The DHS Program User Forum: General » Operationalizing Menopausal Status

Home » Topics » General » Operationalizing Menopausal Status (How to calculate prevalence of menopausal women using V226?)

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

Operationalizing Menopausal Status [message #30202]

Mon, 14 October 2024 06:34

ANWER
Messages: 3
Registered: October 2024

Member

Hello,

We assume we should use V226 to calculate the age at menopause or the prevalence of women in menopause (please advise otherwise). According to the DHS manual, V226 is the computed time since the last menstrual period. This is computed from the response for V215, with durations exceeding the interval since the last birth (V227 = 7, 9) recoded to the response "Before last birth" (code 995) and inconsistent responses flagged on variable V227 (codes 1-6) recoded to 997.

In the data, the variable is categorized as 0-339 (indicates months before last menstruation), along with the following other categories:
IN MENOPAUSE/HAS HAD HYSTERECTOMY 994
BEFORE LAST BIRTH 995
NEVER MENSTRUATED 996
INCONSISTENT 997
MISSING 998

For calculating women's age at menopause, we devise the formulae: V226_age = (V012 - V226/12). While we can, by definition, recode categories 994, 996, AND 997 as "not menopausal", our queries are:
1. How may we deal with observations in 995(BEFORE LAST BIRTH)? They comprise a significant share of respondents (8%-12% in South and Southeast Asian countries), so we do not want to drop them.
2. Are these women (995 in V226) really in menopause or could they also be just in post-partum Amenorrhea?
3. If they are menopausal, how do we replace "995" for these observations with the time since their last menstruation to correctly calculate V226_age = (V012 - V226/12)?

Thank you very much in advance!

Best,

Anwer

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30213 is a reply to message #30202]

Tue, 15 October 2024 14:05

Janet-DHS
Messages: 938
Registered: April 2022

Senior Member

Thank you for submitting your question. Could you please provide some more detailed information so we can better advise you?
Can you let us know:
• Which survey you are using (Include country name and year)?
• Which data files you are referring to?
• Which software you are using (Stata, SPSS, R, etc.)?
If you are trying to match a Table in a final report, please also indicate which table and which estimate you are trying to match.

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30227 is a reply to message #30213]

Fri, 18 October 2024 15:17

ANWER
Messages: 3
Registered: October 2024

Member

We used IR data files from 53 DHSs conducted across 12 South and Southeast Asian countries from 1986 to 2022 and analyzed the data in RStudio. One of our objectives is to explore the prevalence trends of premature (occurring before age 40) and early menopause (occurring between age 40 and 44) in these countries. We understand our estimates would not match those, if any, in the DHS reports as we modified the DHS definition of menopause extending the no-menstruation period from six to twelve months which aligns with the clinical definition of menopause. This necessitates appropriately dealing with "Before last birth" categories in V226s. We look forward to hearing from you!

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30275 is a reply to message #30227]

Fri, 25 October 2024 14:36

Janet-DHS
Messages: 938
Registered: April 2022

Senior Member

Following is a response from DHS staff member, Tom Pullum:

I'm not sure I understand your question, but I believe you are asking how to interpret code 995, "before last birth" for the variable v226, "time since last period (months)". That code means that the woman has not menstruated since her latest live birth.

Here are the Stata lines we use to construct the binary outcome, using the 6-month definition. "<6 months" would be the values 0,1,2,3,4,5, and that's the reason for "v226>5" in the construction. I think the only thing you would need to do is change that to "v226>11".

     
              gen fe_meno = 0 if  v013>3

              replace fe_meno = 1 if (v226>5 & v226<997) & v213==0 & v405==0  & v013>3

You will see that nothing special is done for code 995. "v405=0," in addition to code 995, would mean that she has not menstruated since her last birth and she is past the interval for postpartum amenorrhea. I think the change to 12 months would not affect the construction in any way other than the change from 5 to 11. I can imagine that you might want to include a condition for the interval since the last birth, but I don't think that's needed.

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30296 is a reply to message #30275]

Wed, 30 October 2024 10:52

ANWER
Messages: 3
Registered: October 2024

Member

Hi Tom, thank you very much for your response.

I understand you mean women belonging to 995 in V226 are truly menopausal as they had their last menstruation before their last births and they were past the interval for postpartum amenorrhea. This clarifies the first part of our queries (though there is an issue that I will get back to you later).

The next problem is how we determine women's age at menopause as our interest is not only determining their menopausal status (binary outcome). You will agree with us that this is straightforward using our formulae: V226_age = (V012 - V226/12), given that we exclude special responses (994, 995, 996, and 997) from the sample. But, what do you suggest we do if we are to determine the age at menopause of these women belonging to special categories in V226? This is particularly important because these women comprise a significant share of samples in SA/SEA countries, so excluding them would bias the prevalence (of premature and early menopause) estimates. Is there any way that the original variable V215 might be helpful in this regard?

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30309 is a reply to message #30296]

Thu, 31 October 2024 17:09

Janet-DHS
Messages: 938
Registered: April 2022

Senior Member

Following is a response from DHS staff member, Tom Pullum:

You could improve your estimate of age at last menstruation as follows. The woman's age in months at the time of the survey is v008-v011. Her current age in years is given by v012=int((v008-v011)/12). Therefore her months of age at last menstruation is v008-v011-v226 and her completed years of age at that time is int((v008-v011-v226)/12).

You are asking how to calculate age at menopause for women in the special categories of v226:

994 in menopause
995 before last pregnancy
996 never menstruated
997 inconsistent
998 don't know

I picked a survey (Philippines 2022) and entered "tab v012 v226 if v226 >=994". 998 does not occur (a response is required). The previous post dealt with 995. 996 is rare and mostly young women. The problematic categories are 994 and 997. I see that 994 is almost completely women age 40+ and 997 is almost never women age 40+. But I will not offer any suggestions for how to estimate when women with codes 994 or 997 last menstruated. Age at menopause is not a DHS indicator but I hope there is a literature that would give guidelines, or help you develop guidelines, for how to code these categories.

Report message to a moderator

Re: Operationalizing Menopausal Status [message #30978 is a reply to message #30202]

Sun, 11 May 2025 04:06

Jeannaldez
Messages: 1
Registered: May 2025

Member

To replace "995" with time since last menstruation, you could consider using an imputation method based on available data (like age at last birth or a related variable), though this will depend on the available dataset and your assumptions.

Report message to a moderator

Previous Topic:	Using Special DHS data
Next Topic:	Find data

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Thu May 15 09:21:35 Coordinated Universal Time 2025