The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Child Health » HBR ownership indicator derivation in Ethiopia DHS 2016
HBR ownership indicator derivation in Ethiopia DHS 2016 [message #12990] Thu, 31 August 2017 12:06 Go to next message
dbrown is currently offline  dbrown
Messages: 3
Registered: July 2015
In DHS where there is no facility trace-back exercise, the h1 variable tends to take on the values:

0 no card
1 yes, card seen
2 yes, card not seen
3 no longer has card

based on a combination of responses to the following two questions:

• Do you have a card where (NAME'S) vaccinations are written down? If YES, May I see it please?
with response set: YES, SEEN / YES, NOT SEEN / NO CARD
• Did you ever have a vaccination card for (NAME)?
with response set: YES / NO

All fine. Using h1 values as shown above and given the way the questions are asked, users have been able to derive an indicator for ever HBR ownership and for current HBR ownership from the variable h1 as follows:

ever HBR ownership = 1 (yes) if h1 == 1 | h1 == 2 | h1 ==3; = 0 (no) if h1 == 0
current HBR ownership = 1 (yes) if h1 == 1; = 0 (no) if h1 == 0 | h1 == 2 | h1 == 3

In this instance, we only give credit for current ownership if the HBR is seen.

Now to the issue. In the 2016 Ethiopia DHS PUF, variable h1 maintains information from the questionnaire items 504A and 505A (which generally conform to the above except that a look for other documented evidence in the HH was completed -- very good!) as well as information on whether documented evidence was observed in a health facility as part of the trace back that was conducted. The values for h1 in the 2016 Ethiopia DHS take on the values:

0 no card
1 yes, card seen
2 yes, card not seen
3 no longer has card
4 yes, card seen from health facility

Again, this is all fine and does not present any problems for the derivation of current HBR ownership indicator as noted above.

However, because the response values to question 505A in the Ethiopia questionnaire are not provided in the PUF and given the addition of the value = 4 response in h1, it is no longer possible to derive the ever HBR ownership indicator in the same manner as before. Why?

The problem is that PUF users no longer have the information on where the children that did not have a HBR seen in the HH would fall out in categories of "no card ever received" vs "card was received but I just cannot show it to you now" (e.g., it's locked up) or "card was received but I no longer have the card". In order for users to be able to derive the ever HBR ownership indicator, they must be able to ascribe those for whom facility-based documentation was identified to one of two groups: group 1: never received a card or group 2: received a card but either can't show you or no longer have it. The latter group of course would go into the numerator of the ever HBR ownership indicator.

So, the question is whether the variable h1 can be restructured to be consistent with past releases or also include the response values to question 505A in the PUF.

Thanks for having a look.

Re: HBR ownership indicator derivation in Ethiopia DHS 2016 [message #12998 is a reply to message #12990] Tue, 05 September 2017 19:09 Go to previous message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 800
Registered: January 2013
Senior Member
Hi David, We are looking into this and will produce a version of the dataset that preserves this information. I'm not sure of the timeframe right now, but will let you know when I know more. Cheers. Trevor
Previous Topic: Child's age in months
Next Topic: Ethiopia treatment sources: more detail on country-specific labels?
Goto Forum:

Current Time: Sat Jul 20 02:21:33 Coordinated Universal Time 2024