Artificial adaptive systems and fuzzy measures for translation of clinical trial results to the bedside.

Medical science relies uniquely on statistical evidence from large clinical trials or laboratory experiments to deal with uncertainties regarding clinical decisions. The statistical evidence is stated in probabilities. Probability theory is based on the logical rules set forth by Aristotle: the law of noncontradiction, excluded middle, and identity. Thus, medical science assumes that variables such as disease and treatment relate or interact only in certain ways and the statistical evidence defines the presence of these relations. Experiments seek to find these certain predefined relations among variables, relations that then predict in terms of probabilities their behavior over time. Living systems are nonlinear and complex, which means that the relations among variables not only change over time, but their behavior over time changes in response to the change in one another in unpredictable ways. A patient is a living system and as such is a dynamic system of the type described. Because patients have different genetic and contextual makeup, there is no reason to assume they share a common dynamic. One goal of medical science should be to discover the behavioral dynamic of variables of interest in a given patient to individualize treatment appropriate to that system. Ideally this would be in real time so that diagnostic and treatment decisions adapt to dynamic change. Artificial adaptive systems on the one hand and fuzzy theory on the other are two scientific approaches toward this end.