Characteristics of childhood cancer survivors predicted their successful tracing

J Clin Epidemiol. 2004 Sep;57(9):933-44. doi: 10.1016/j.jclinepi.2004.01.005.


Objective: The Childhood Cancer Survivor Study is a retrospective cohort study that was initiated to explore late effects of childhood cancer and its therapies. We evaluated the characteristics of those requiring tracing and factors that influenced tracing success.

Study design and setting: Medical record review identified 20,051 eligible individuals from 25 institutions in the United States and Canada. Of these, 13,021 had a current address in the medical record at the treating institution, and 7,030 had an incorrect address and required tracing by a commercial firm.

Results: Tracing was successful for 4,188 persons (60%). Younger age at contact, shorter length of time since last contact, having a middle initial available, an uncommon last name, and socioeconomic factors were found to predict successful tracing. Compared to those with a current address available in medical records, subjects successfully traced were less likely to have accessed health care during the previous 2 years; and more likely to be current smokers, obese, and to report moderate to severe impairments (pain, functional status, and activity).

Conclusion: These findings provide an empirical basis concerning determinants and predictors of tracing success. If tracing had not been performed in this cohort, spurious associations may have been obtained for some health outcomes of interest.

Publication types

  • Multicenter Study
  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Adolescent
  • Adult
  • Bias
  • Canada / epidemiology
  • Child
  • Child, Preschool
  • Comorbidity
  • Epidemiologic Methods
  • Female
  • Health Status
  • Humans
  • Infant
  • Infant, Newborn
  • Male
  • Medical Records / standards*
  • Neoplasms / epidemiology
  • Neoplasms / mortality
  • Neoplasms / therapy*
  • Socioeconomic Factors
  • Survivors / statistics & numerical data*
  • United States / epidemiology