ABSTRACT
Diagnostic procedures for the assessment of model adequacy are as important in parametric modelling as they are when the Cox regression model is used in the analysis of survival data. Procedures based on residuals are particularly relevant, and so we begin this chapter by dening residuals for parametric models, some of which stem from those developed for the Cox model, described in Chapter 4. This is followed by a summary of graphical procedures for assessing the suitability of models tted to data that are assumed to have a Weibull, log-logistic or lognormal distribution. Other ways of examining the t of a parametric regression model are then considered, along with methods for the detection of in uential observations. We conclude with a summary of how the assumption of proportional hazards can be examined after tting the Weibull proportional hazards model.
