| Re: Maternal Mortality Ratio [message #25183 is a reply to message #25146] |
Fri, 09 September 2022 16:08  |
Janet-DHS
Messages: 893
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member Tom Pullum:
You didn't give a link, but the trends report you are referring to is here: https:// www.unfpa.org/featured-publication/trends-maternal-mortality -2000-2017. It happens that I am a member of the Technical Advisory Group for the UN inter-agency group that produces such reports, as well as being on the DHS staff. I'm very familiar with the differences between the estimates. The DHS estimate comes from just the 2018 survey and describes an interval, the 7 years before the 2018 survey. The calculation is complex but it doesn't involve any adjustments, or reconciliation with earlier surveys, etc.
The UN uses many data sources, including the 2018 DHS survey, but also including earlier DHS surveys and other sources. It models a trend line that actually extends back before 2000. It gets some leverage from trends in covariates such as the percentage of births with skilled birth attendants. A big difference is that the UN estimate for 2017 actually refers to a point in time, July 1, 2017, rather than an interval. It therefore has a more recent time reference than the midpoint of the 7-year interval before the 2018 survey.
Another big difference is that the key piece of information from the DHS survey that the UN uses is the proportion of deaths (to sisters) that are pregnancy-related. The UN estimates then get the necessary all-cause death rates (needed for the MM Rate) and fertility rates (needed for the MM Ratio) from the latest version of World Population Prospects, which is prepared by the UN Population Division. Those estimates use based in part on DHS survey data but also use other sources and are also modeled.
In other words, the UN estimates involve many data sources, including DHS surveys, and are adjusted in many ways. If I had to choose, I would definitely prefer the UN estimates.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
