Adverse drug reactions and avalanches: Life at the edge of chaos

Abstract

Although many reports have described the incidence of adverse drug reactions, none have explained their variable severity or why they happen. Because human physiology shares many of the features of other complex adaptive systems, reactions to drug therapy were examined mathematically for specific patterns to show (1) that the severity of adverse drug reactions follows a distribution seen in other complex adaptive systems, called a power law distribution, and (2) that preventable reactions occurred for reasons fundamentally different from those that underlie the nonpreventable reactions. Two reports detailing adverse drug reaction incidence and severity were evaluated: a meta-analysis of prospective adverse drug reaction studies and a prospective cohort study. Incidence of drug reaction was plotted as a function of severity and fit to an equation. The incidences of overall and nonpreventable drug reaction, plotted as a function of severity, followed a similar power law distribution regardless of sample size or the nature of the population or drugs studied. An exception to this was the preventable reactions, which were described by a different type of equation. Response to pharmacotherapy exhibits many properties of systems with self-organized criticality. An exception to this is the preventable reactions, which seem to be fundamentally different from the nonpreventable ones. These observations suggest that the presence and the distribution of severity of reaction to pharmacotherapy is a consequence of our adaptation as biological systems, and although adverse reactions can be made less frequent, a certain percentage will not be preventable. ©2005 the American College of Clinical Pharmacology.