Learning from Cancer Incidence: Secular Trend, Lead Time, and Overdiagnosis in Prostate Cancer Screening

• Activity Number: 496
• Type: Topic Contributed
• Date/Time: Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Ruth Etzioni*+ and Donatello Telesca
• Companies: Fred Hutchinson Cancer Research Center and University of Washington
• Keywords: lead time ; maximum likelihood ; surveillance data ; PSA screening

Abstract

Introduction of a cancer screening test leads to a rise in disease incidence followed by a decline. The size and duration of the rise inform about the lead time associated with the test. The lead time is critical in predicting overdiagnosis, which occurs when screening detects disease that would not otherwise have been diagnosed. Previously, we used simulation to make informal inferences about lead time and overdiagnosis in prostate cancer due to PSA screening. In this paper, we develop a likelihood approach that uses surveillance data to formally estimate the lead time due to PSA screening. A key model input is the secular trend in disease incidence, which is the incidence that would be have been expected in the absence of PSA. We use the model to estimate lead time given a projected secular trend, but also to estimate age-specific secular trends given a range of lead times.