In all fairness, cohort studies are rare to begin with, due to the massive requirement of time, money, and effort to track the individuals over time. No to mention how they're mentally unstable individuals who are prone to do as they wish (as you've mentioned). There's then also the factor that
@Caesar mentioned, as there's no current golden standard for this kind of procedure. It's a mess.
I mean, okay, let's assume they're not such a mentally unstable individuals who like to do as they wish.
We need to group them based on the technique used, which is already a pain. Even if we were to ignore that and categorize based on surgeries/procedures, there are a lot of variables to factor in. Are they under other kind of medication? Comorbidity? How long did they take HRT before the surgery? Have they taken another surgery, as well?
It's a pain. Even if we pick a very specific group in order to reduce as much bias as possible, and reduce as many variables as possible which are uncertain regarding their relationship with the outcome, can we even obtain the data? Let's be honest, we all know it's unlikely they'd be willing to do that. As for lost to follow up cases, I guess we can just treat them as censors and analyze it with survival analysis techniques. But then again, how are we supposed to define the event, in this case...
Oh well, we all know it's just cash farm for butchers and the medical industry, anyway. The whole gaining data thing is just me trying to look at it on the positive side.