As I understand, one of the factors in this accident was that the problem was recognized late, with very little time to react. Would it be possible (reasonable?) to change the procedures so that this kind of problem can be found earlier, for example by avoiding "long" periods at reduced thrust?