The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Mortality » What are Steps and Syntax to Calculate Under Five Mortality using SAS
What are Steps and Syntax to Calculate Under Five Mortality using SAS [message #12869] Tue, 25 July 2017 14:19 Go to previous message
Shams16 is currently offline  Shams16
Messages: 4
Registered: July 2017
Member
Background information for question

I am very new to DHS data. I am currently working on a project where I want to calculate U5 mortality for the last five years, and for the last 10 years. I read the DHS Guide to Statistics (Page 93) plus the codebook. Guide explains how to calculate mortality, I believe it is based on Ruststein's methods.

It seemed easy at first to do it in SAS. But when I started I was not successful. I need to understand the steps to do this calculation in SAS. What I did not understand in the Guide to Statistics (201) is the Steps C & D. It is difficult further as the guide does not give any numerical example to put the theory in perspective.

What I have done so far:

    I have consulted several sources to find a paper with example - No success

    I know most of the variables which will be required for calculation and have located them in AFBr70FL file. Note: I am using 2015 DHS data for Afghanistan.

      I found a very good macro crafted by some researchers at Harvard - if works perfectly - I can use it, but 1) the marco is very complected I don't understand what it is doing, 2) I don't want to use these codes if I do not understand the concept since I wont be able to defend my findings.

        I found few resources for STATA and SPSS - but I was not able to replicate them in SAS

        What I need?
        I need the steps in SAS that I have to follow to achieve this Job. Plus if you could share your codes (not macro). Or guide me with each step using the DHS variables and SAS syntax, I will highly appreciate.
        Theoretically, I know I have to create a restrict subset of data to certain time period, then create 0, 12, 35, 611, 1223, 2435, 3647, and 4859 months groups, then dig into each group and identify deaths and alive and depending on their contribution to either (cohort A, B, C) count them 1/2 or full numerator and denominators.(Here my brain freezes: how to identify who to count full and who to count half).
        I also know that simple SAS procedures such as PROC FREQ, and PROC MEANS or PROC SURVEYFREQ should do the job!!! but I need a guidance and assistance. Thank you in advance.

        FYI: there are not much of resources for DHS analysis in SAS - Any idea why? I don't know STATA byond basics!

        Regards
        Shams



         
        Read Message
        Read Message
        Previous Topic: How are tl and tu defined in DHD reports (Stat manual p93)
        Next Topic: Pregnancy History
        Goto Forum:
          


        Current Time: Tue Dec 3 15:21:21 Coordinated Universal Time 2024