The DHS Program User Forum: Mortality » how to look for the number of pregnant women and children died

Home » Topics » Mortality » how to look for the number of pregnant women and children died

Show: Today's Messages :: Show Polls :: Message Navigator

how to look for the number of pregnant women and children died [message #20533]

Mon, 16 November 2020 05:02

kkkkkk
Messages: 12
Registered: November 2020

Member

Hi! I am currently working on a project about maternal education and mortality. I am looking for the total number of women who are pregnant/gave birth in the past 5 years and the total number of child died under age of 5 in a survey. I tried to look it up in the final report and key findings but they don't seem to be there. How should I do it? Thank you in advance.

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21557 is a reply to message #20533]

Wed, 18 November 2020 15:11

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

The third table in the infant and child mortality chapter of the main report gives the under-five mortality rates by mother's education. These rates are given for the past 10 years, rather than the past 5 years, to get more cases.

You can also look at STATcompiler, where many of the indicators are broken down by background characteristics, including mother's education.

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21559 is a reply to message #21557]

Wed, 18 November 2020 15:46

kkkkkk
Messages: 12
Registered: November 2020

Member

Thank you! What I am looking for in specific is the actual figure recorded in the survey (5-year period) so I can work around with the dataset for a regression analysis. For this, is the dataset the only option I can look into? Which variable should I look at when I am trying to see how many mothers gave birth and how many children died in the preceding 5 years? I am trying to make a binary variable of children died under the age of 5 = 1 and 0 otherwise. Do you have any suggestion on computing that?

[Updated on: Wed, 18 November 2020 15:51]

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21561 is a reply to message #21559]

Wed, 18 November 2020 16:03

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Number of births in the past five years is v208 in the IR file. We do not have a variable for whether there was a child death in the past five years. If a child in the KR or BR file died, it has b5=0. However, the KR and BR files have children as units, not mothers or women.

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21565 is a reply to message #21561]

Thu, 19 November 2020 07:30

kkkkkk
Messages: 12
Registered: November 2020

Member

I see that b6 gives information about how long has the child lived. Can I use that to construct a binary variable, ie under 5 mortality = 1 and 0 otherwise? And how do I interpret the value? From the manual, '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.' What does it mean by it lies between 1 - 23 months?

[Updated on: Thu, 19 November 2020 07:33]

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21566 is a reply to message #21565]

Thu, 19 November 2020 08:03

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is another response from DHS Research & Data Analysis Director, Tom Pullum:

You can definitely construct such a binary variable based on b6 (or b7, which is a recode of b6 and is easier to use). However, it is important for the analysis of such a variable to note that some children will be alive and younger than the cutoff age at death. They have not had full risk of dying before that age. They are "censored". The DHS calculation of under-five mortality rates takes this into account. You cannot match DHS rates without taking this into account. See the Guide to DHS Statistics and the under-five mortality programs on out GitHub site.

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21570 is a reply to message #21566]

Thu, 19 November 2020 15:04

kkkkkk
Messages: 12
Registered: November 2020

Member

In fact I am not trying to calculate the rate but do you mind to explain a bit more on the 'censored' part?

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21571 is a reply to message #21570]

Thu, 19 November 2020 17:09

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is another response from DHS Research & Data Analysis Director, Tom Pullum:

This is a fairly technical subject. For DHS data it comes up when you want to use relatively recent births and deaths to estimate a child mortality rate. For example, you cannot estimate the probability of dying before the first birthday using only children who were born in the past year. Children who have not died are still at risk of dying before their first birthday. Below is a definition from the internet--you can find more. This type of censoring is also called "right" censoring, distinguished from "left" censoring.

"Censoring / Censored observation: If a subject does not have an event during the observation time, they are described as censored. The subject is censored in the sense that nothing is observed or known about that subject after the time of censoring. A censored subject may or may not have an event after the end of observation time."

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21583 is a reply to message #21571]

Fri, 20 November 2020 18:57

kkkkkk
Messages: 12
Registered: November 2020

Member

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21584 is a reply to message #21583]

Fri, 20 November 2020 18:59

kkkkkk
Messages: 12
Registered: November 2020

Member

Thank you! Now I am moving on to look at the number of interviewed women who have given birth in her life. I see that I can make use of bidx to do it but it is in the IR file. Is there a variable or any way I can compute that from the br file? Also, if I am using the br file for my regression, I may get repeated information regarding mother's educational level because she might give birth to more than 1 child. Is that a problem and how can I solve that?

Report message to a moderator

Re: how to look for the number of pregnant women and children died [message #21590 is a reply to message #21584]

Mon, 23 November 2020 10:48

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

The number of children ever born is v201. It is in both the IR and BR files. For each woman, that matches with the maximum value of bidx and the maximum value of bord. I suggest that you look at the labels for all the variables with variable names that start with v20 or v21.

In the BR file, all the data for the mother will be the same for all of her children's records. It is not a problem that this information is repeated.

Report message to a moderator

Previous Topic:	Time Spent at Place of Delivery m61
Next Topic:	Number of under-five deaths

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Dec 25 14:15:58 Coordinated Universal Time 2024