The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » producing HAZ distributions
producing HAZ distributions [message #15196] Thu, 14 June 2018 22:58 Go to next message
tdusingize is currently offline  tdusingize
Messages: 4
Registered: September 2017
Location: New Zealand
Member
I would like to see the trends in child nutritional status using DHS datasets from different years, say, DHS 2000, DHS 2007, DHS 2014. Can anyone suggest me a statistical software that could help me to produce HAZ distributions of the 3 years in one single graph?

I have been using ENA for SMART but this can only give me distribution for one single year and I would like to have the distribution combined in one graph to a better visualization.

Thanks,
Theo
Re: producing HAZ distributions [message #15245 is a reply to message #15196] Tue, 19 June 2018 09:46 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
A response from Technical Specialist, Dr. Shireen Assaf:
Quote:

Dear Theo,

In Stata, you can show the histograms side by side using the hist command with the by option. To do this you would need to append the PR files from different years and create a year variable. The you can use the following command: hist haz, by(year)
Also, there is an alternative suggestion described in the links below. You would also need to append the PR files from different years before attempting this.
https://stats.idre.ucla.edu/stata/faq/how-can-i-overlay-two- histograms/
https://www.stata.com/new-in-stata/transparency-in-graphs/

Be sure to replace any values greater than 600 for the haz (hc70), waz (hc71), and whz (hc72) before plotting the histogram

Best,

Shireen Assaf
Technical Specialist
The DHS Program

Re: producing HAZ distributions [message #15254 is a reply to message #15245] Tue, 19 June 2018 18:58 Go to previous messageGo to next message
tdusingize is currently offline  tdusingize
Messages: 4
Registered: September 2017
Location: New Zealand
Member
Dear Shireen Assaf,

Thanks for your reply and for the explanation.I appreaciate.

However, I am not familiar with STATA and I was wondering if one could:
1) extract child sex, age, weight and height data from the PR files (i.e., copying these variables from, for example SPSS data set;
2) clean the data (by cleaning I mean take out the 9999, other, refused, and not present);
3) use WHO Anthro or ENA for SMART, to generate WAZ and HAZ;
4) copy HAZ data and use other software (for example R) to plot WAZ or HAZ?

I read somewhere on this forum that HAZ before 2005 were generated using NCHS reference, not WHO 2006 Standards. From my understanding, I think I need to re-calculate HAZ for 2000, 2005 using WHO 2006 standards for me to make a comparison with the HAZ of 2010 and 2015.

Another challenge is that I am not getting exactly the same number of children as in DHS reports after cleaning the data.

Kindly let me know your thought on these as well.

Thanks,
Theo


Re: producing HAZ distributions [message #15256 is a reply to message #15254] Wed, 20 June 2018 04:37 Go to previous messageGo to next message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear Dr.Shireen Assaf, Thank you very much!! I have learned a lot from your post and reply.
With Best Wishes,Hassen


Hassen Ali(Chief Public Health Professional Specialist)

[Updated on: Wed, 20 June 2018 04:40]

Report message to a moderator

Re: producing HAZ distributions [message #15317 is a reply to message #15254] Thu, 28 June 2018 09:10 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
A response from technical expert, Dr. Shireen Assaf,
Quote:

Dear Theo,
For any software you would need to use the PR files and use the variables hc70, hc71, hc72. For surveys before 2005 we have HW files that can be merged with the PR files to get the 2006 WHO reference values. The HW files contain the haz, waz, and whz according to the WHO reference. Then you can generate haz, waz, whz as follows (you can see the logic that you can adapt for your software of choice).
generate haz=hc70/100
replace haz=. if hc70>600
replace haz=. if hv103==0

Do the same as above for waz with hc71 and whz with hc72 . hv103 is the variable to indicate if the children are de facto or not. The DHS tables report anthropometric measures for de facto children (see notes on the bottom of the DHS tables).
The same software you use to clean and construct the variables should allow you to then plot the data.

We will only be able to answer questions on how to use the DHS data but cannot help with questions on which software to use or how to use it.

Hope these instructions will help.

Thank you.

Best,

Shireen Assaf

Re: producing HAZ distributions [message #15347 is a reply to message #15317] Tue, 03 July 2018 02:30 Go to previous message
tdusingize is currently offline  tdusingize
Messages: 4
Registered: September 2017
Location: New Zealand
Member
Hi Shireen,

Thank you so much. This helps a lot.

Cheers,
Theo
Previous Topic: Mother's Nutritional Status.
Next Topic: Conindex results not matching.
Goto Forum:
  


Current Time: Sat Dec 7 15:32:54 Coordinated Universal Time 2024