search

opensafely / tpp-sql-notebook

2 stars 0 forks source link

*DISEASE* Other Heart Diseae #83

Open annaschultze opened 4 years ago

annaschultze commented 4 years ago

For the ACEi/ARB and ICS studies, we want to separate out HF and other ischaemic heart diseases from the chronic heart disease code list. Once the HF list is finalised (#80), this should be reasonably straight forward.

CarolineMorton commented 4 years ago

There is also a myocardial infarction list created by @hjforbes here #64 that could be added as well.

annaschultze commented 4 years ago

Thanks! I've created a list from the chronic heart disease list, minus the codes we agreed were codes for heart failure.

IHD_CTV3_DRAFT.xlsx

I then compared this to the MI list. There are a few (~10-15) codes in the MI list not in the CHD list, primarily relating to term 'Q wave'. Would be good to confirm whether these should be in either list.

MI_CTV3_COMPARE.xlsx

After this - we'll need to decide whether we want this to capture a specific concept (such as IHD), or whether we are content to just let it be 'all other chronic heart diseases, as identified for that code list, apart from heart failure'. My sense is that it'll be easier to try and capture a distinct concept, such as IHD (although it may be that these completely overlap).

annaschultze commented 4 years ago

Notes from call with LT: Decision to leave this variable as 'other' heart diseases as the primary purpose of it is confounding control. It'll therefore be a subset of the CCD list, with decisions pending about whether or not to incorporate the additional MI codes identified during the comparison of the CCD and MI lists.

annaschultze commented 4 years ago

Update from conversation with @HelenCEBM : The CCD list did not include Q waves because they don't necessarily indicate a past MI and so they therefore required a diagnosis code for a past MI. Therefore, to maintain consistency, we will not incorporate into this list either. Posting the subset for sign-off, if @laurietomlinson can provide clinical sign-off.

annaschultze commented 4 years ago

Final Sign-off

DEFINITION:

Example Output: patient_id date_of_ohd ohd_bin
234234 01.02.2019 1
23490 04.03.2008 1

CODE LISTS:

OHD_CTV3_FINAL.xlsx

This codelist was made by:

  1. Using a finalised code list for chronic cardiac disease (#29)
  2. Identifying codes relevant to heart failure (#80)
  3. Creating a subset of CCD that consists of all codes from [1], not in [2].

FLEXIBILITY NEEDED BETWEEN STUDIES: None anticipated.

EFFECTS ON COHORT SELECTION: None anticipated. Should be noted that this list is intended for confounding control, and thus the specific concepts may not be transferable to other studies.

POTENTIAL BIASES: None anticipated.

CLINICAL SIGN OFF & DATE: @CarolineMorton 07/05/2020 15:27

EPIDEMIOLOGY SIGN OFF & DATE: @annaschultze 05/05/2020 19:32

SHARED WITH WIDER TEAM:

FINAL SIGN OFF DATE: