478_num IS NOT NULL) AS has_unknown,
    (D_597869478_D_434243220 IS NOT NULL) AS has_months
FROM `nih-nci-dceg-connect-prod-6d04.FlatConnect.module2_v1`
GROUP BY has_years, has_unknown, has_months
ORDER BY count DESC;

Row	count	has_years	has_unknown	has_months
1	2302	false	false	false
2	14	false	false	true
3	11	false	true	false
4	2	true	false	false

SELECT
    COUNT(*) AS count,
    (D_597869478_D_597869478_num IS NOT NULL) AS has_unknown,
    (D_597869478_D_434243220 IS NOT NULL) AS has_months
FROM `nih-nci-dceg-connect-prod-6d04.FlatConnect.module2_v2`
GROUP BY has_unknown, has_months
ORDER BY count DESC;

Row	count	has_unknown	has_months
1	65807	false	false
2	451	false	true
3	423	true	false

Supporting context
Current data indicates that when values are present, they appear in only one field (years, months, or unknown) with no overlap between variables.

[sonyekere]

Hi @gloria-trivitt does just checking, does Analytics have a milestone or timeframe in mind for DevOps to assist with the investigation?

[gloria-trivitt]

Hi @sonyekere, I do not currently have a defined timeline. I will defer to @FrogGirl1123 for guidance on prioritization and scheduling.

[gloria-trivitt]

Hi @sonyekere,

I spoke to Nicole about the timeline; would it be possible to schedule this for the May or June cycle?