Last update:


Arch Hellen Med, 21(2), March-April 2004, 112-122


The concept of cause in epidemiology

Laboratory of Clinical Epidemiology, School of Nursing, University of Athens, Athens, Greece

Event causation has been the primary focus of scientific and epistemological discussion in contemporary period, and its analysis has been controversial. The following four approaches have been prominent: regularity analysis, counterfactual analysis, manipulation analysis, and probabilistic analysis. According to a popular form of nomological analysis, the cause of a phenomenon is the events that are described in the set of antecedent conditions of a correct causal explanation of this phenomenon (semantic concept of causation). The semantic concept of causation is the opposite of the concept that is used in everyday life (pragmatistic concept of causation). The obscurity of the term “cause” in everyday life could be met only in the context of a rational causal model. Mackie offered a methodical analysis of the pragmatistic concept of causation. According to his analysis, cause of an event (P) is usually another event (A) that fulfills the “INUS condition”. That is, A is an insufficient but necessary component of a minimal condition, which is unnecessary but sufficient for P. Mackie makes it clear that a cause is not always an INUS condition. A could be the cause of P also in the situations: (a) when A is a component of the sole minimal sufficient condition, (b) when A is a minimal sufficient condition by itself, (c) when A is the sole minimal sufficient condition. The term “at least INUS condition” is used to include all cases in which A is a cause of P. Rothman introduced a pragmatistic approach to the concept of causation in the Health Sciences as the model of “causal pie”. This description has many similarities with the model of “at least INUS condition”. According to this model, cause is an event that alone or in conjunction with other events causes a disease. The minimal set of components that inevitably causes the disease is the sufficient cause and each component is a component cause. The model of “at least INUS condition” could be used to fill out the model of “causal pie”. Thus, the second model could be considered as a complete model of causation and it could offer more profound understanding of the concept of cause in the Health Sciences.

Key words: “Causal pie” model, Component cause, INUS condition, Necessary cause, Necessary condition, Sufficient cause, Sufficient condition.

© Archives of Hellenic Medicine