A two-stage protocol for verifying vital status in large historical cohort studies

When access to the Social Security Administration's Master Death Claim File was restricted in the mid-1980s, researchers were left with no time- and cost-effective protocol for verifying the vital status of large historical cohorts. A two-stage tracing protocol was designed to overcome this restriction. Stage I relies on national-scale sources to focus on the complete and accurate identification of deaths among persons unconfirmed as alive and assumes that persons not identified as deceased are alive. Stage II tests the "alive" assumption by extensively tracing a random sample of cohort members with unconfirmed vital status. Stage II provides unbiased estimates of the proportion of deaths among the assumed "alives" in the cohort (misclassification rate) and the proportion of persons untraceable in the total cohort. This paper describes our two-stage protocol and an application to a large, ongoing occupational cohort study.