Home » Countries » Other countries » Any Anaemia Ghana DHS 2022 (I got a different value)
| Any Anaemia Ghana DHS 2022 [message #30656] |
Tue, 14 January 2025 22:23  |
backclac
Messages: 28
Registered: August 2023
|
Member |
|
|
Hello
I am analyzing anaemia using KR files from the GDHS 2008, 2014, and 2022. The prevalence of "any anaemia" among 6-59 month-olds in rural areas (v025) differs from the GDHS report. Below is the STATA do.file syntax used:
* Set the survey design variables
*weighted
gen wt822=v005/1000000
*Preparing the dataset as a survey Dataset
svyset [pweight=wt822], strata(v023)psu(v021)
*****************RECODE VARIABLES**********************************
* dataset for only rural Ghana
tab v025
keep if v025==2
tab v025
*Outcome
*tab hw57
**outcome variables**
tab hw57
* adding all forms of anaemia together as one ALLanemia22
recode hw57(1 2 3=1 "anemia") (4 9=0 "not_anemic"), gen (ALLanemia822)
tab ALLanemia822
My results: 1258, 1061 and 1334 anaemic children(6-59 months) for the years 2008, 2014 and 2022, respectively
Where did I go wrong? Please advise me.
[Updated on: Tue, 14 January 2025 22:24] Report message to a moderator |
|
|
|
|
|
|
|
| Re: Any Anaemia Ghana DHS 2022 [message #30701 is a reply to message #30663] |
Tue, 21 January 2025 15:59  |
Janet-DHS
Messages: 932
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
Biomarkers such as anemia, which is based on blood tests conducted on children during the household survey, can be found in both the PR and KR files. It is not always which file was used to construct a table.
You are apparently trying to match table 11.7 in the final report on the Ghana 2008 survey. It appears that this table was constructed with the KR file, because it contains the total of 2313 children age 6-59 months, but I see that there is indeed a problem with this table. Using the KR file (GHKR5AFL.dta), the following command should give the bottom row of the table, except for "any", which is obtained by adding up "severe", "moderate", and "mild":
tab hw57 if hw1>=6 [iweight=v005/1000000]
. tab hw57 if hw1>=6 [iweight=v005/1000000]
anemia |
level | Freq. Percent Cum.
------------+-----------------------------------
severe | 164.444313 7.11 7.11
moderate | 1,016.7676 43.95 51.06
mild | 478.457266 20.68 71.74
not anemic | 456.775979 19.74 91.49
9 | 196.975826 8.51 100.00
------------+-----------------------------------
Total | 2,313.421 100.00
Comparing with the bottom row of Table 11.9, this matches the total frequency, 2313. However, the percentages in the "mild", "moderate", and "severe" categories are a little too low and there are 197 (weighted) cases with the illegal code "9".
The other panels of the table should be obtained by inserting covariates after "tab". For example, the urban/rural panel should be obtained with
tab v025 hw57 if hw1>=6 [iweight=v005/1000000], row
Unfortunately, all the panels are a little off, in terms of both frequencies and percentages.
I believe that something was done behind the scenes, or outside of the KR file, to re-assign those 197 cases (the unweighted number is 193) in terms of values on the covariates as well as anemia category. I could pursue that hypothesis, but I just don't think it's worth the time.
Next, I believe that your more immediate interest is in the 2022 survey, and I think you want to match Table 11.12 in the final report on that survey.
I tried to match it with the KR file, but it turns out that the 2022 version of the anemia table comes from the PR file (GHPR8CFL.dta). You can match the bottom row of this table, except for the "any" column, with this command:
tab hc57 if hc1>=6 & hc1<=59 & hv103==1 [iweight=hv005/1000000]
tab hv025 hc57 if hc1>=6 & hc1<=59 & hv103==1 [iweight=hv005/1000000], row |
|
|
|
| Re: Any Anaemia Ghana DHS 2022 [message #30709 is a reply to message #30701] |
Tue, 21 January 2025 20:44 |
backclac
Messages: 28
Registered: August 2023
|
Member |
|
|
The feedback is helpful and I appreciate it.
Kindly help me address the following;
1. This study will consider the anaemia tables(data) from 2008, 2014, and 2022. We would be glad if you could help us resolve all challenges or advise us on what to do.
2. Is it appropriate to do a trend or comparison analysis using the KR file in one and the PR file in a different year
Please advise me. Thank you
|
|
|
|
Goto Forum:
Current Time: Thu Jan 23 18:17:29 Coordinated Universal Time 2025
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
