Lots of the oddball prisoners housed in male prisons would indeed be safer if housed in women’s prisons. As you say.
But the other part of that issue is what risks to the existing women in the women’s prisons would be imported into the women’s prisons along with those now-safer inmates formerly housed in the male prison.
Any valid approach to harm reduction would need to look at both. Using statistics, not doctrine, as their guiding principle.
I think @Miller nailed it. This whole topic is yet another RW freakout