Cohort follow-up: the 21st century procedures

The basic logic of designing an occupational cohort study has changed little since William R. Gaffey outlined the issues of follow-up, measurement of exposure, and analysis of data. However, many new avenues of tracking workers for epidemiological studies have been developed since Gaffey wrote his paper in 1973. Many disease registries also perform follow-up of subjects for vital status determination, so the procedures used with this process are common to the two applications. This article speaks to cohort construction for this occupational research as well as describes the 2007 methods for vital status follow-up. Rises in concern about work-related disease risks and the scientific resources for performing these studies coincided with the computer revolution. Government and private sources of data on vital status have changed in several ways over the 35 years since Gaffey's seminal paper. Some systems make the process of follow-up more rapid and productive, and some barriers have been imposed as societal concerns for privacy have risen. We describe the process of linking 5 sources of data to compile a roster of 6,820 workers employed at the Paducah Gaseous Diffusion Plant from 1953 to 2003. The record linkage processes achieved a final death cohort of 1672 deaths--the ascertainment of these deaths (by time period) was 1379 (1979-2003) and 293 (1953-1978); follow-up then was 100% for this cohort.