Defense of epidemiologic research


a stock image of an illustration of 2 researches looking at a large computer.

A recent article published in Science Advances demonstrates the rigor and substantial advancements that have occurred in the field of epidemiology over the past 27 years. Led by Rollins researchers—including 33 Emory PhD students in epidemiology—the article revisited an influential article authored by journalist Gary Taubes that questioned the utility of nonrandomized epidemiologic research.

The new article, led by Lauren E. McCullough, associate professor of epidemiology, and Maret L. Maliniak 14MPH 24G, systematically evaluated the current evidence for 53 examples of associations Taubes claimed were doubtful in his 1995 Science article.

The researchers found that 25 percent of the associations Taubes deemed questionable are now widely viewed as causal, resting on a substantial body of epidemiologic research and tremendous innovation in the field during the past 27 years.

Among Taubes’ list of questionable associations that are now viewed as causal, and with public health policies to address them, are the links between alcohol consumption and breast cancer, residential radon exposure and lung cancer, and the use of tanning devices and melanoma.

For 27 years, Taubes’ paper has been cited more than 1,000 times and has been used to cast doubt on the value of epidemiologic research. This new article—to the authors’ knowledge—is the first to address all of Taubes’ example assertions head-on and serves to dismantle Taubes’ argument while verifying the rigor and reproducibility of epidemiologic research results.—Kelly Jordan