The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » child nutrition BDHS-2011
child nutrition BDHS-2011 Mon, 14 July 2014 02:05
 sohelruhrd Messages: 19Registered: July 2014 Location: Bangladesh Member
I used Bangladesh DHS-2011 "Children's Data - Children's Recode (KR)" data set to estimate prevalence of stunting, wasting and underweight among children.
I found the prevalence of stunting, wasting and underweight are as 40.6%, 15.4% and 35.8% respectively. But the BDHS 2011 report shows the prevalence as 41.3%, 15.6\$% and 36.4% respectively.

i used following STATA commands:

// child stunting calculation
codebook hw70
tab hw70 if hw70>9990,m
tab hw70 if hw70>9990,m nolabel
gen HAZ=hw70
replace HAZ=. if HAZ>=9996
histogram HAZ
gen stunted=.
replace stunted=0 if HAZ ~=.
replace stunted=1 if HAZ<-200
tab stunted

// child wasting calculation
codebook hw72
tab hw72 if hw72>9990,m
tab hw72 if hw72>9990,m nolabel
gen WAH=hw72
replace WAH=. if WAH>=9996
histogram WAH
gen wasted=.
replace wasted=0 if WAH ~=.
replace wasted=1 if WAH<-200
tab wasted

// child underweight calculation
codebook hw71
tab hw71 if hw71>9990,m
tab hw71 if hw71>9990,m nolabel
gen WAZ=hw71
replace WAZ=. if WAZ>=9996
histogram WAZ
gen underweight=.
replace underweight=0 if WAZ ~=.
replace underweight=1 if WAZ<-200
tab underweight

So, what's wrong with my estimate. Can anybody help me to find the problem.

regards,
Shafiur
Re: child nutrition BDHS-2011 [message #2566 is a reply to message #2565] Mon, 14 July 2014 02:29
 Reduced-For(u)m Messages: 292Registered: March 2013 Senior Member

You need to use the survey weights to get the DHS report numbers. See the FAQ under "using data files"

http://dhsprogram.com/faq.cfm

Here is some example code for setting up the weigthts, and then you would want to use "svy: tab stunted" to calculate stunting rates.

Example Stata code:
*generate weight
generate weight = v005/1000000
*make unique strata values by region/urban-rural (label option automatically labels the results)
egen strata = group(v024 v025), label
*check results tab strata
*tell Stata the weight (using pweights for robust standard errors), cluster (psu), and strata:
svyset [pweight=weight], psu(v021) strata(strata)

- See more at: http://dhsprogram.com/faq.cfm#sthash.QTqZEALu.dpuf
Re: child nutrition BDHS-2011 [message #2574 is a reply to message #2566] Wed, 16 July 2014 01:54
 sohelruhrd Messages: 19Registered: July 2014 Location: Bangladesh Member
Thank you very much for reply.

regards,
Shafiur
Re: child nutrition BDHS-2011 [message #2832 is a reply to message #2574] Fri, 29 August 2014 10:09
 DHS user Messages: 111Registered: February 2013 Senior Member
I use the Stata codes below in an attempt to reproduce the Nigeria DHS 2013 Child Anthropometry table using the PR6 files. However, my results are slightly different and total is 26,255. Please advise.

use "/Users/user/Documents/NDHS2013/NGPR6AFL.dta
*generate weight
generate weight = hv005/1000000

// child stunting calculation
codebook hc70
tab hc70 if hc70>9990,m
tab hc70 if hc70>9990,m nolabel
gen HAZ=hc70
replace HAZ=. if HAZ>=9996
histogram HAZ
gen stunted=.
replace stunted=0 if HAZ ~=.
replace stunted=1 if HAZ<-200
tab stunted

// child wasting calculation
codebook hc72
tab hc72 if hc72>9990,m
tab hc72 if hc72>9990,m nolabel
gen WAH=hc72
replace WAH=. if WAH>=9996
histogram WAH
gen wasted=.
replace wasted=0 if WAH ~=.
replace wasted=1 if WAH<-200
tab wasted

// child underweight calculation
codebook hc71
tab hc71 if hc71>9990,m
tab hc71 if hc71>9990,m nolabel
gen WAZ=hc71
replace WAZ=. if WAZ>=9996
histogram WAZ
gen underweight=.
replace underweight=0 if WAZ ~=.
replace underweight=1 if WAZ<-200
tab underweight

tab shstat wasted [iweight=weight], row

Re: child nutrition BDHS-2011 [message #2833 is a reply to message #2565] Fri, 29 August 2014 11:16
 Trevor-DHS Messages: 799Registered: January 2013 Senior Member

To match the results in the NDHS report, you need to select the same cases as used by DHS. These are:
a) De facto children (children that stayed in the household the previous night)
b) Children with complete and in-range measurements on both height and weight and the z-scores calculated from them.

The following code can be added to your Stata program to select the correct cases:

* Children who stayed in the household the previous night and who have height and weight z-scores on all 3 measures
gen eligch = 0
replace eligch = 1 if (hv103==1 & hc70 < 9996 & hc71 < 9996 & hc72 < 9996)

* Tabulate only for the eligible children
tab shstat wasted [iweight=weight] if eligch==1, row
Re: child nutrition BDHS-2011 [message #2834 is a reply to message #2565] Fri, 29 August 2014 11:23
 Trevor-DHS Messages: 799Registered: January 2013 Senior Member
Dear Sohelruhrd

You are trying to calculate the stunting and wasting from the data in the KR file, but this is only for children of women interviewed. To match the results in the Bangladesh recport, you need to use the information for all children in the household found in the PR file. Change the hw variables that you are using below to the equivalent hc variables (for the most part they have the same number). Additionally you need to restrict the children included to
a) De facto children (children that stayed in the household the previous night)
b) Children with complete and in-range measurements on both height and weight and the z-scores calculated from them.
See my other post in this thread for the code to do that.

[Updated on: Fri, 29 August 2014 11:24]

Report message to a moderator

Re: child nutrition BDHS-2011 [message #8424 is a reply to message #2566] Mon, 26 October 2015 16:44
 sohelruhrd Messages: 19Registered: July 2014 Location: Bangladesh Member
Hello,

Could you please help me to get the solution the the following question?

"What were the SD flag limits used to identify implausible values when computing the prevalence of stunting, wasting and underweight in BDHS 2011?

Thank you.

Regards,
Shfiur
Re: child nutrition BDHS-2011 [message #8425 is a reply to message #2565] Mon, 26 October 2015 18:30
 Trevor-DHS Messages: 799Registered: January 2013 Senior Member
You can find the SD flag limits in the following document: http://www.who.int/nutgrowthdb/software/Differences_NCHS_WHO .pdf.
Note that DHS excludes any case flagged for any of the three measures from all calculations.
 Previous Topic: Haz weighted mean Next Topic: More questions on exclusive breastfeeding syntax
Goto Forum:

Current Time: Wed Jun 12 15:38:48 Coordinated Universal Time 2024