An Investigation into the Use of the Relative Standard Error as a Measure of Rate Stability

• Activity Number: 348
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): John Keighley*+ and Sue Min Lai
• Companies: Kansas University Medical Center and University of Kansas
• Keywords: direct age-adjusted rate ; relative standard error ; stability ; cancer incidence

Abstract

Cancer incidence is commonly reported using a direct age-adjusted rate. When rates are calculated for cases within the U.S. the current U.S. standard population is typically used as the reference population. When cancer incidence rates are published, the organizations that report cancer such as state cancer registries or National Center of Health Statistics (NCHS), must consider patient confidentially, rate stability, and the need for accurate information to formulate public policy. The relative standard error (RSE) is one measure of rate stability. It is defined as the standard error of a rate divided by the rate. For a crude rate the values of the RSE range from 0 to 1 with smaller being more stable. The range of values that have been published as cut points for stability, range from 0.20 to 0.30. The New York Cancer Registry has a discussion of the value of the RSE to be considered as stable on their website but the discussion was based on using a crude rate. This study will give a recommendation for a relative standard error cut point for a direct age-adjusted rate based on both a simulation study and data from the Surveillance, Epidemiology, and End Results Program (SEER).