I have a database of diabetic patients. Mostly 80 % are T2DM patients. Now I would like to identify the T2DM patients based on their conditions reported. I mean when patients visit hospital (diabetic specialist clinic), they report/record their conditions.

Is it normal to consider items like Type 2 diabetes mellitus with Ulcer or Rubeosis iridis with T2DM or Retinopathy with T2DM etc as diagnosis for T2DM?

I thought we should only look for Type 2 diabetes mellitus in condition table because this to me seems like a diagnosis for T2DM. But can the above (or similar) items be considered for T2DM diagnosis and if yes, why? Because I see it is being considered

Can you help me understand this?

  • I think this is a pretty complex question that can only be answered knowing your research goals and with deep knowledge of the data source you have. Jan 28 '20 at 21:55
  • 1
    Ok, I'd be cautious in advising your further. I think this is really something you need to address with your research team, especially if these sorts of entries are confounded with your predictors. Analyzing a different data set, I ran into similar issues with dementia diagnoses and ultimately had to explain to the team that there was no way we could analyze dementia in that dataset because the codes were far too unreliable (the problem wasn't that people coded as having dementia lacked dementia, but far more lacked dementia codes that should have had them). Jan 28 '20 at 22:54
  • 1
    All your 3 highlighted diagnoses clearly word out T2DM, so it's obviously T2DM. But if the diagnosis says only "diabetes mellitus" without saying is it type 1 or type 2, there's no way you can say it just from the complications.
    – Jan
    Jan 29 '20 at 7:43
  • 1
    Please read something about diabetes 2. You can't really work with this without knowing the basics. There's an excess of insulin in T2DM already, unlike in T1DM. But you should really study this first.
    – Jan
    Jan 29 '20 at 15:09
  • 1
    The diagnosis is T2DM. With experience you'll realize that in practice not everything is so formally written as you expect. The diagnoses above are shortened descriptions of a T2DM + complication. So, a patient has a "T2DM" (this is the diagnosis) and a doctor believes that the (stomach) ulcer is associated with diabetes. As Bryan Krause mentioned above, there may be more informality in the written documents than you would expect.
    – Jan
    Jan 29 '20 at 15:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.