The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset use in SPSS » Antenatal care vs infant mortality -Kenya
Antenatal care vs infant mortality -Kenya [message #1572] Fri, 14 March 2014 07:08 Go to next message
Malachi Arunda is currently offline  Malachi Arunda
Messages: 30
Registered: February 2014
Member
Dear Expert
1. How do I select only LAST BORN infants data from the CHILDREN recode data (born 12-59 months before the survey-spss)?
2. The AGE AT DEATH data is missing and I'm looking at death from (0-11months)or live beyond 1 year, how do I deal with this.I need to select ONLY those infants who die before their 1st birthda?
3. If I select Last born infants and do cross tabs for Tetanus toxoid and Iron Folic, will it match with ONLY those Last born infants selected?
Thank you
Re: Antenatal care vs infant mortality -Kenya [message #1575 is a reply to message #1572] Fri, 14 March 2014 09:36 Go to previous messageGo to next message
Sarah-DHS is currently offline  Sarah-DHS
Messages: 54
Registered: February 2013
Senior Member
Dear User,

We are working on getting an answer to your question. Which Kenya survey are you using?

Thanks,
Sarah at DHS

[Updated on: Fri, 14 March 2014 09:39]

Report message to a moderator

Re: Antenatal care vs infant mortality -Kenya [message #1578 is a reply to message #1575] Fri, 14 March 2014 13:08 Go to previous messageGo to next message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
Malachi,

For Question 1:
Restrict to BIDX == 1. You can confirm that BIDX == 1 is the most recent birth by tabulating the birth order and birth index variables.

. tab bord bidx if v201==5

.....birth.|
.....order.|.............birth.column.number
....number.|.........1..........2..........3..........4.|... ..Total
-----------+--------------------------------------------+--- -------
.........2.|.........0..........0..........0..........3.|... ......3.
.........3.|.........0..........0.........45..........0.|... .....45.
.........4.|.........0........184..........0..........0.|... ....184.
.........5.|.......358..........0..........0..........0.|... ....358.
-----------+--------------------------------------------+--- -------
.....Total.|.......358........184.........45..........3.|... ....590


Here you see that among women with 5 kids, the kids that are birth order=5 (youngest kid) has birth column number (bidx) equal to 1 (the Recode Manual also states that BIDX is reversely coded from BORD).


For Question 2:
From the DHS5 Recode Manual:
Quote:

Age at death of the child as reported in the questionnaire. The first digit of the age at death
gives the units in which it was reported: 1 - Days, 2 - Months, 3 - Years, 9 - Special
responses. The last two digits give the age at death in those units. Age at death is usually
reported in days if it was less than one month, in months if it was less than two years and
otherwise in years. If the last two digits contain a value greater than 90 then this is a special
response. For example, 298 means the age at death was a number of months, and the exact
number was unknown, but lies between 1 and 23 months.
BASE: Dead children (B5 = 0).

Link here: http://www.dhsprogram.com/pubs/pdf/DHSG4/Recode5DHS_23August 2012.pdf

Based on this, for children who died before their first birthday, you will most likely want to restrict to those with b6<300. Note, in the 08/09 child recode dataset, it looks like 167 children with bidx==1 are dead.

For Question 3:
I think these questions are only asked of the most recent birth, so if you restrict to bidx==1, you should be all set. You can confirm with by checking the birth column number of those for whom m1 (tetanus injection before birth). Note, the ANC vbls are typically only asked of the most recent birth, so for older children, you can expect them to be missing. For example:

. tab bidx if m1==.

......birth.|
.....column.|
.....number.|......Freq......Percent........Cum.
------------+-----------------------------------
..........2.|......1,673.......83.78.......83.78
..........3.|........302.......15.12.......98.90
..........4.|.........18........0.90.......99.80
..........5.|..........3........0.15.......99.95
..........6.|..........1........0.05......100.00
------------+-----------------------------------
......Total.|......1,997......100.00


As you can see, the missing values are all birth index >1.


hth,
RHS

[Updated on: Fri, 14 March 2014 13:09]

Report message to a moderator

Re: Antenatal care vs infant mortality -Kenya [message #1581 is a reply to message #1578] Fri, 14 March 2014 15:44 Go to previous messageGo to next message
Sarah-DHS is currently offline  Sarah-DHS
Messages: 54
Registered: February 2013
Senior Member
Dear User,

Please let us know if the previous post did not answer your question.

Thank you,
Sarah at DHS
Re: Antenatal care vs infant mortality -Kenya [message #1589 is a reply to message #1575] Fri, 14 March 2014 18:57 Go to previous messageGo to next message
Malachi Arunda is currently offline  Malachi Arunda
Messages: 30
Registered: February 2014
Member
Many thanks, I will be waiting. Using SPSS 2008-09 Children recode.

Small Addition: If Antenatal care data M12 was collected for only LAST BORN CHILDREN 12-59 months. Now when I'm cross tabulating with Children alive-yes/No data..this gives me all the children yet I only need want to compare those whose mothers attended antenatal care, how do I do this. Thank you.
Re: Antenatal care vs infant mortality -Kenya [message #1590 is a reply to message #1589] Fri, 14 March 2014 19:46 Go to previous messageGo to next message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
If you restrict to BIDX==1 (i.e., delete observation if BIDX>1), then your dataset includes only the most recent born children. From then on, any of your tabulations re: mother's antenatal care, for example, the number of ANC visits, whether she received injections, iron, etc., will be tabulated only on those children.

Using 2008-9 child recode data, I see that 1,997 children in that dataset were NOT the most recent, so when I drop those observations, I am down to 4,082 most recent births. Again, you can confirm that the ANC variables are collected only for the most recent birth, by asking the software to tabulate BIDX where ANC variable (ex: m13) is missing.

Note that the age range for BIDX==1 is 0-59 months, not 12-59 months. You will need to do some data management to create categories of child's age if you are interested only in the 12-59 subset.

Similarly, you will need to create a variable on whether the mother attended ANC or not in order to get the cross-tabulation between ANCxdeath. Note, because the number of children who died are quite small (165), you may have difficulties finding statistical significance.

My examples are in Stata but they are directly translatable to SPSS.


hth,
RHS
Re: Antenatal care vs infant mortality -Kenya [message #1591 is a reply to message #1589] Fri, 14 March 2014 19:51 Go to previous messageGo to next message
Malachi Arunda is currently offline  Malachi Arunda
Messages: 30
Registered: February 2014
Member
Thank you Hth and Sarah, I have read the details of your response and its proving to be very helpful. Please let me keep working, I will let you know if there is a problem.
Re: Antenatal care vs infant mortality -Kenya [message #1592 is a reply to message #1591] Sat, 15 March 2014 08:15 Go to previous messageGo to next message
Malachi Arunda is currently offline  Malachi Arunda
Messages: 30
Registered: February 2014
Member
In Spss dataset 2008-09, there is MIDX,I am assuming you mean the same thing with BIDX. The guidance you've given me has made me smile. thank you
Re: Antenatal care vs infant mortality -Kenya [message #1593 is a reply to message #1592] Sat, 15 March 2014 11:17 Go to previous messageGo to next message
user-rhs is currently offline  user-rhs
Messages: 132
Registered: December 2013
Senior Member
Dear Malachi,
Always happy to help. Based on the FRQ file that comes in the KEKR52SV.ZIP, there is a variable called BIDX in the SPSS version of the dataset. BIDX should equal MIDX, but it's always good to double check (example: by crosstabulating them).

When in doubt about variable definitions, consult the DHS Recode manual for the survey round in question. Kenya 08/09 falls under DHS5. Link: https://www.dhsprogram.com/publications/publication-dhsg4-dh s-questionnaires-and-manuals.cfm.


RHS

[Updated on: Sat, 15 March 2014 11:29]

Report message to a moderator

Re: Antenatal care vs infant mortality -Kenya [message #1599 is a reply to message #1593] Sun, 16 March 2014 17:00 Go to previous message
Malachi Arunda is currently offline  Malachi Arunda
Messages: 30
Registered: February 2014
Member
Thank you, I got them both now
Previous Topic: How to create the variable "family arrangement"
Next Topic: analysis of data of children
Goto Forum:
  


Current Time: Fri Mar 29 12:00:58 Coordinated Universal Time 2024