Shawn,
There is no substitute for training on the helideck itself - training to a square painted on a runway provides a lateral visual cueing environment that is just unrealistic (hence my earlier remark about testing for Cat A procedures being totally unrelated to the operational environment).
Training normally consists of normal approaches and departures and, except in controlled conditions, does not include engine failures. The fact is, the norm (for twins) in offshore flying is Performance Class 2 - in most cases with exposure (for the benefit of Geoff, that does not apply to the AW 139). Ipso facto whilst the profiles to minimise the exposure can be rehearsed, engines should not be failed (except as stated above, under strict and controlled conditions - i.e. low take-off mass).
The exposure regime is changing as the AW139 enters service but expect the power reserves of even that machine to erode when it moves from 6.4t to 6.8t. At that stage it is likely, along with other helicopters, to be operating in enhanced PC2 (PC2e); when operating to PC2e, and with the correct rig, with a clear obstacle sector, and the wind in a favourable direction, limited engine failure training could be carried out.
The real answer is training in the simulator but, as Geoff will confirm and I experienced when observing the testing of PC2e procedures, the fidelity is still not as good as it might be and therefore (and sorry for the pun) somewhat artificial. Even in the simulator, introducing engine failures when operating with exposure will only lead to a crash (the very definition of exposure indicates that any failure during the critical time is likely to lead to a catastrophic outcome).
We are not yet in possession of the facts with regard to this accident; but even when we are, it is unlikely that Category A procedures were being practiced because, under operational conditions, it is unlikely that they would have been in use.
I share your sentiments about conducting a check flight to/from a hospital site - slightly more risk than was absolutely necessary under the circumstances. However, as is the case with all retrospective analyses, the problem and solution are obvious to all.
Where skill is shown is when risk assessment for this type of operation (including training) takes into account the the hazard and probability of release and avoids that for which the consequences are most severe. To put this into the context of this discussion, such a risk assessment might indicate that whilst training might be acceptable to a helideck (with the presence of the deck crew and emergency personnel/equipment) it might not be to a hospital helipad. On the other hand, simulator training might be the best option when abnormal procedures are taking place.
Jim