The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Biomarkers » Blood Pressure and Diabetes Bangladesh 2017-18 (Biomarker )
Blood Pressure and Diabetes Bangladesh 2017-18 [message #21794] Tue, 22 December 2020 02:08 Go to next message
bakerchowdhury
Messages: 25
Registered: April 2014
Member
Hello,
I am using the Bangladesh DHS 2017-18 to look at hypertension and diabetes. I am trying to match the tables in Stata by using Household Member Recode (BDPR7RFL) file.

First I am selecting the participants who answered yes to "sh18/sh19/sh20" variables which generate 6691 men and 8015 women; who were eligible for bp and glucose measurements.

However, I am unable to match the numbers in table 13.3.1 second column "Prevalence of raised blood pressure (hypertension)^1" by residence or division. I am using the "sb333aa" and "sb333ab" for systolic and diastolic pressure, respectively. However, I do not find any variable to consider hypertension medication.

For diabetes, I am using "sb335b" and "sb327a" to create the diabetes variable in table 13.5.51 second column "Prevalence of diabetes^1".

I created weight [wgt=hv005/1000000] and then used svyset[pw=wgt],psu(hv021) strata(hv022). To tabulate the frequencies I used svy:tab hv025 diab and svy:tab hv025 hp. Here diab, and hp are for diabetes and hypertension respectively.

I wondered if anyone would be able to advise me if I am using the right dataset/sample weights/ most importantly right variables to determine diabetes or hypertension.

Any suggestions for the variable that would be used would be much appreciated, thank you :)
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #21813 is a reply to message #21794] Wed, 23 December 2020 06:34 Go to previous messageGo to next message
SaraDHS is currently offline  SaraDHS
Messages: 46
Registered: December 2020
Member
You are using the correct datafile and weighting variables. The BP medication variable is sb318a. The attached do file shows how to recreate the values for the variables in tables 13.3.1, 13.3.4, 13.5.1, and 13.5.4. The classifications can be tricky.

Sara Riese, PhD Senior Demographic and Health Researcher, DHS Program
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #21835 is a reply to message #21813] Mon, 28 December 2020 18:52 Go to previous messageGo to next message
bakerchowdhury
Messages: 25
Registered: April 2014
Member
Thank you so much for your detailed response. Could you please help in replicating the percentages of Figure 13.6 and 13.10 on the report page and 2019 and 221, respectively. I am unable to match them.

Thank you again.
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #21871 is a reply to message #21835] Mon, 04 January 2021 06:46 Go to previous messageGo to next message
SaraDHS is currently offline  SaraDHS
Messages: 46
Registered: December 2020
Member
The "awareness" variable is sb317b for blood pressure and sb326b for diabetes. If you use that variable, along with the classification (bpclass_m/bpclass_w/diabclass_m/diabclass_w), medication (bpmed/diabmed), and hypertension/diabetes (hypert_m/hypert_w/diab_m/diab_w) variables created in the do file, you should be able to recreate those figures. "Controlled" is anything not Hyperglycemia for diabetes classification, and the "optimal" and "normal/high normal" classifications for blood pressure.

Best,
Sara


Sara Riese, PhD Senior Demographic and Health Researcher, DHS Program
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #22918 is a reply to message #21871] Thu, 03 June 2021 05:34 Go to previous messageGo to next message
gueter89@yahoo.fr is currently offline  gueter89@yahoo.fr
Messages: 10
Registered: November 2018
Member
I had a major problem with the documents related to DHS because they do not mention information about the variables of hypertension and diabetes. Neither the Guide nor the Recoding Manual provides information on these two variables. For several weeks, I have been using the Haiti database (PR file) and I still cannot identify the variables related to diabetes. Moreover, the sample for which we surveyed blood pressure, diabetes, nutritional status.
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #23175 is a reply to message #22918] Tue, 27 July 2021 05:49 Go to previous messageGo to next message
SaraDHS is currently offline  SaraDHS
Messages: 46
Registered: December 2020
Member
Dear Gueter,

Thank you for your message. Yes, the blood pressure and diabetes questions, as well as the larger module they are included in, the non-communicable disease module, are optional and have not been included as of yet in Github code or the guide to statistics. We hope to include them in the future.

I have been looking into the specific issue with the Haiti data, and unfortunately it appears that an error was made during file construction, and many of the diabetes and hypertension variables were not included in the Individual Men's and Women's file. By time this was recognized, the original data file had been erased. sbh263c and shb381a (glucose measurement for men and women is still available in the PR file.

Thank you for the question and apologies that we don't have a better answer.

Sara



Sara Riese, PhD Senior Demographic and Health Researcher, DHS Program
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #30017 is a reply to message #21813] Wed, 11 September 2024 02:20 Go to previous messageGo to next message
bakerchowdhury
Messages: 25
Registered: April 2014
Member
Dear Dr. Riese,
Thank you so much for your continuous support with the data. I am currently reviewing the 2022 BDHS dataset, specifically focusing on the diabetes and hypertension variables, and am working with the Household Member Recode (BDPR81FL) file.

I have noticed that systolic, diastolic, and blood glucose values are recorded in separate columns for men and women. For instance, `wbp24` and `wbp25` represent systolic and diastolic values for women, while `mbp24` and `mbp25` are for men. Additionally, `wbp19` and `wbp20` denote blood pressure medication for women and men, respectively. Similarly, `sb267g` and `sb367g` represent blood glucose values for women and men, while `sb240` and `sb340` indicate diabetes medication for men and women, respectively.

Could you please advise on the best approach to consolidate this data into single columns for:
- Systolic values for both men and women
- Diastolic values for both men and women
- Blood pressure medications for both men and women

along with other characteristics such as age, sex, region, education, wealth index, and BMI, so that I can compare these with the previous survey (2017-18 BDHS) iteration? Additionally, I am looking to obtain an overall prevalence of diabetes and hypertension rather than calculating them separately for each gender.

Below is the code I have used to create the hypertension and diabetes variables for each gender separately, but I would like to have them into a single dataset for both genders.

use "BDPR81FL", replace


rename wbp24 w_systolic
rename wbp25 w_diastolic
sum w_systolic w_diastolic

gen validbp = 0
replace validbp =1 if w_systolic!=. & w_diastolic!=.
tab validbp,m

drop if validbp==0

gen w_bpclass = 0 if hv104 ==2 & validbp==1
replace w_bpclass = 1 if (w_systolic<120 & w_diastolic<80) & validbp==1
replace w_bpclass = 2 if (w_systolic>=120 & w_systolic<130) | (w_diastolic>=80 & w_diastolic<85) & validbp==1
replace w_bpclass = 3 if (w_systolic>=130 & w_systolic<140) | (w_diastolic>=85 & w_diastolic<90) & validbp==1
replace w_bpclass = 4 if (w_systolic>=140 & w_systolic<160) | (w_diastolic>=90 & w_diastolic<100) & validbp==1
replace w_bpclass = 5 if (w_systolic>=160 & w_systolic<180) | (w_diastolic>=100 & w_diastolic<110) & validbp==1
replace w_bpclass = 6 if (w_systolic>=180) | (w_diastolic>=110) & validbp==1

svy: tab w_bpclass

With this code the women hypertension prevalence matches however, I am getting no of women 7899, however report shows 7724 (Table 14.3.1)

Thank you again for your invaluable assistance.
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #30032 is a reply to message #30017] Fri, 13 September 2024 11:40 Go to previous messageGo to next message
SaraDHS is currently offline  SaraDHS
Messages: 46
Registered: December 2020
Member
Hello,
In order to consolidate the BP and other data into a single variable for both men and women you could do something like:

gen systolic=.
gen diastolic=.
replace systolic=w_systolic if hv104==2
replace systolic=m_systolic if hv104==1
replace diastolic=w_diastolic if hv104==2
replace diastolic=m_diastolic if hv104==1

All the variables of age, sex, region, education, wealth, and BMI are available in the PR dataset for the 2022 Bangladesh DHS, as well as for the 2017-18 DHS, with the same variable names, so you should be able to compare these easily.

Best,
Sara


Sara Riese, PhD Senior Demographic and Health Researcher, DHS Program
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #30042 is a reply to message #30032] Sat, 14 September 2024 03:34 Go to previous messageGo to next message
bakerchowdhury
Messages: 25
Registered: April 2014
Member
Thank you for your reply. The BP code worked. However, I couldn't find BMI variable. Could you please mention which variable in 2022 BDHS I need to use for BMI. Also, what condition do I need to use to get some individual characteristics (e.g. anxiety) from IR file?

Thank you!
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #30053 is a reply to message #30042] Mon, 16 September 2024 16:45 Go to previous messageGo to next message
SaraDHS is currently offline  SaraDHS
Messages: 46
Registered: December 2020
Member
Hello,
you can find the BMI variables if you search "body mass index" in the variable manager.

As for merging characteristics from the IR file onto the PR file, I would recommend searching the user forum for resources. Here, for example, is another thread that has lots of details about this merge: https://userforum.dhsprogram.com/index.php?t=msg&th=1249 0&start=0&

Best of luck!
Sara


Sara Riese, PhD Senior Demographic and Health Researcher, DHS Program
Re: Blood Pressure and Diabetes Bangladesh 2017-18 [message #30056 is a reply to message #21794] Tue, 17 September 2024 05:18 Go to previous message
matthewhancock is currently offline  matthewhancock
Messages: 1
Registered: September 2024
Member
For hypertension, use systolic and diastolic pressure thresholds to identify cases. Check if there's a variable for hypertension medication to include in your analysis. For diabetes, ensure you're applying the correct glucose level thresholds. Your use of weights and survey settings seems correct. Verify that you're using the right variables and thresholds as specified in the dataset documentation.
Previous Topic: Air France BRI Terminal
Next Topic: Difficulty in finding blood pressure measurement values for male
Goto Forum:
  


Current Time: Fri Dec 6 11:15:51 Coordinated Universal Time 2024