Multi-causal Relationships in their Socio-political Context

Identification of causal relationships has a lot to do with control. Understanding which phenomenon determines another increases the ability of human beings' to cope with their environment. At present, scientists and philosophers have difficulties in specifying when a relationship between two events is causal (Pearl 2000). In epidemiology, different models of causality exist, which Vineis and Kriebel (2006) divide into two classes. The first is characterized by a linear monocausal pattern of explanation, based on the concept of necessary cause (that is, the disease does not develop in the absence of exposure to the agent). The second is characterized by the concept of causal web (that, a concurrence of different conditions is required to induce disease). A widely used multicausal model is Rothman's 'pie' mode, in which a sufficient causal complex (a pie) is represented by a combination of several component causes (Rothman and Greenland 2005). The disease appears when the pie is completed. Such multi-causality provides the opportunity for removing only one or a small number of factors for preventing harm (Gee 2003). Other causality models have been developed in epidemiology but they have many limitations, mainly related to the low degree of accounting for interactions between causal factors, to their dynamic character over time, and to the differences between individual and population levels (Valleron 2000; Vineis and Kriebel 2006). The usual scientific approach to untangling such complex cause–effect relationships is to isolate one possible causal factor by statistically accounting for all others (EEA 2005).