The collateral damage would be considerable. How many victims of partner violence tracked to their new home and attacked would be too many to make it "worth it"

The collateral damage would be considerable. How many victims of partner violence tracked to their new home and attacked would be too many to make it "worth it"? And why should a security researcher be the one to make such a decision?