If hhkids=fekids=1 then female=1. If hhkids=1 amd fekids=0 then female=0. The female variable is redundant.
Given the specification the correct interpretation is female with kids is impacted more than people without kids (regardless of gender).
I would specify model with female, fekids and makids variables. This way you can compare the last two