The Unknown Denominator Problem in Population Studies of Disease Frequency


To calculate incidence or prevalence of disease, at risk population size is typically conceptualized as a denominator. However, due to many reasons including individual migration, record availability, and population distinction, obtaining precise measurements are difficult or unknown. Using spatial ecological studies of injury incidence as an example, PISC Past Trainee Christopher Morrison, PhD, et al. identify and evaluate five approaches that researchers have used to address “unknown denominator problems”: ignoring, controlling for a proxy, approximating, controlling by study design, and measuring the population at risk. They present a case example and recommendations for selecting a solution given the data and the hypothesized relationship between an exposure of interest, a count outcome, and the population at risk.

Go to Article

Back to News