The question of whether a selected psychological pattern precedes or follows the onset of coronary heart disease has served as a vehicle for a discussion of the fundamental issue, namely, the ability to draw causal inferences from observed associations in cross-sectional studies. The possibility of an independent contribution of the psychological pattern to the development of coronary disease, hence a possible etiological role, has been clarified and its implications have been illustrated. The crucial distinctions between "predictive" and "etiologic" stochastic models have been reviewed. When such models are used to predict phenomena, the validity of their underlying assumptions is not mandatory. However, when stochastic models are used to explore etiologic relationships, their value depends on the validity of the underlying assumptions. In addition, the users of formal mathematical models must exercise caution to avoid the pitfall of excluding relevant knowledge available from epidemiologic observation. To do so may prevent selection of the most likely inferences to explain particular causal mechanisms.
ASJC Scopus subject areas