| Pregnancy History [message #9211] |
Thu, 25 February 2016 01:07  |
|
|
I did not find variable indicating total number of pregnancies in pregnancy history of ever married women. I searched out from V224 to V243 but could not get the information about total number of terminated pregnancies. Its available only in binary nature (V228, V234).
Can i get total pregnancies by adding all children ever born to a women + total number of pregnancy losses + whether women currently pregnant or not?
I need:
1. Total number of pregnancies
2. Total number of lost pregnancies.
3. Whether the pregnancy resulting in (i) live birth (ii) miscarriage (iii) abortion (iv) spontaneous. Nominal variable
Which file should i adopt for above calculations!
1) BR-Birth Records
2) KR-Kid Records
3) IR-individual women records
Kindly help me for above problems.
Waqas Imran
|
|
|
|
| Re: Pregnancy History [message #9232 is a reply to message #9211] |
Thu, 25 February 2016 17:37  |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear user,
Your query has been forwarded to one of our experts. For your information, these are calendar variables. Take a look at your VCAL variables in the dictionary VCAL_VS3 has the value labels.
Thank you!
[Updated on: Thu, 25 February 2016 17:37] Report message to a moderator |
|
|
|
|
|
| Re: Pregnancy History [message #9245 is a reply to message #9242] |
Fri, 26 February 2016 09:33 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
Looking at the VCAL variable value labels in CSPro this is what I see:
Calendar (Colum VCAL_VS3
Birth B
Pregnancy P
Abortion A
Miscarriage C
Still Birth S
Not using 0
Female Sterilization 1
Male Sterilization 2
IUD 3
Injections 4
Implants/Norplant 5
Pill 6
Condom 7
Standard day methods (SDM) 8
Lactational amenorrhea (LAM) 9
Rhythm - Periodic abstinence L
Withdrawal M
Other modern method X
Other traditional Y
Missing ?
|
|
|
|
| Re: Pregnancy History [message #9248 is a reply to message #9211] |
Fri, 26 February 2016 16:42 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
Here is a response from senior data processing expert, Mr. Guillermo Rojas:
Quote:
In DHS we have two ways of collecting respondents children: A birth history or a pregnancy history. In surveys where we collect birth histories, it isn't possible to deduct the number of pregnancies because we only have terminations (stillbirths, abortions, miscarriages) for pregnancies in the last five years. In surveys with pregnancy histories, to determine the number of pregnancies besides the current pregnancy status, it is important to take into consideration those pregnancies that ended in multiple outcomes (twins, triples, and so on).
|
|
|
|
| Re: Pregnancy History [message #9265 is a reply to message #9248] |
Mon, 29 February 2016 12:34 |
|
|
A very helpful information shared.
I have another query that was the last part as i mentioned in my first thread on this issue, i.e.,
Whether the pregnancy resulting in (i) live birth (ii) miscarriage (iii) abortion (iv) spontaneous. Nominal variable
Here i wish to have variable with multinomial structure indicating the above four categories explicitly. In PDHS (2012-13) we have a table 5.15, chapter fertility and page number 80. Here the variable is presented in multinomial form. I could not find it in datafiles.
another thing i want to make clear.
Which of the following file will be suitable for taking account of abortion status i.e. (i) live birth (ii) miscarriage (iii) abortion (iv) spontaneous
1. BR File
2. KR File
3. IR File
Thank You
Waqas Imran
|
|
|
|
| Re: Pregnancy History [message #9475 is a reply to message #9265] |
Wed, 30 March 2016 20:07 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
A response from senior data processing expert, Mr. Guillermo Rojas:
Quote:
This is how that variable is constructed:
If S216 = 1 or S217 = 1 then Live birth
elseIf S216 = 2 then stillbirth
elseIf S218 =1 then abortion
elseIf S218 <> 1 then miscarriage
Those variables exist in file PKIR61FL, but in that file they have different variables for each child. For example S216_01 for the last child, S216_02 for the next to last and so on. The same applies for S217 and S218. The reason is because we don't have a pregnancy history file, what we have is a birth history (for children born alive) file (BR).
|
|
|
|
| Re: Pregnancy History [message #9910 is a reply to message #9475] |
Mon, 06 June 2016 05:12 |
|
|
D'r Liz,
I have worked out on file. I have found the S216&1-21 abd S217$1-21 in PKIR61FL File but i did not find the variables S218 there in PKIR61FL.
Variables S216$ provides following information:-
1. Born Alive
2. Born Dead
3. Lost Before Full Term
Can i consider above as:-
1. Born alive = live births
2. Born dead = Still birth
3. Lost Before Full Term = Miscarriage
I AM STILL FINDING SOME INFORMATION ABOUT ABORTION/MISCARRIAGES BECAUSE I COULD NOT FIND THE VARIABLE S218 EVEN IN FLAT (FL) FILES.
WAITING FOR REPORNSE
THANK YOU.
Waqas Imran
|
|
|
|
| Re: Pregnancy History [message #9925 is a reply to message #9910] |
Mon, 06 June 2016 15:11 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
There was an error in the earlier posting. Here is the proper coding:
Referring to S218 when it should be S228
If S216 = 1 or S217 = 1 then Live birth
elseIf S216 = 2 then stillbirth
elseIf S228 =1 then abortion
elseIf S228 <> 1 then miscarriage
|
|
|
|
|
|
| Re: Pregnancy History [message #10043 is a reply to message #9926] |
Sun, 19 June 2016 13:58 |
|
|
D'r Trever,
I am actually trying to develop some binary variable as below:-
Var1=If still birth=1 and 0 otherwise.
Var2=If miscarriage=1 and 0 other wise.
Var3=if Abortion=1 and 0 otherwise.
and a multinomial response variable as below:-
Var4=Livebirth=0
Var4=Miscarriage=1
Var4=abortion=2
var4=stillbirth=3.
Above outcome is explained in PDHS2013;Chapter Fertility; Page 80 Table 5.15. It is presented there as Pregnancy outcome.
S228 is not developed for Pakistan.
What would be its alternative in PKIR61FL file.
Thanksyou and regards,
Waqas Imran
Waqas Imran
|
|
|
|
|
|
| Re: Pregnancy History [message #10046 is a reply to message #10045] |
Mon, 20 June 2016 02:22 |
|
|
We have multiple separate viariables for each child against the following:-
idx97_ ord97_ s215_ s216_ s217_ s219_ s220m_ s220y_ ///
s221_ s222_ s223_ s224_ s225u_ s225n_ s226m_ s226y_ s227_ s228_.
Do i need to have to merge all children as below:-
For S215:-
Compute S215=$sysmiss.
If S215$1=1 or S215$2=1 or ..... or S215$21=1 S215=1.
If S215$1=0 or S215$2=0 or ..... or S215$21=0 S215=0.
Exe.
The same above strategies is for other variables.
I am asking above because the STAT codes you provided before showing only one child with addition '_' attribute with each variable.
More over only one entery is there in shared codes but Data file has 21 entries for each of variables mentioned above.
Thank You
Waqas Imran
|
|
|
|
|
|
| Re: Pregnancy History [message #10058 is a reply to message #10055] |
Tue, 21 June 2016 02:40 |
|
|
Currently the reshaped file providing the records of 57688 (unweighted) and 56748 (weighted by wt=v005/1000000) pregnancies.
Currently there are 14350 pregnancies ending in five years preceding the survey in Table 5.15; page 80; Chapter Fertility; PDHS 2013.
How can i achieve 14350 pregnancies?
Thank you.
Waqas Imran
|
|
|
|
|
|
| Re: Pregnancy History [message #10064 is a reply to message #10061] |
Tue, 21 June 2016 13:43 |
|
|
I learned alot from your support. Kindly tell me your full name so that i may acknowledge your name under acknowledgement heading of my paper.
Thank You so much.
Waqas Imran
|
|
|
|
|
|
| Re: Pregnancy History [message #10601 is a reply to message #10045] |
Mon, 15 August 2016 22:46 |
chr8850
Messages: 20
Registered: July 2016
|
Member |
|
|
Hi Trevor,
I was trying to generate a file of pregnancies using the Pakistan 2012-13 DHS and have replicated the code you provided in this thread using IR data but have also restricted to all pregnancies in the five years preceding the survey. I have pasted below the summary of pregnancy outcomes I get and additional code i ran. When I look at only pregnancies of 7 months or more I still get some miscarriages included and the number of stillbirths does not match what is in the DHS report (387 instead of 412). Any advice on what i might have done incorrectly or am missing? Thanks
keep if s220c > v008-60
tab pregoutcome [iw=v005/1000000]
pregoutcome | Freq. Percent Cum.
------------+-----------------------------------
Live birth | 11,977.381 83.47 83.47
Stillbirth | 399.217658 2.78 86.25
Abortion | 247.508767 1.72 87.98
Miscarriage | 1,725.5345 12.02 100.00
------------+-----------------------------------
Total | 14,349.642 100.00
ta pregoutcome if s227 >=7 [iw=v005/1000000]
pregoutcome | Freq. Percent Cum.
------------+-----------------------------------
Live birth | 11,977.381 96.58 96.58
Stillbirth | 386.60506 3.12 99.70
Abortion | 3.03187 0.02 99.72
Miscarriage | 34.470525 0.28 100.00
------------+-----------------------------------
Total | 12,401.489 100.00
|
|
|
|
| Re: Pregnancy History [message #10774 is a reply to message #10601] |
Sat, 10 September 2016 18:24 |
Trevor-DHS
Messages: 802
Registered: January 2013
|
Senior Member |
|
|
Yes, there are inconsistencies in the reporting of the outcome and the duration of pregnancy, with some pregnancies reported as ending with miscarriages at 7 months or more, and some pregnancies reported as stillbirths, but with a duration of only a couple of months. These are almost certainly reporting errors, but the data have not been edited to remove them. You will need to decide how you want to handle them in your analysis.
If you compare the numbers in your first table above with those in table 5.15 then the results match. I suspect though that you are comparing to table 8.5. This table is produced in a very different way, based on the calendar data, and following the approach used in most DHS that do not use a pregnancy history. In these surveys, stillbirths are calculated based on reported duration of pregnancy in the calendar of 7 months or more. Based on your second table above, that would be 424 births. However table 8.5 reports just 412 births. Because of the way the calendar is collected, some stillbirths are not included in the calendar, and thus table 8.5 may slightly undercount stillbirths. The calendar only permits one code to be captured in each month of the calendar. In the case of stillbirths, if there are twins and one is a stillbirth and one a live birth, then only the live birth is captured in the calendar. Similarly if the pregnancy was twins and both were stillbirths, then again, only one stillbirth is captured in thye calendar. Thus in these two cases, stillbirths would be missed and not counted in the numbers in table 8.5
|
|
|
|
| Re: Pregnancy History [message #10852 is a reply to message #10774] |
Sun, 25 September 2016 00:49 |
chr8850
Messages: 20
Registered: July 2016
|
Member |
|
|
Thanks so much for all of that info Trevor. Is it usual practice to code any fetal deaths where the gestational age (s227) is unknown as a miscarriage?
Based on the following code on another forum (msg #10623). Why does DHS not just drop these deaths rather than code as miscarriage?
gen type = .
replace type = 1 if s216==1 | s217==1
replace type = 2 if type==. & s227 >= 7 & s227 < 98
replace type = 3 if type==. & (s227 < 7 | s227 >= 98) & s228 == 1
replace type = 4 if type==. & (s227 < 7 | s227 >= 98)
lab def type 1 "Live birth" 2 "stillbirth" 3 "Abortion" 4 "Miscarriage"
lab val type type
lab var type "Type of pregnancy"
|
|
|
|
|
|
| Re: Pregnancy History [message #12304 is a reply to message #10045] |
Thu, 27 April 2017 08:35 |
gafap
Messages: 3
Registered: April 2017
|
Member |
|
|
Dear all,
I have a related question. I'm following the code before for reshaping the data into a pregnancy level dataset. I start from the IR (women) dataset, for Armenia 2010 DHS. After correcting for variable names, the code I run is below.
I want to replicate table 8.1 from the DHS 2010 report. But I cannot even get the number of pregnancies right. In report, the total number of pregnancies in the 3 years prior to survey are 1,457. After my code, I get 1,368. Hence I am missing some pregnancies somehow. Can you help?
Thanks!
************************************************************ *********************************
use "AMIR61FL.DTA", replace
gen weights=v005/1000000
* Rename variables to permit reshaping
ren bidx_0* bidx_*
ren bord_0* bord_*
ren s2*_0* s2*_*
ren b0_0* b0_*
* keep the variables ytou need - you may need more than this list
keep caseid weights v013 v008 v102 s209a s215a_* s222c_* s214y_* s214m_* s214c_* bidx_*
reshape long s215a_ s222c_ s214y_ s214m_ s214c_ bidx_, i(caseid) j(order)
ren s2*_ s2*
ren bidx_ bidx
* drop empty pregnancies
drop if bidx == .
cap drop bidx_1 -bidx_24
* Keep those who were aborted (born) within the last 3 years from interview
cap drop timesincepreg
gen timesincepreg = v008 - s214c
tab s215a if timesincepreg<=36
s215a | Freq. Percent Cum.
------------+-----------------------------------
live birth | 913 66.74 66.74
still birth | 3 0.22 66.96
miscarriage | 81 5.92 72.88
abortion | 371 27.12 100.00
------------+-----------------------------------
Total | 1,368 100.00
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
