After many incidents or accidents of the past years across high risk industries (transport, chemical, nuclear) , it appears that, when reading investigations reports, changes always play an important part in the genesis of these events (i.e. our own investigations do show this very well too, Le Coze, 2010). Yet the problem of change is quite complicated because it includes a wide range of issues, better identified in favour of hindsight. It is indeed much more difficult to predict beforehand these changes that will combine together to lead to a specific unexpected accident scenario. Several problems can be related to this difficulty, problems that one finds when dealing with any complex systems linking technological, psychological, cognitive, social, cultural, economical and political dimensions together (Le Coze, 2008a). One of them is that changes are permanent. Changes happen all the time. But their effects are not always immediate. They might need for instance to be associated with other changes to generate an adverse outcome. Another problem is that there is not necessarily proportionality between the magnitude of a change and its effects. A last one that is mentioned here is that one change somewhere might have some effects somewhere else.