Let's hear it for
Eric Jacobs, PhD, senior scientific director, Epidemiology Research (pictured here)!
The American College of Epidemiology (ACE) named an article he was the lead author on as the best overall paper of the year published in the journal Annals of Epidemiology. Co-authors, as he was quick to point out, include ACS epidemiologists Christina Newton, MSPH; Ying Wang, PhD; Peter Campbell, PhD; and Susan Gapstur, PhD; and Dana Flanders, MD, an epidemiologist from Emory University.
The award is given each year to recognize and support high-quality epidemiologic research published in the Annals of Epidemiology, which carefully reports and addresses questions of substantial public health or methodological significance.
The article by Eric and others - "Ghost-time bias from imperfect mortality ascertainment in aging cohorts" - describes a potential problem in many research studies that ACS epidemiologists identified while analyzing data from our Cancer Prevention Study II (CPS-II). (CPS-II is a large epidemiologic study of over a million men and women, followed since 1982, that has resulted in hundreds of scientific publications about the causes of cancer.)
As explained by Eric, "In 2016, a team of ACS researchers was analyzing data from CPS-II to compare the influence of social isolation on death from cancer and other causes in black and white Americans. For the most part, social isolation was linked with higher death rates, as reported in a separate publication. However, the researchers noticed this was not true in a relatively small subgroup of participants, those who were aging into their late 90s or even 100s, a result they found suspicious. The researchers eventually identified the problem. While the vast majority of deaths in epidemiologic studies are correctly identified through computerized linkage with national death certificate data, a few participant deaths are inevitably missed, for example due to misreported Social Security numbers or the use of nicknames. These participants can live forever as “ghosts” in the researcher’s dataset and distort results among the very oldest participants in unexpected ways. The ghost-time bias paper presents simulations that show such distortion can happen under conditions that occur in many ageing study cohorts, and discusses ways for researchers to recognize and avoid it, thereby producing higher quality health research.”
Each year's award-winning paper is selected by a panel of judges from both the ACE Publications Committee and Annals of Epidemiology, underscoring the close affiliation of the College with the Journal.
Congratulations to Eric, Christina, Ying, Peter, and Susan!