Home » Countries » India » NFHS 1998-1999 Diarrhoea Questions
| NFHS 1998-1999 Diarrhoea Questions [message #2737] |
Tue, 19 August 2014 09:37  |
lee7573
Messages: 3
Registered: August 2014
|
Member |
|
|
Hello there,
I am working with the NFHS - 2 (1998-1999) dataset for India.
For the National Family Health Survey for Bihar 1998-1999. I got a question regarding the appendices located here:
http://www.rchiips.org/nfhs/data/bh/bhappab.pdf
I was wondering how the estimates were created for:
1. Had diarrhoea in past 2 weeks (Children under age 3 years)
2. Treated with ORS (Children under age 3 with diarrhoea in past 2 weeks)
3. Taken to a health facility/provider for diarrhoea (Children under age 3 with diarrhoea in past 2 weeks)
Also is the variables used to create the indicator for the percentage with ARI taken to a health facility or provider on Table 6.17 in the India National Family Health Survey 1998-1999 (NFHS-2) the same as point 3 above.
Thanks
|
|
|
|
|
|
| Re: NFHS 1998-1999 Diarrhoea Questions [message #2751 is a reply to message #2737] |
Wed, 20 August 2014 15:34  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior Data Processing Specialist, Noureddine Abderrahim:
1- The inclusion/exclusion of a child for any type of analysis is defined by the period in which the child is born. To include a child we take his/her date of birth and see whether it is included in the period of analysis which in this case is the period 1-35 months preceding the survey. To compute the age in months or the number of months passed since the child's birth we look at his/her date of birth and the date of interview. The date of birth of the child is asked in question 215 of the India questionnaire and the date of interview is given in the cover page of the questionnaire. In the recode data files that DHS distributes, , these two dates are transformed into what is called the Century Month Code or CMC for ease of use. The CMC is the number of months since the beginning of the last century until the date of interview. The number of months would be the number of years since the year 1900, times the number of months in a year which is 12, plus the number of months passed in the current year. The CMC for date of birth for children is B3 and the CMC for the date of interview given by the variable V008. You may want to consult the DHS Recode Manual, that can be downloaded from the www.DHSprogram.com website. For the child health indicators, you need to select only the surviving children age 1-35 months from among the two most recent births.
2- For analysis of data on child health, use of the data where the child is the unit of analysis will be most efficient. The most appropriate data file for child and maternal health information is the one with KR in its name, namely: IAKR42xx (where xx is FL for ASCII Flat file including a data description for STATA, SPSS, STATA, SV for SPSS system file, SD for SAS system file, and DT for STATA system file).
3- Diarrhea: This indicator is taken from the question 469 of the India questionnaire and it corresponds to question 454 model A questionnaire or standard questionnaire model A of the phase 3 of DHS. This standard questionnaire is the base of questions used by the NFHS-2 and all the other surveys that were conducted during this phase. The question 469 is recoded for distribution into variable H11. The estimate is a frequency of the answer "Yes, last two weeks", code 2, in the variable H11 for all children born in the period of analysis.
4- Treated with ORS: Among the children who had diarrhea in the 2 weeks preceding the survey, the estimate of children who received an ORS treatment is the percentage who answered "YES" to question 475 first part. The recode variable that corresponds to this question is H13. The answer "yes" to 475 is recode "Yes: ORS - probed", code 2
5- Taken to a health facility: All the answers to question 474 from to A to S excluding pharmacy codes Q and R were grouped in one recode variable called H12Z. The individual answers are stored in the variables H12A to H12X. Among the children who had diarrhea and born within the period of analysis, the estimate of the children taken to a health facility is a frequency of the answer "Yes: medical treatment", code 1, to H12Z.
6- Taken to a health facility for children with fever/ARI: The estimate is based on question 466 (breathes faster than usual with short, rapid breath), or H31B in the recode data file, for children with fever in question 465, or H31 in the recode data file. The answers to "Taken to a health facility" uses the question 468, or H32Z. The individual questions for the facilities are stored in recode variable H32A to H32X.
|
|
|
|
| Re: NFHS 1998-1999 Diarrhoea Questions [message #2917 is a reply to message #2751] |
Tue, 09 September 2014 14:01 |
lee7573
Messages: 3
Registered: August 2014
|
Member |
|
|
Hi, thanks for answering my questions quickly. Your comments were especially helpful in analyzing the dataset.
I do have one more question regarding the NFHS 1998-1999 dataset. I am unable to get the prevalence rate of fever to match the NFHS-2 Report on Table 6.17 for Uttar Pradesh. I got an estimate of 27.7%, whereas the report says 27.8%. I was wondering if this could be due to the differences in the weighted denominator used to form the estimate. One interesting note is that the weighted denominator is not matching for me throughout the prevalence rate of ARI, diarrhea and fever for all youths aged months 0-35. I got a weighted denominator of 4050.05255 for ARI and diarrhea, whereas 4045.0725 for fever. I was wondering if you happen to know how to account for the 5 youths that were missing between the two estimates? This is something that my client is interested in.
FYI, this is how I set up my analysis:
The dataset I used is: UPKR42FL.DTA
The variables I used to set up my survey are v022 v001 to set up the strata and cluster. The weight I used is weight = v005/1000000 and defined the kids aged in month as age_month = v008 - b3.
To define fever:
if h22 in (0,8,9) then fever = 0; *no fever or don't know;
else if h22 in (1) then fever = 100; /*had fever within two weeks */
Thanks again for all your assistance!
|
|
|
|
Goto Forum:
Current Time: Wed Nov 27 14:51:19 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
