Very tricky I'd have thought. Even assuming you could cancel out all the variable factors that exist in front-line medical treatment, you'd have to have two cohorts of doctors, one with the 'live subject' training and one without and compare outcomes for specific classes of injuries. However there are obvious ethical issues there. Plus the sample size (i.e. the number of injured with any specific type of wound) is likely to be quite small because despite the timbre of the news coverage, UK losses in Afghanistan aren't actually that great - they're terrible for the people involved and their families, obviously, but they are a small fraction of the number of people who get killed in road accidents over the same period for example.
In the end, the only evidence you're likely to get is the assessment of the doctors in question - do they feel their skills have been improved by the live subject training they did? Not exactly scientific, even given it would be sincere.
In the end I think the argument is similar to the use of animals in medical science - you can not really care at all, you can have a stance that it's acceptable if the benefit is great enough and the suffering inflicted on the animal is minimised as much as possible, or you can just say it's unacceptable to do regardless of the benefit. This is a moral judgement which is personal.
I do still have issues with it though on the simple basis that I looked at it and thought to myself 'If we were told someone else's army was doing that, there's a fair portion of the population that would be outraged'.
Jon