This is why you invert and otherwise "look at it from a different angle" - instead of per capita or otherwise, you calculate "percentage chance a woman will be X" in various countries. If 100% of murders in Vatican City are women from a relationship, but they haven't had one since 1819, a woman is "safer" there than someplace where only 10% of murders are women, but the population is much, much larger.
Either way it's statistically low, but murder is easy to track, other forms of abuse and "generally this place is shitty" are not.
Null's main point stands, however - there are large percentages of South Korean people (and women) who want to get the fuck out of Korea.
The other problem is with terms; what one country counts as "abuse" may only be physical lasting bodily harm, whereas another may include yelling, emotional, etc.
In general the only way to solve it is local and large families; then if an abusive situation develops and gets out of hand it gets "resolved" or someone disappears.