The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » Mothers BMI
Mothers BMI [message #1556] Thu, 13 March 2014 09:47 Go to next message
user is currently offline  user
Messages: 11
Registered: March 2014
Member
Hi!

I am trying to measure the association between Childhood undernutrition and mother's BMI from Nepal DHS survey- 2011. I am using the dataset Household member Recode. I categorized variable ha40 in to three categories (for mothers BMI) and also i have categorized z-scores for Stunting, underweight and wasting....... When i tried to tabulate Stunting/Wasting/Underweight with mother's BMI, STATA says no observations...

I need help to solve the problem. I will be grateful to get some suggestions!!
Thank you.
Pawan
Re: Mothers BMI [message #1559 is a reply to message #1556] Thu, 13 March 2014 12:34 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 292
Registered: March 2013
Senior Member

Could you post your Stata code on the board?

Also, you may want to consider using the child recode. That will have one observation per child, whereas the hh member recode includes everyone in the household including people not eligible for anthropometry (I think).

Re: Mothers BMI [message #1560 is a reply to message #1559] Thu, 13 March 2014 14:20 Go to previous messageGo to next message
user is currently offline  user
Messages: 11
Registered: March 2014
Member
*I choose Household Member Recode dataset because, the findings are compatible with the DHS report (i think this is the recommended one for child nutrition): the prevalence was shown different when i tried child recode..
I tried the child recode and it was possible to perform the task on that.......

This is the code for the tabulation i used..
*
generate stunting
generate stunting=.
replace stunting =1 if hc70>=-600 & hc70<-200
replace stunting =0 if hc70>=-200 & hc70<600
tabulate stunting

*generate Mother's BMI
generate mother_nutrition=.
replace mother_nutrition=1 if ha40>=1660 & ha40<1850
replace mother_nutrition=2 if ha40>=1850 & ha40<2500
replace mother_nutrition=3 if ha40>=2500 & ha40<=3917
tabulate mother_nutrition
**
tabulate stunting mother_nutrition

Thank you.
Re: Mothers BMI [message #1561 is a reply to message #1560] Thu, 13 March 2014 14:40 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 292
Registered: March 2013
Senior Member

Hey there.

Just looking at the recode manual - is that a typo with hc70? Is it possible you want hc8? (hc7 is actually percentile, not z-score, so maybe it was a double-typo?)

I have always used the child recode for these. Also, in the HH member recode, you'd need to limit the age of your sample, I think, since some mothers (and maybe some fathers) would have anthro measurements too.

I'm guessing that, since your results worked with the child recode, you're happy using that? Or did you want to for sure get the HH recode working?

You should also probably weight these to get the DHS guide results, so you'd need the svyset command, and they svy: prefix before tabluate.

You good now?
Re: Mothers BMI [message #1564 is a reply to message #1561] Thu, 13 March 2014 16:10 Go to previous messageGo to next message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
User--the reason for the "no observations" message is that the people who have non-missing values for your maternal nutrition vbl don't overlap w/ the ppl that have non-missing values for your stunting vbl, you can confirm this via

tab stunting mother_nutrition,miss

Now, when I do child anthropometry, I've always used the HW variables instead of the HC variables. The difference, according to Tom Pullum (via Bridgette-DHS, link here: http://userforum.dhsprogram.com/index.php?t=rview&th=137 &goto=262#msg_262), is that HC is taken from ALL children in the household, whereas HW is taken from children whose mothers are in the household. Like Reduced-For(u)m, I use the Child Recode dataset when I do this.

For mother's BMI, I use V445. Note that this is in the women's recode file, so I guess you can miss the children of mothers who weren't the main respondent to the women's questionnaire in this way. When you use HW70 and V445, there is some overlap between mother's nutrition and child stunting:


. generate mother_nutrition=.
(5306 missing values generated)
.
. replace mother_nutrition=1 if v445 >=1660 & v445 <1850
(362 real changes made)
.
. replace mother_nutrition=2 if v445 >=1850 & v445 <2500
(1860 real changes made)
.
. replace mother_nutrition=3 if v445 >=2500 & v445 <=3917
(249 real changes made)
.
. tabulate mother_nutrition

mother_nutr
ition...Freq....Percent.Cum.

1.......362.....14.65...14.65
2.......1,860...75.27...89.92
3.......249.....10.08...100.00
Total...2,471...100.00

. generate stunting=.
(5306 missing values generated)
.
. replace stunting =1 if hw70>=-600 & hw70<-200
(982 real changes made)
.
. replace stunting =0 if hw70>=-200 & hw70<600
(1353 real changes made)
.
.tabulate stunting

stunting.Freq.......Percent......Cum.

0.......1,353.......57.94.......57.94
1.........982.......42.06......100.00
Total...2,335......100.00

.tabulate stunting mother_nutrition,miss

mother_nutrition
stunting.......1...........2.........3........M......Total

0..............157.........990.......159.........47..1,353
1..............172.........717.......64..........29....982
M...............33.........153.......26.......2,759..2,971
Total..........362.......1,860.......249......2,835..5,306

. tabulate stunting mother_nutrition

mother_nutrition
stunting.......1.........2.........3.........Total

0..............157.......990.......159.......1,306
1..............172.......717........64.........953
Total..........329.....1,707.......223.......2,259


hth,
RHS

[Updated on: Thu, 13 March 2014 16:22]

Report message to a moderator

Re: Mothers BMI [message #1570 is a reply to message #1564] Thu, 13 March 2014 19:44 Go to previous message
user is currently offline  user
Messages: 11
Registered: March 2014
Member
Thank you very much!! That was very helpful......
Previous Topic: Complementary foods - how to interpret dataset contents
Next Topic: Replicating table counts for children under 5 in Bangladesh DHS
Goto Forum:
  


Current Time: Thu Mar 28 21:52:24 Coordinated Universal Time 2024