I have wondered about this question for some time.
There are many jobs where there is just not enough resources to do the job properly as intended, so people do what is reasonable within the resources they are alocated at the time. This may be increased from time to time but will always lag behind what is considered necessary. We therefore use such techniques as risk analysis to make sure we do the important high risk jobs first and work down the list in the time available. If we can show that what we had done within the resources available was REASONABLE then I believe we would have a good case.
As far as going away from the manufacturers recommedations I think we would be on very doggy ground unless we could show that it was either common practice within the industry or that we had collected data over so many years and could show from it that this was a sensible and reasonable course of action.
Just aome thoughts to think about!!