We propose dynamic systems models as one component of the epidemiologic toolbox. Systems models reflect the fact that diseases are caused within complex molecular, biological, and social systems, with positive and negative feedback. Such models predict empiric observations, provide a framework for clarifying what new data is needed, allow for complex interactions between variables at levels from the subcellular to the community, and incorporate known feedbacks between systems elements at these various levels. In all of these ways, they have the capability to advance the science of epidemiology.