Paper short abstract:

Systems science was heralded as a possible solution to impasses that have riddled epidemiology throughout its history, but this did not really come through. We argue that deeper differences in the metatheoretical framework underpinning are responsible for that.

Paper long abstract:

Beginning in 2009, the NIH created an initiative, called "Institute of Systems Science and Health", based on the idea that "[s]ystems science methodologies provide a way to address complex problems, while taking into account the big picture and context of such problems." One important area of application is the field of epidemiology, central to public health, for establishing causality and evaluating interventions. Throughout its history epidemiology has been subject to an inner tension between, grossly speaking, two different approaches to research: either focusing on the proximal aspects of processes, which resulted in risk factor epidemiology; or a more comprehensive approach incorporating the context. This new methodological development was heralded as an opportunity for the "contextualists" to overcome the dominance of their rivals, fusioning the "social" with the "biologic" in a somewhat chimeric, undefined "biosocial" synthesis. With the passing years, however, this future is yet to materialize. We contend that such changes require more than the incorporation of new methods, but a real transformation in its worldview, in the meta-theoretical components of scientific thinking described by authors such as Fleck (though styles), Kuhn (paradigms) and, more recently, Hacking (styles of reasoning). Such meta-theoretical components comprise, among other things, an epistemology that is adequate to its methodological approach. Epidemiological thinking seems still to be connected to a deterministic framework which includes a conception of science marked by various forms of naive realism, which do not mesh well with the complex approach.