[Realist protocol for implementing integrated care pathways in France (waiver 51 LFSS 2018)]

The paradox of real world researches supposedly conducted in real life is that they did not succeeded in freeing themselves from the techniques of the randomised trials which they pretended to escape. In the successionist perspective of classical experimental or quasi-experimental methods, the cause always precedes the effect in a linear manner, and any interference that is likely to threaten the stability of this relationship must be neutralised by mobilising the appropriate statistical techniques. In complex systems where everything moves at the same time due to multiple interrelationships that make it impossible to construct a counterfactual, these elements are no longer considered as confounding factors that need to be controlled, but as decisive factors in the smooth running of the experiment. The protocol presented in this article proposes an alternative evaluative technique mobilising the teachings of critical realism, which seems to us to be the most appropriate for understanding what happens "in the black box" recording the events that occur between the implementation of the Article 51 experiments and the observed results. The role of the evaluator is to put the actors back at the heart of the change, since it is achieved (or not) according to their reactions and the contextual elements. This credible explanatory theory allows us to understand: how does it work? For whom does it work? Why do losers lose, winners win? and under what circumstances?