Cancer Risk in the Environment: Public Policy Issues in Screening

• Activity Number: 486
• Type: Invited
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Karen Kafadar*+ and Philip Prorok
• Companies: Indiana University and National Cancer Institute
• Keywords: randomized screening trial ; model uncertainty ; lead time ; length biased sampling ; discrete event simulation

Abstract

Screening for diseases that may have environmental initiators or interactions is often viewed as a beneficial and successful strategy for identifying environmental risks and for reducing mortality. However, as with any clinical treatment or intervention, benefit cannot be assumed, and screening can entail both costs and harms, so screening as a "treatment" must undergo evaluation. An evaluation requires a definition of treatment benefit, the design of studies to measure that benefit with low bias and variance, and the development of methods for for estimating the potential benefit. Often, models on which evaluation is based and factors in the model are unobservable (e.g., time of introduction into the environment of pollutants). We discuss these issues in the context of cancer screening, including robustness of models and sources of bias and variability in outcomes of screening studies.