intrepid2566, page-27

Of course people with TBI can die, but it is very difficult to predict which ones will, such is the variability of the initial presentation. So for doctors in any one trauma unit, it is impossible to know whether, for instance, the lesser number of deaths in the last 12 months is due to statistical variation or to an input, e.g. NNZ2256. That is why a large randomised clinical trial is required, preferably multi centre, and controlled with standard of care. The truth of the matter will not be known until the trial is unblinded and analysed, and nobody here will know the results until 2015. Full stop. And if the effect is small, it still may be of value (the absolute effect of statins on mortality is small yet they are among the most widely-prescribed medications) as long as there is some cost efficacy to the medication's use.