I think the point is that without accurate data that reflects the detail (i.e. do trans women need to be their own separate category than cis women etc… are the risks the same? Are frequencies of certain diseases? etc…), anything that depends on actuarial data is going to inherently be inaccurate or non-representational for trans folks.
You could look at this like “for what nefarious purposes are they trying to single out trans people?”, or it could be looked at as “if there’s no info on trans people, assumptions (very likely incorrect ones) are being made, and that’s a kind of marginalization as well…”. Disease surveys and insurance pricing are just two things what come to mind as being very dependent on this kind of info.
