Abstract
Survival data can contain an unknown fraction of subjects who are "cured" in the sense of not being at risk of failure. We describe such data with cure-mixture models, which separately model cure status and the hazard of failure among non-cured subjects. No diagnostic currently exists for evaluating the fit of such models; the popular Schoenfeld residual (Schoenfeld, 1982. Partial residuals for the proportional hazards regression-model. Biometrika 69, 239-241) is not applicable to data with cures. In this article, we propose a pseudo-residual, modeled on Schoenfeld's, to assess the fit of the survival regression in the non-cured fraction. Unlike Schoenfeld's approach, which tests the validity of the proportional hazards (PH) assumption, our method uses the full hazard and is thus also applicable to non-PH models. We derive the asymptotic distribution of the residuals and evaluate their performance by simulation in a range of parametric models. We apply our approach to data from a smoking cessation drug trial.
Original language | English (US) |
---|---|
Pages (from-to) | 340-350 |
Number of pages | 11 |
Journal | Biostatistics |
Volume | 14 |
Issue number | 2 |
DOIs | |
State | Published - Apr 2013 |
Keywords
- Accelerated failure time
- Long-term survivors
- Proportional hazards
- Residual analysis
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty