Abstract
Data from a completed randomized trial in breast cancer are used to demonstrate and quantify the variation in estimated survival curves and log-rank statistics at different times throughout a trial. False 'plateaux' are common, as are wide fluctuations in chi2 values obtained from the log-rank test when there are few events. We show how analyses conducted at different times can demonstrate different effects. Long follow-up is often necessary to allow correct interpretation of results. We discuss the assumption of proportional hazards and the consequences of making that assumption inappropriately. We show how checking whether hazards are proportional can help in avoiding erroneous conclusions.
This is a preview of subscription content, access via your institution
Access options
Subscribe to this journal
Receive 24 print issues and online access
$259.00 per year
only $10.79 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Gregory, W., Bolland, K., Whitehead, J. et al. Cautionary tales of survival analysis: conflicting analyses from a clinical trial in breast cancer. Br J Cancer 76, 551–558 (1997). https://doi.org/10.1038/bjc.1997.424
Issue Date:
DOI: https://doi.org/10.1038/bjc.1997.424