Hi Nile,
Hopefully I can better clarify your question.

1) The dataset BFPR70 is the "person" recode meaning that it is information about all household members. So you are correct that 39,529 is the total number of people surveyed from the 6,500 households
2) To answer your second question, if you are wanting to look at ITN usage by the entire population you are correct in using the BFPR70 file.

If this is what you are interested in you should use the indicator "Proportion of the population who slept under an ITN the previous night".
Numerator: Number of individuals who slept under an ITN the previous night
Denominator: Total number of individuals who spent the previous night in surveyed households

When examining the denominator closely you will see that the denominator is limited to individuals who spent the previous night in surveyed households. This is because the Person Recode (PR) file contains information about everyone recorded in the household schedule. So this is anyone who usually lives in the household (de jure) OR stayed there the night before (de facto). However, we only have ITN utilization data on those who stayed in the household the night before the survey (de facto). When we calculate this indicator we are going to have to make sure we restrict the PR data to only those who slept in the household the previous night. (variable hv103= slept last night)

To calculate this in Stata you could use the following code

use " BFPR70FL.DTA"
g wgt = hv005/1000000
g itn=(hml12==1|hml12==2)
tab itn if hv103==1 [iweight=wgt]

If run correctly you should see that 67% (n=37,340) of the de facto population slept under an ITN the previous night in Burkina Faso (2014 MIS).

You will see that another user had a very similar question using the Nigeria 2013 dataset