Epidemiology

Katherine Keyes, PhD ‘10 Selected by the APHA's Epidemiology Section to Receive the Student Award at 2009 Annual Meeting in Philadelphia

The award-winning research by Katherine Keyes, PhD ‘10 focuses on the estimation of age, period, and birth cohort effects. Age-period-cohort (APC) analysis has long been important in scientific inquiry, providing critical information concerning the etiology of tuberculosis, peptic ulcer, and alcohol dependence.  Early APC analysts relied on broad conceptual population-health frameworks and strong tests of creative hypotheses to inform analytic choices. More recently, APC analysis has been dominated by the development of statistical approaches to partition variance, absent the conceptual rigor of the APC pioneers.

The overarching aim of Katherine’s research program is to reinvigorate and extend the utility of age-period-cohort analyses for answering substantive etiologic questions about common health outcomes by examining the shifting of prevalence over time. To achieve these research goals, Katherine has been mentored by Dr. Guohua Li in the Department of Epidemiology, developing methodological tools to analyze data collected over time, and applying their methodologies to diverse health outcomes such as homicide, suicide, adolescent substance use, breast cancer, and obesity.

Award-Winning Project

For this year’s APHA meeting, Katherine reanalyzed data from a classic study by epidemiology pioneer Wade Hampton Frost. This year marks the 70th anniversary of the publication of Frost’s classic paper, providing an excellent opportunity to both reflect on the importance of Frost’s work and to highlight the methodological advances being made 70 year later. Frost’s study of tuberculosis mortality in Massachusetts, published in 1939, was one of the first demonstrations of a strong cohort effect in disease mortality.

Frost’s findings stimulated decades of research into cohort effects for many health outcomes, and remains a classic work in age-period-cohort effect literature. Katherine reanalyzed the Frost data using a novel median polish approach to quantify the cohort effects, providing a statistical quantification of the decreasing rate of tuberculosis mortality across generations born after the turn of the century.