Ghost-time bias from imperfect mortality ascertainment in aging cohorts

Eric J. Jacobs, Christina C. Newton, Ying Wang, Peter T. Campbell, W. Dana Flanders, Susan M. Gapstur

Research output: Contribution to journalArticlepeer-review

5 Scopus citations

Abstract

Purpose: Many cohort studies in the United States link with the National Death Index to detect deaths. Although linkage with National Death Index is relatively sensitive, some participant deaths will be missed. These participants continue to contribute person-time to the data set after their death, resulting in bias, which we refer to as ghost-time bias. We sought to evaluate the influence of ghost-time bias on mortality relative risk (RR) estimates. Methods: Simulations were performed to determine the magnitude of ghost-time bias under a variety of plausible conditions. Results: Our simulations demonstrate that ghost-time bias can be substantial, particularly among the elderly, where it can reverse the direction of the RR. For example, we conducted a simulation of a cohort of men beginning follow-up at age of 70 years, assuming 5% missed deaths and a true RR of 2.0. In this simulation, observed RRs were 1.89 during the year the cohort was aged 85 years, 1.60 during the year the cohort was aged 90 years, and 0.61 during the year the cohort was aged 95 years. We also provide results from actual cohort data that are consistent with ghost-time bias. Conclusions: Ghost-time bias may meaningfully affect mortality RR estimates under conditions that can plausibly occur in aging cohorts.

Original languageEnglish (US)
Pages (from-to)691-696.e3
JournalAnnals of Epidemiology
Volume28
Issue number10
DOIs
StatePublished - Oct 2018
Externally publishedYes

Keywords

  • Aging
  • Bias
  • Cohort studies
  • Epidemiologic methods
  • Mortality

ASJC Scopus subject areas

  • Epidemiology

Fingerprint

Dive into the research topics of 'Ghost-time bias from imperfect mortality ascertainment in aging cohorts'. Together they form a unique fingerprint.

Cite this