Thank you for your question and your code. You are correct in that the Tchad data does require an extra restriction for this indicator. In this survey mosquito net questions were only asked to households that were not selected for the men's questionnaire. So you need to add the restriction hv027==0 to your tabulation. You can read more about the sampling (ÉCHANTILLONNAGE) on page 7 and 8 in the final report.

I tried to replicate your code on other surveys and I am not matching denominators. I would suggest you re-examine your code for other surveys since you should be using the PR file to replicate this indicator. Here is the code you can modify for TChad 2014-15

use "TDPR71FL.DTA", clear
gen itn=(hml12==1|hml12==2)
gen wt= hv005/1000000
tab itn if hv103==1 & hv104==2 & hml18==1 & hml16>=15 & hml16<=49 & hv027==0 [iweight=wt]