Diagnostic checks in mixture cure models with interval-censoring
(2016) , 27 pages
Files
ISBADP2016_14_legrand_Diagnosticchecksinmixturecuremodelswithinterval-censoring_web.pdf
Open Access - Adobe PDF
- 1.33 MB
Details
- Authors
- Abstract
- Models for interval-censored survival data presenting a fraction of "cure" or "immune" patients have recently been proposed in the literature, in particular extending the mixture cure model to the case of interval-censoring. However, little is known about the fitt of such models to a given data application. We thus propose to extend the classical Cox-Snell residuals to such models to assess assumptions about the survival distribution. Moreover, as covariates may, in mixture cure models, impact either the probability to experience the event, and/or the survival distribution of the uncured patients, we define deviance residuals allowing to detect non-linearity in covariates in each part of the model. Simulation studies show the behavior of these residuals; they are then applied to an Alzheimer's disease database studying the occurrence of Mild Cognitive Impairment, which may be a precursor of Alzheimer's disease. This event is typically detected between two visits, and is thus interval-censored. Furthermore it is known that not all of the patients will experience this event, leading to a fraction of “cure" or “immune" patients.
- Affiliations
Citations
Scolas, S., Legrand, C., Oulhaj, A., & El Ghouch, A. (2016). Diagnostic checks in mixture cure models with interval-censoring (ISBA Discussion Paper 2016/14). https://hdl.handle.net/2078.5/185359