The DHS Program User Forum: Mortality » What Should be Used "And (&)" "or "OR (|)

Home » Topics » Mortality » What Should be Used "And (&)" "or "OR (|)

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

What Should be Used "And (&)" "or "OR (|) [message #20388]

Tue, 27 October 2020 05:08

shujaat.smc@gmail.com is currently offline

shujaat.smc@gmail.com
Messages: 75
Registered: July 2020

Senior Member

Dear DHS Representative,

Before performing descriptives and inferential analysis, after declaring my data set (BR file PDHS 2017-18). I run the following command for my univariate analysis;

svy linearized, subpop(if bidx==1 & b0==0 & v210!=1) : stcox ib(0). HepBTest

Quest: I am using subpopulation for those neonates who were last born to the mothers included in the survey (bidx==1), who was the product of singleton pregnancy (b0==0), and who were not born in the month of conducting the survey (v210!=1). Kindly confirm and please also explain is this correct to use "AND (&)" in between my subpopulation variables. This query pops up in my mind after seeing the coding on the git-hub " https://github.com/DHSProgram/DHS-Indicators-Stata/blob/mast er/Chap09_RH/RH_DEL.do" over here the "OR (|) is used.

Quest: What is meant by "if age>=period" I am asking this question because in my analysis up till now I have confidence in my sup-population, however after seeing the above git-hub link I in doubt regarding AND(&) OR(|) use. Also, I have not used any stata command if age>=period up till now.

I am working on the determinants of neonatal mortality in Pakistan (BR-file 2017-18) using STATA 13.1.

Waiting for your reply.

Best Regards

Dr. Hussain

Report message to a moderator

Re: What Should be Used "And (&)" "or "OR (|) [message #20496 is a reply to message #20388]

Thu, 12 November 2020 09:11

Shireen-DHS
Messages: 140
Registered: August 2020
Location: USA

Senior Member

Dear Dr. Hussain,

Based on your description of your population of interest it seems you should be using & which is AND not OR. However, I am not sure why you would need to exclude birth that occurred in the month of the interview using v210. I cannot comment on your model or your analysis.

For the age/period question, this comes from the RH_age_period.do file on our GitHub site: https://github.com/DHSProgram/DHS-Indicators-Stata/tree/mast er/Chap09_RH

Age is the age of the child and period here is defined as the last five years before the survey. For some indicators the estimate is reported in the last 3 years or last 2 years (check the final report for your indicator of interest) so we create the variable period so the user can only change it one time.

So for BR files it is:
gen age = v008 - b3
*if you have b19 in the file you need to use that for age instead so age would be : age=b19
*only recent surveys have b19
gen period = 60
* 60 is 5 years in months

Thank you.

Best,
Shireen Assaf
The DHS Program

Report message to a moderator

Re: What Should be Used "And (&)" "or "OR (|) [message #20501 is a reply to message #20496]

Thu, 12 November 2020 09:35

shujaat.smc@gmail.com
Messages: 75
Registered: July 2020

Senior Member

Dear Shireen,

Thanks for your reply. I am excluding those birth which occurred in the month of conducting the survey (v210!=1) because according to my understanding about the DHS statistics (book/guide); These newborns lie in cohort C that is considered to be partially exposed. They will not be eligible for the numerator.

Restricting my analysis to the most recent birth a mother had in the preceding five years before the conduct of the survey (bidx==1) is due to the reason that P-DHS only takes ANC information for the last or most recent birth a woman had.

Restriction for singleton livebirth (b0==0) is to avoid the information bias in my estimates.

Please comment do these reasons are acceptable to the scientific community.

Best Regards

Dr. Hussain

Report message to a moderator

Re: What Should be Used "And (&)" "or "OR (|) [message #20515 is a reply to message #20501]

Fri, 13 November 2020 08:03

Bridgette-DHS
Messages: 3214
Registered: February 2013

Senior Member

Following is a response from Senior DHS Specialists, Tom Pullum & Shireen Assaf:

These restrictions are ok, but we recommend that you be explicit about them in the text or in a footnote. The restriction to bidx=1 is unavoidable, since the data were only collected for the youngest child, but when relevant that restriction should always be stated, because some other indicators apply to all births in the past five years or all surviving births. You justify the restriction to singletons as a desire to avoid "information bias". You could be more specific about what that means. A child in a multiple birth is definitely at higher risk but would always be included in a neonatal mortality rate, for example. Otherwise the rate is biased downwards.

DHS treatment of births in the month of interview is to include them in the counts of children ever born and children born in the past five years, and include them as the most recent birth for questions about the most recent birth. They are not included in fertility rates, e.g. the fertility rates for the last three years. Births and exposure for the fertility rates begin with the month before the month of interview, because the month of interview has less than a full month of exposure, as you say.

Report message to a moderator

Re: What Should be Used "And (&)" "or "OR (|) [message #20517 is a reply to message #20515]

Fri, 13 November 2020 08:18

shujaat.smc@gmail.com
Messages: 75
Registered: July 2020

Senior Member

Dear Tom Pullum & Shireen,

Thanks for your reply, now I am more confident about my subpopulation for analysis and will give the reasons in my thesis & manuscript for this restriction.

I will also discuss your reply with my thesis committee and in case of any further questions, I will share it on the forum.

Best Regards

Dr. Hussain

Report message to a moderator

Previous Topic:	Comparison Between Birth weight Variable m19 & m18
Next Topic:	Age of Mother at Time of Most Recent Birth

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Tue Dec 24 13:09:05 Coordinated Universal Time 2024