The article presents short historical review of the concepts of causality. Then it deals with contemporary concepts of causal dependence with special reference to epidemiological studies. In particular Lewis counterfactuals theory and contemporary probabilistic theories were analyzed including applications of DAG's (directed acyclic graphs), which in the last decade are frequently applied for presentation of complicated study designs in epidemiology. Authors high appreciation explanatory role of graphic presentation of relationships between variables and the role of conditional probability of events respecting Markov conditions and Bayesian premises, does not change his opinion, that statistical methods are insufficient for final assessment of causal dependence and some subjective element of learned judgment of the scientist has to be always present. In Authors opinion causal approach to associations between are crucial as a base for therapeutic approach and for public health interventions. This is why he is against consequent indeterministic approach.