Estimating Subject-Specific Treatment Differences for Risk-Benefit Assessment with Event-Time Data in the Presence of Competing Risks

• Activity Number: 292
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Brian Claggett*+ and Lihui Zhao and LJ Wei and Lu Tian and Davide Castagno
• Companies: Harvard School of Public Health and Harvard School of Public Health and Harvard University and Stanford University and University of Turin
• Keywords: Clinical trial ; Nonparametric estimation ; Personalized medicine ; Survival analysis

Abstract

To evaluate treatment efficacy using event-time data from a comparative study, one usually makes inference about a summary measure which quantifies an overall treatment difference. However, a positive result based on such a measure does not mean that every future subject should be treated by the new therapy. It is desirable to identify subjects, using baseline covariates, who would benefit from the new treatment from a risk-benefit perspective. In this paper, we propose a systematic approach to achieve this goal with competing risk event-time data. First, we utilize data from a similar, but independent, study to build a parametric score for stratifying the current study patients. We then use the present study to obtain a nonparametric estimate of the treatment difference, with respect to each event, for any fixed score. Confidence interval and band estimates are constructed to quantify uncertainty in our inferences for the treatment differences over the score. To illustrate the new proposal, we use data from two cardiovascular studies for evaluating specific beta-blockers. The score is based on time to death, and the competing events are death, MI, hospitalization and toxicity.