|
|
Re: survival analysis of neonatal mortality [message #30088 is a reply to message #30079] |
Sat, 21 September 2024 02:29 |
kidus28sewagegn@gmail.com
Messages: 3 Registered: September 2024
|
Member |
|
|
Dear janat-dhs
I am asking for help with the DHS dataset for Sub-Saharan Africa, specifically for the period from 2010 to 2022.
In the course of my analysis, I discovered that the datasets from 2015 to recent years have a 'century day code', also known as CDC, but for the years 2010-2015, the datasets only provide a 'century month code', also known as CMC.
Although the 'century day code' provides more precise information to calculate the age in days, I am having difficulty deriving the age in days for the period 2010-2015 due to the absence of this precise information.
For instance, when using the women's dataset from the 2011 Ethiopian DHS (ETIR61FL.DTA), I encountered the same problem. The 'CMC' variable only provides information on the century and month (0-12), which restricts the calculation of age to an approximate number of months.
I kindly seek your advice or suggestion on possible ways to work around this limitation, as using alternative Sub-Saharan African DHS datasets yielded a similar outcome.
Thank you for taking the time to read and respond to my query.
Best regards,
kidus
|
|
|
Re: survival analysis of neonatal mortality [message #30118 is a reply to message #30088] |
Fri, 27 September 2024 11:32 |
Janet-DHS
Messages: 878 Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
The century day code only appears for b18, in the KR and BR files, for surveys for which day of birth was collected in the birth history. Day of birth (within the month of birth) appears as b17. The cdc code is elapsed days since some reference date, which I believe is January 1, 1960 for the algorithm DHS uses.
If you are studying neonatal mortality, I don't think you need to use that code at all. You just need age at death, in days, for deaths in the first 28 days after birth. That variable is b6, for codes in which the first digit is 1 (signifying that age at death is given in days). Days 1-28 corresponds with completed days 0-27, given by b6 values from 100 through 127.
|
|
|