Post-diagnosis weight change, physical activity, and survival among women with breast cancer: a longitudinal study with missing data Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 21, 2019
  • Bradshaw, Patrick T.
    • Affiliation: Gillings School of Global Public Health, Department of Epidemiology
  • Nearly forty thousand women per year are diagnosed with breast cancer and there are currently over 2 million female breast cancer survivors in the United States. Whether survival after breast cancer diagnosis is influenced by modifiable lifestyle factors, including post-diagnosis weight change and physical activity, is unclear. These associations were examined using data from a population-based follow-up study of 1,508 women diagnosed with first primary in situ or invasive breast cancer between August 1, 1996 and July 31, 1997 in Long Island, New York. During baseline and follow-up interviews, women self-reported their height, weight, recreational physical activity levels and other factors. Additional information on clinical factors was ascertained through medical records and the New York State Cancer Registry. Vital status was determined using the National Death Index; through the end of 2005, 308 women were deceased, with 164 due to breast cancer. Approximately one-third of the subjects did not complete the follow-up interview. To address the issue of potentially non-ignorably missing data, I developed a selection model for survival analysis with time-varying covariates. A sensitivity analysis using the data on post-diagnosis weight change illustrated that a standard analysis resulted in reduced statistical efficiency and differences in magnitude of effect when compared to the selection model. Mortality was positively associated with both post-diagnosis weight loss and weight gain, regardless of the time since diagnosis or pre-diagnosis body size. More detailed analyses showed that previously reported associations of mortality with pre-diagnosis body mass index (BMI) and adult weight change were attenuated after accounting for post-diagnosis weight change, while associations with post-diagnosis weight change remained. Mortality was inversely associated with recreational physical activity, regardless of pre-diagnosis activity levels, timing of post-diagnosis activity, or pre-diagnosis BMI. Since weight gain and reduction in physical activity are common after breast cancer diagnosis, these findings that suggest weight maintenance and physical activity enhance survival among breast cancer survivors may be especially important.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Gammon, Marilie D.
Degree granting institution
  • University of North Carolina at Chapel Hill
  • Open access

This work has no parents.