Have done a bit of experimentation with this. We have some customers operating regularly at 7000' pressure altitude at plus 35 deg C, at fairly high AUM who have to depart from helipads with a fence round them. During recurrent training in the simulator, they wanted to see what was the best plan for a departure with possible engine failure i mind, given that they were over the WAT limits for Cat A.
Whatever profile they flew, they were going to crash, but we managed to find a profile which was a cross betwen the helipad and the short field concept that genrally allowed a more controlled and survivable crash than any of the documented procedures.
It took a bit of playing with - you cant necessarily assume any technique is going to work until you try it with a real failure! Generally, something close to a Cat A profile seems to give you minimum hurt if you have the failure early in the departure, but you have the option of weighing this up against reducing your exposure time, and hoping the failure will come later if it happens.
I reckon the best plan is to get in a simulator and do some experimentation at your typical weights and DA.