Jonathan Samet, MD

BETHESDA, MD — As technology advances and the ability of researchers to collect and compute larger and larger amounts of data increases, the face of epidemiology is slowly but surely changing.

This new kind of epidemiology involves many scientists partnering to study vast amounts of data and tease out scientific conclusions that would have been impossible only 10 years ago. As more health care systems move to electronic health records, and existing systems like VA begin to fully leverage their EHRs to track health trends, this trend in epidemiology is likely to continue.

Speaking on the topic of “big epidemiology” last month on the campuses of NIH, Jonathan Samet, MD, a professor at the University of Southern California and newly appointed NCI advisory committee member, said he defined “big” not only in terms of study size but in the questions these studies seek to answer. From the health perspective, “big” can refer to the size of the disease burden, the population exposed, the costs to society and the future risk of disease. 

“And there are some big questions that society expects epidemiologists to answer,” Samet said. Those include: “Should I learn about my genes? What chemicals cause cancer? How is pollution damaging?” and “What happens to veterans as a result of military service?”

“Even the question of whether cell phones cause cancer — we should be able to generate some information that will provide a sufficiently suitable answer to this question on safety,” Samet said.

From Small to Big

“Not so long ago, scientists could answer big questions with small studies — looking at lung-cancer risk and smoking, and breast-cancer risk and so on. In those days, one investigator could start one study and finish it,” Samet said.

Epidemiology had its start in the middle of the 19^th century, with London scientists examining cholera transmission. In the early days, epidemiologists were focused on germs and micronutrients, Samet said. This changed mid-century after scientists recognized the dangers of epidemics. Then there was a shift to social epidemiology at the turn of the century, and now the science is moving on to new things.

“We’re now in an era of transition, which is driven in part by technology,” Samet said. “We have very large data sets now and the ability to do computation like never before.”

In his own work on air pollution, Samet began using a compilation of data sets collected by various organizations, rather than relying on just one. “We turned to national resources. We looked at the monitoring system maintained by the Environmental Protection Agency, and we looked at Medicare data to examine hospitalization and air pollution, and we looked at data from all available cities and not just one,” Samet said. “We were able to look at how mortality is impacted by pollutants in the air, based on data from 54 million people. That’s something we could not have done 10 or 15 years ago. My first air pollution study was in the 1980s, and examining one city was computationally challenging at the time."
Page 2