The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Child Health » Tuberculosis and Childhood Tuberculosis (Not matching with the state report)
Tuberculosis and Childhood Tuberculosis [message #24569] Thu, 02 June 2022 01:35 Go to next message
NKS is currently offline  NKS
Messages: 2
Registered: June 2022
Member
Hello!

I am trying to estimate the point prevalence of TB per 100,000 population for all as well as childhood TB from PR file of NFHS-5 (2019-21). The variable I have used is sh29aa. I am using STATA 14 for the analysis with the following command:

gen tb_mtreated =0
replace tb_mtreated =1 if sh29aa>=1 & sh29aa<=3
proportion tb_mtreated if hv102==1 [iw=hv005/1000000]
proportion tb_mtreated if hv102==1 & hv024==23 [iw=hv005/1000000], over (shdist)

The point prevalence by age groups, rural/urban and total is matching with the national report at India level. However, when I am doing the same analysis with same command at state level, I found huge differences in point prevalence by age groups and rural/urban at state level. Also, I would like to mention that the total number of sample is matching with the state reports but not the estimates of point prevalence.

I humbly request to solve this problem on priority basis.

[Updated on: Thu, 02 June 2022 05:40]

Report message to a moderator

Re: Tuberculosis and Childhood Tuberculosis [message #24582 is a reply to message #24569] Fri, 03 June 2022 08:47 Go to previous messageGo to next message
Janet-DHS is currently offline  Janet-DHS
Messages: 899
Registered: April 2022
Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

The problem may be that you are using hv005 and iweights (iw). Please repeat, using the state-specific weight for within-state estimates. Please let us know if you still get a discrepancy.
Re: Tuberculosis and Childhood Tuberculosis [message #24614 is a reply to message #24582] Wed, 08 June 2022 03:27 Go to previous message
NKS is currently offline  NKS
Messages: 2
Registered: June 2022
Member
Thank you for the quick response.

I have already applied the state weight and IW. The number of usual residents for states which is "N" is matching with the state report while point prevalence by age-group and gender is not. For instance, point prevalence among <15 years for UP was estimated as 34/100,000 from the data but in the report, it is 356/100,000.

Here is the full command of STATA, which I have employed for the state of Uttar Pradesh (if hv024==9):

gen tb_mtreated =0
replace tb_mtreated =1 if sh29aa>=1 & sh29aa<=3
proportion tb_mtreated if hv102==1 [iw=hv005/1000000]

recode hv105 (0/14=1 "0-14 years") (15/59=2 "15-59 years") (else=3 ">=60 years"), gen (age)

proportion tb_mtreated if hv102==1 & hv024==9 [iw=shweight/1000000], over (age)
proportion tb_mtreated if hv102==1 & hv024==9 [iw=shweight/1000000], over (hv104)
proportion tb_mtreated if hv102==1 & hv024==9 [iw=shweight/1000000]

//Results from STATA//

proportion tb_mtreated if hv102==1 & hv024==9 [iw=shweight/1000000], over (age)

Proportion estimation Number of obs = 364,194

_prop_1: tb_mtreated = 0
_prop_2: tb_mtreated = 1

_subpop_1: age = <15
_subpop_2: age = 15-59
_subpop_3: age = >=60


Over Proportion Std. Err. [95% Conf. Interval]

_prop_1
_subpop_1 .9996582 .0000553 .9995307 .9997511
_subpop_2 .9975473 .0001065 .9973295 .9977473
_subpop_3 .994669 .0003812 .9938673 .9953665

_prop_2
_subpop_1 .0003418 .0000553 .0002489 .0004693
_subpop_2 .0024527 .0001065 .0022527 .0026705
_subpop_3 .005331 .0003812 .0046335 .0061327

Request you to kindly look at the discrepancies.
Previous Topic: duplicate caseid
Next Topic: Birth Month
Goto Forum:
  


Current Time: Sat Dec 7 08:44:20 Coordinated Universal Time 2024