+ Reply to Thread
Results 1 to 6 of 6

Thread: What method to choice?

  1. #1
    Points: 3,175, Level: 34
    Level completed: 84%, Points required for next Level: 25

    Posts
    22
    Thanks
    0
    Thanked 0 Times in 0 Posts

    What method to choice?




    Dear TalkStat friends,

    I have merged two databases (a=22,000 cases;b=12,000 cases). The two databases are the following a= diabetes screening project; b=heart failure project). In the two databases, 621 individuals are present in both projects (in other words i know at what year they had a diabetes diagnosis and at what year they had heart failure diagnosis).

    My question is, how would you suggest to do the following?

    I want to see if there are any sign difference between (years since diabetes diagnosis and heart failure. Some of the subjects have the heart failure (meassured by ejection fraction (%) before the diabetes diagnosis.

    First i was thinking to do a kruskall wallis (data not norma-d) and that each group represent eg 1 year from diabetes diagnosis to heart failure, 2 year from, 3 year from and so on. (range 0-10 years) but some subjects will have the diagnosis of heart failure before diabetes and will be minus(-) eg 1 year, 2 year and so on.

    I was thinking just making two groups (years from diabetes to heart failure) and one (years from heart failure to diabetes).

    Any suggestions?

    Best regards
    Gunnar

  2. #2
    Omega Contributor
    Points: 38,374, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: What method to choice?

    Your question and description are not quite clear.


    Is b a subset of a?
    Stop cowardice, ban guns!

  3. #3
    Points: 3,175, Level: 34
    Level completed: 84%, Points required for next Level: 25

    Posts
    22
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: What method to choice?

    hmm don't know how to explain. let's say that two professors are doing research in London. One professor invite subjects to his university, all ages, men and women during 2010 and 2015 (a) and another professor in london invites also same kind of subjects to his research (b). Since some subejcts will go to (a) only, (b) only and some both the 2 professors meet and match to see if the same individual has visit both of them. Let's say they find out that 100 subjects have attend both research teams (a) + (b). Now they want to study if there are any significant difference between (years since diabetes diagnosis and heart failure. Some of the subjects have the heart failure (meassured by ejection fraction (%) before the diabetes diagnosis.

  4. #4
    Points: 16, Level: 1
    Level completed: 31%, Points required for next Level: 34

    Posts
    7
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: What method to choice?

    As I understand:
    (a) contains a set of patients and dates of diabetes events
    (b) contains a set of patients and dates of heart failures

    Some patients may have had both diabetes and heart failures.

    Correct?

    And the goal now is to find a correlation between date of diabetes and date of heart failure, right?

    So the expected trend may be that diabetes leads to heart failure within x time, and you want to find some proof of that?

  5. #5
    Points: 16, Level: 1
    Level completed: 31%, Points required for next Level: 34

    Posts
    7
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: What method to choice?

    Quote Originally Posted by peteritv View Post
    As I understand:
    (a) contains a set of patients and dates of diabetes events
    (b) contains a set of patients and dates of heart failures

    Some patients may have had both diabetes and heart failures.

    Correct?

    And the goal now is to find a correlation between date of diabetes and date of heart failure, right?

    So the expected trend may be that diabetes leads to heart failure within x time, and you want to find some proof of that?
    If so, I have a few questions:
    a) If the intersection of (diabetes) and (heart problems) is very small compared to the total number of patients, then what relevance does your research have?
    b) From the intersection of (diabetes) and (heart problems), WHY would you want to make a distinction between (a) --> (b) and (b) --> (a)???

    If just 1% of all (diabetes) also have (heart problems), what is the relevance in that?
    If just 1% of all (heart problems) also have (diabetes), what is the relevance in that?

    And furthermore:
    I expect that patient x may have more than 1 entry in both (diabetes) and (heart problems)?
    How do you expect to cope with that?

    Can you eleborate a little further?
    Last edited by peteritv; 10-25-2016 at 05:40 PM.

  6. #6
    Points: 3,175, Level: 34
    Level completed: 84%, Points required for next Level: 25

    Posts
    22
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: What method to choice?


    Thanks,

    I want to make a distinction between a--b and b--a since it might be that diabetes itself, will accelerate the development of heart problems (or protect) and in those patients with a later diagnois of diabetes (first heart problems then diabetes) that it might be a difference in those diagnosed first with diabetes then heart problems and vice versa.

    hope its clearer

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats