Diabetes reported after baseline assessment - diabetes type when follow-up reported in 1B or 1C

[hcadavid]

Pairing rule issue

Description

• Is diabetes_type_adu_q_1_a an open-ended question? @baukearends and @squareb: which SNOMED diabetes type should we return if a value is given here?
  • diabetes_followup_adu_q_1 was collected in 1B, 1C, 2A, 3A, 3B, but t1d_followup_adu_q_1 and t2d_followup_adu_q_1 were collected only on 2A, 3A, 3B. So, which diabetes type should we use when 'diabetes_followup_adu_q1' is reported on 1B or 1C?

Specification

https://github.com/MyDigiTwinNL/Lifelines2Medmij-Mapping-tool/blob/8ca870f650b17373440c56020a5364556c72fe24/src/lifelines/Diabetes.ts#L163-L186

[baukearends]

First question: This depends on the answers participants can give. Probably the answers will include several more rare types of diabetes that cannot be classified as type 1 or 2 (MODY, LADA). @squareb Is there a way to see the possible answers to this question? And a more general question, is there a way to look these up ourselves?

Second question: One way to find out would be to look at the 'future' answers on t1d_followup_adu_q_1 and t2d_followup_adu_q_1. If this doesn't work, the most probable answer will be type 2 diabetes as type 1 diabetes develops generally, but not exclusively, at a young age. This is a topic that may warrant some more discussion.

[squareb]

@baukearends @hcadavid Question diabetes_type_adu_q_1_a is indeed an open-ended question. It contains a variety of (unique) answers filled in by the particpants. I'm not able to list them all, but examples may include: "pregnancy diabetes", "pre-diabetic" "hypoglycemia". You may need to recode this data in order to get the information you require. For baseline +- 475 participants have filled in an answer for this variable.

[hcadavid]

@baukearends @squareb I'll leave this issue open while we get access to the data, so we can analyze the responses to diabetes_type_adu_q_1_a

[hcadavid]

@baukearends I extracted the open-ended responses from diabetes_type_adu_q_1_a, sorted and saved them on our tmp01 space (hcadavid/analysis/diabetes_type_adu_q_1_a_open_ended_responses.txt file). Could you look at it and check how could we recode them, or to which kind of diabetes (id any) could be mapped?

[baukearends]

I took a first look at the data. A significant portion of the data is coded in values of 0 to 5. @squareb, do you know what these mean? They are not listed in the _1a_q_1enumerations.csv file.

[squareb]

The variables diabetes_type_adu_q_1_a is an open-ended question, so you're correct that it is ot listed in the enumerations file.

How do you extract the column from the data? If I cut the data and save it in a new file, I'm also getting values between 0 to 5, but if I look at the datafile without extracting the column I'm not getting these values. This may need some further investigation.

[hcadavid]

@squareb On issue #3, which seems to be related to this, I'm describing how I'm extracting the data.

https://github.com/MyDigiTwinNL/CDF2Medmij-Mapping-tool/issues/3#issuecomment-1739378204