Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • composite measure not adding up?

    Hi,

    Relativley new to stata- I am using a data set with 619 participants, I created a composite measure called CDS( chronic disease syndrome) that should be 1= if ANY one of the following ( diabetes, obesity, or hypertension) and 0 if all diabetes, obesity AND hypertension =0. In order to do this, I made obesity, diabetes and hypertension binary. I expected my composite measure CDS = 1 to equal to the sum of yes/1 in each individual in the composite

    My issue is that when I tabulate the composite measure and count the composite number of yes, I get a number that is NOT equal to the tabulation of each individual yes

    ex: cdm, chronic diabetes, YEs -27
    chtn, chronic hypertension, yes=24
    obese_bi_1= binary obesity variable, yes= 112

    expected CDS_1 if any CDM, CHTN, obese=1 to be 112+24+27 , but instead it 136,

    I tried to show some of my output, below, generating the CDS variable, and then some simple tab, cross tabs so you can see the total of th individual variables

    .
    . **Generate a composit measure for exposure for preg 1 (CDS)--- where Chronic disease syndrome (com
    > posite measure) === Chronic htn, chronic dm, obesity
    . generate CDS_1=1 if chtn_bi_1==1| CDM_1==1|obese_bi_1==1
    (483 missing values generated)

    . replace CDS_1=0 if chtn_bi_1==0 & CDM_1==0 & obese_bi_1==0
    (483 real changes made)

    . label define CDS_1_name 0"No CDS" 1"Any CDS"

    . label values CDS_1 CDS_1_name

    . tab CDS_1

    CDS_1 | Freq. Percent Cum.
    ------------+-----------------------------------
    No CDS | 483 78.03 78.03
    Any CDS | 136 21.97 100.00
    ------------+-----------------------------------
    Total | 619 100.00


    | HDP_2
    chtn_bi_1 | No HDP HDP | Total
    ---------------+----------------------+----------
    No Chronic HTN | 536 59 | 595
    | 97.45 85.51 | 96.12
    ---------------+----------------------+----------
    Chronic HTN | 14 10 | 24
    | 2.55 14.49 | 3.88
    ---------------+----------------------+----------
    Total | 550 69 | 619
    | 100.00 100.00 | 100.00


    tab CDM_1

    CDM_1 | Freq. Percent Cum.
    ------------+-----------------------------------
    No CDM | 592 95.64 95.64
    CDM | 27 4.36 100.00
    ------------+-----------------------------------
    Total | 619 100.00



    . tab obese_1 //categorical obese variable

    obese_1 | Freq. Percent Cum.
    ------------------------+-----------------------------------
    Not obese | 316 54.02 54.02
    overweight | 157 26.84 80.85
    obese or morbidly obese | 112 19.15 100.00
    ------------------------+-----------------------------------
    Total | 585 100.00

    .
    .
    .
    . gen obese_bi_1 = 1 if obese_1 == 2
    (507 missing values generated)

    . replace obese_bi_1 = 0 if obese_1==.| obese_1==1 | obese_1 ==0
    (507 real changes made)

    . label define obese_bi_1_names 0"not obese" 1"obese"

    . label values obese_bi_1 obese_bi_1_names //binary obese variable



    | obese_bi_1
    sex_1 | not obese obese | Total
    -----------+----------------------+----------
    F baby | 258 57 | 315
    | 50.89 50.89 | 50.89
    -----------+----------------------+----------
    M baby | 249 55 | 304
    | 49.11 49.11 | 49.11
    -----------+----------------------+----------
    Total | 507 112 | 619
    | 100.00 100.00 | 100.00


    .

  • #2
    Hi Michelle,
    I dont think there is anything wrong with how you estimated your composite index, but rather on what you think it should measure.
    The bottom line is that if a person has more than 1 out of the three conditions you are using, he/she will will not be double counted in your composed measure.
    a simple tab between your individual measures will show you that a few people have more than one conditions at a time.
    HTH
    Fernando

    Comment


    • #3
      you are absolutely right -- I proved it by making a more granualr measure and now i can account for everything


      CDS_1_cats | Freq. Percent Cum.
      ---------------------------------+-----------------------------------
      No CDM,chtn,obesity | 483 78.03 78.03
      chronic HTN alone | 9 1.45 79.48
      chronic DM alone | 12 1.94 81.42
      obesity alone | 91 14.70 96.12
      chtn & chronic DM | 3 0.48 96.61
      obesity & chtn | 9 1.45 98.06
      obesity and chronic DM | 9 1.45 99.52
      chronic DM, chronic HTN,&Obesity | 3 0.48 100.00
      ---------------------------------+-----------------------------------
      Total | 619 100.00
      Last edited by Michelle Ogunwole; 19 May 2019, 07:23.

      Comment

      Working...
      X