| Linkage by administrative region x type of health facility [message #26455] |
Wed, 22 March 2023 09:46 |
Aline Semaan
Messages: 3
Registered: March 2023
|
Member |
|
|
I am looking to link DHS and SPA data, but the SPA survey was conducted among a sample of the facilities, meaning that the most appropriate linkage method to DHS data is at the sample domain i.e. the region (as mentioned in DHS Spatial Analysis Reports 10).
The SPA facilities sampling strategy was stratified by Region and Type of health facility (e.g. Gov hospital, gov health center, private hospital, private health center, etc.).
The indicator I will assess in DHS is duration of stay in the health facility after birth. This inherently limits the DHS sample to women who had their most recent livebirth in a health facility. We have information in the DHS about the type of health facility where the women gave birth. I plan to conduct a multilevel analysis to identify health-system level factors (e.g. number of maternity beds per 1000 pregnant women) influencing length of stay at the regional level.
My question is, instead of linking the two surveys by region alone, is it technically correct to link the surveys by region and facility type? Would this interact with the representativeness and sampling weights of the DHS data? And what are methodological considerations in this case?
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
