Modeling the epidemiologic individual

Abstract

Modern epidemiological methods often elide the distinction between individuals and populations in practice. Health data and outcomes gathered from a population can be, and often are, applied to a specific person, guiding preventive, diagnostic, and therapeutic interventions. This article looks at a key site for the origin of this elision, the Framingham Heart Study, and shows how a novel methodological 'calculator' for individual risk of future disease emerged from what was originally designed as a community-based epidemiological study. The article explains how the methodological transformation of epidemiology and biostatistics was surprisingly driven by methods emerging from outside of traditionally trained epidemiologists, particularly through statisticians trained in non-medical areas of the human sciences, including economics, sociology, and demography. It therefore also explains how and why epidemiologists became far more statistically sophisticated and the field more dependent on statistical methods by the 1970s than they had been in the 1940s.

Keywords: Framingham; Jerome Cornfield; epidemiology; risk; statistics.

Dundee Coronary Risk-Disk, circa 1993. Source: Science Museum Group item 1993-171 (Science Museum Group Collection Online: https://collection.sciencemuseumgroup.org.uk/objects/co94356/dundee-coronary-risk-disk-dundee-scotland-1993-diagnosis-kit-cardiovascular).