Home » Topics » Mortality » Calculating Under 5 Mortality Rate
| Calculating Under 5 Mortality Rate [message #9327] |
Fri, 11 March 2016 11:46  |
gemma1224
Messages: 2
Registered: March 2016
|
Member |
|
|
I'm trying to estimate U5 mortality rates in India (98,06) and Kenya (98,08).
I'm using this code in Stata and am getting close, within 1-2 people, of the published DHS rate, but not exact matches. Can anyone help me see what I am missing?
Very appreciative of any suggestions.
I've created this code for the BR DHS file following http://siteresources.worldbank.org/INTPAH/Resources/Publicat ions/459843-1195594469249/HealthEquityCh3.pdf and http://legacy.measuredhs.com/help/datasets/
gen hypage=(v008-b3)
gen survivelength=.
replace survivelength=hypage
replace survivelength=b7 if b5==0
gen dead=(b5==0)
ltable survivelength dead [fw=v005] if hypage <60 , int(0,1,3,6,12,24,36,48,60) failure
|
|
|
|
|
|
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #9867 is a reply to message #9327] |
Wed, 01 June 2016 21:31  |
dhs_bee
Messages: 1
Registered: June 2016
|
Member |
|
|
I have a similar code as gemma1224 for KDHS and my U5 estimates a 1-2 people away from the DHS report.
gen hypage=(v008-b3)
gen timonths=.
replace timonths=hypage
replace timonths=b7 if b5==0
gen dead=(b5==0)
ltable timonths dead [fw=v005] if hypage <60 , int(0,1,3,6,12,24,36,48,59) failure
Is there a fix for it?
I'm using stata13
Thanks!
|
|
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #10863 is a reply to message #9867] |
Mon, 26 September 2016 15:50 |
lonrwo1
Messages: 3
Registered: September 2016
|
Member |
|
|
Dear Sir or Madam,
For MSc Epi dissertation; I'm running a poisson regression (including lexis expansion/stsplit for agegroup) analysis of Kenya DHS 2009 for U5 mortality. ltable U5 rate, then using the follow-up time variable (survive length) in stset before moving onto lexis expansion (stsplit) etc.
When using the following code for synthetic l table U5 rate calculation method,
gen hypage=(v008-b3)
gen survivelength=.
replace survivelength=hypage
replace survivelength=b7 if b5==0
gen dead(b5==)
ltable survivelength dead if hypage <60, int(0,1,3,6,12,24,36,48,69)
The overall U5 mortality rate as per the DHS report is calculated (74/1,000). However when I add by(province) option the stratum specific rates for province are completely different from province specific rates reported in the DHS report (page 106). I attach dofile and dhs report for reference.
Could you shed any light on this?
Thanks,
Rob
|
|
|
|
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #10874 is a reply to message #10872] |
Tue, 27 September 2016 12:16 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
A response from Dr. Tom Pullum:
Quote:
The calculation of the under-five rates is vastly more complex. DHS uses a synthetic (or life table) approach. Deaths and risk are assessed within an interval of time, such as the five years before the survey, and within eight sub-intervals of age. The method is described in the Guide to DHS Statistics https:// www.dhsprogram.com/pubs/pdf/DHSG1/Guide_to_DHS_Statistics_29 Oct2012_DHSG1.pdf.
It happens that within the next couple of weeks I will be putting a Stata program onto the DHS website that calculates these rates. I'm afraid you will have to wait until then.
Thank you for your post.
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #10880 is a reply to message #10874] |
Wed, 28 September 2016 01:25 |
lonrwo1
Messages: 3
Registered: September 2016
|
Member |
|
|
Thank you Liz/Dr Pullum,
Thank you very much indeed for your reply. Unfortunately I'm working to an end of the week deadline, so the stata program in two weeks will sadly be of no use to me.
My understanding was that the code I am running to generate U5 mortality rate matches what is described by Dr Tom in his post and also on section C page 94 of the Guide to DHS Statistics.
gen hypage=(v008-b3)
gen survivelength=.
replace survivelength=hypage
replace survivelength=b7 if b5==0
gen dead=(b5==0)
ltable survivelength dead if hypage <60, int(0,1,3,6,12,24,36,48,60) failure
When running the above code I get the following lifetable output
Beg. Cum. Std.
Interval Total Deaths Lost Failure Error [95% Conf. Int.]
------------------------------------------------------------ -------------------
0 1 6079 176 61 0.0291 0.0022 0.0252 0.0337
1 3 5842 38 219 0.0355 0.0024 0.0311 0.0405
3 6 5585 38 315 0.0423 0.0026 0.0375 0.0477
6 12 5232 47 652 0.0515 0.0029 0.0461 0.0575
12 24 4533 53 1119 0.0641 0.0033 0.0579 0.0710
24 36 3361 14 1147 0.0688 0.0036 0.0622 0.0761
36 48 2200 6 1093 0.0722 0.0038 0.0651 0.0800
48 60 1101 1 1100 0.0739 0.0042 0.0662 0.0825
------------------------------------------------------------ -------------------
You can see that the cumulative failure for the last interval 48 - 60 months is 0.739, x 1000 =73.9 or 74 per 1,000. This is the U5 mortality rate reported in the KDHS 2009 - Fig 8.2 page 107, so this seems to me, to confirm that this is the correct method to use. My confusion comes from the fact that when including other variables e.g. by(province) to the ltable code i.e.
ltable survivelength dead if hypage <60, by (province) int(0,1,3,6,12,24,36,48,60) failure
the U5 rates (cumulative failure for last age interval x 1,000) for each province are;
1. very different from what is reported in the report(## per 1,000)
2. alongside being different the observed difference in rates is also not proportionate to what is reported in the report e.g. Central Province (No. 2) is 51 per 1,000 in the report but increases to 79 when using the above ltable, compared to Nyanza province (No.5) which is 151 per 1,000 in the report but reduces to 126 when using the above ltable method. It something to do with the way DHS calculate the stratum specific (e.g. province) rates?
I would be most grateful if you could provide any further insight into this apparent discrepancy.
Kind Regards,
Rob
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #10905 is a reply to message #10880] |
Mon, 03 October 2016 14:07 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
A response from senior technical specialist, Dr. Tom Pullum:
Quote:
The table you are referring to (Table 8.2 on page 107 of the report on the Kenya 2008-09 survey) is correct. I just re-calculated the numbers. However, the following observations may be relevant.
First, did you notice in the title of the table that it refers to the TEN years before the survey?
Second, the table itself does not include an estimate for the total (that is, all Kenya) for the ten years before the survey. I calculate that number to be 84 (83.72). That number is not mentioned anywhere, in the text or a table or a figure.
Third, I'm sorry to see that Figure 8.2, beneath Table 8.2, has an error in it. For all Kenya it gives an estimate of 74. That's obviously wrong because it is not in between the urban and rural estimates of 74 and 86, respectively. The value of 74 is for the FIVE years before the survey, given in table 8.1. The correct value, should be 84 (as I said). I don't know whether you are trying to match the 74 for the five years before the survey or the 84 for the ten years before the survey.
Fourth, and this will affect your calculations, the DHS estimate for the five years before the survey is not based just on children born in the past five years. It is based on any exposure to any of the eight component age intervals in the past five years. Similarly for the past ten years.
I can't tell what it is that you want to do. Do you want under-five rates for some covariates that are not in table 8.2? Do you want to repeat table 8.2 for the past five years, rather than the past ten years? Or something else? I would like to be able to help but I don't really know what you are trying to do.
|
|
|
|
| Re: Calculating Under 5 Mortality Rate [message #10944 is a reply to message #10905] |
Mon, 10 October 2016 20:54 |
Olutosin
Messages: 5
Registered: October 2016
Location: Nigeria
|
Member |
|
|
hello please sir i want to use NDHS 2013 dataset for my project work. i am to carry out survival analyses ie( the risk of death between 0-11 months and 12-59 month) on infant and child mortality using NDHS2013 data set, as well as cox proportional hazard analyses in my work, i want to use environmental factors like toilet facility,source of drinking water,uses of solid fuel, wealth index,sex,region,residence etc as my independent variable on infant an child mortality. how can i perform the analyses using SPSS and STATA help me out please...........
Olutosin
|
|
|
|
|
|
|
|
|
|
Goto Forum:
Current Time: Mon Dec 2 17:59:10 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
