Guarantee Time Bias: A Discussion About the Analysis of Survival by Tumor Response and Other Comparisons of Time-to-Event by Outcome Variables

• Activity Number: 594
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): James R. Anderson*+ and Robert W. Makuch*+ and Anita Giobbie-Hurder*+ and Nicholas Jewell*+
• Companies: University of Nebraska Medical Center and Yale School of Public Health and Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute and University of California Berkeley
• Keywords: guarantee time bias ; clinical trials ; survival analysis ; landmark analysis

Abstract

Despite the publication of the landmark 1983 Journal of Clinical Oncology paper by Dr.s Anderson, Cain and Gelber, which highlighted inappropriate analytical approaches and suggested alternative unbiased techniques for the analysis of survival by tumor response and other comparisons of time-to-event by outcome variable, there are still no shortages of similarly flawed approaches carried out since. Partly due to the fact that those are plausible approaches seemingly make sense yet statistically flawed, or due to the fact that statisticians were not educated or informed enough against the flawed approaches. It is also very tempting for statisticians to use early biomarkers to predict time to future events, yet many are not aware of the correct approached to eliminate estimation bias. It is imperative that statisticians are made award of the situation and be educated about the right approaches. This panel discussion session will provide a platform for the world leading experts to offer their insight on this important subject, and an educational opportunity to a broad audience of statisticians so that they are well informed to carry out appropriate analyses.