Relative survival techniques are useful when cause-specific death information is not accurate. We present a transformation approach which instead gives for each individual an outcome measure relative to the appropriate background population. It provides additional information on relative survival, and gives new options in regression analysis. The regression models for the new outcome measure are different from existing models, thus providing new possibilities in analysing relative survival data. One distinctive feature of our approach is that we adjust for expected survival before modelling.
COBISS.SI-ID: 18287577
Additive regression models are preferred over multiplicative models in analysis of relative survival data. While there is an abundance of methods to check the goodness of fit of multiplicative models, the respective arsenal for additive models is almost empty. We propose here a variety of procedures for testing for constant as opposed to time-varying additive effects, based on partial residuals defined similarly to Schoenfeld residuals familiar for Cox model diagnostics.
COBISS.SI-ID: 19987161
A coefficient of explained randomness was presented by Kent. Kent and O'Quigley developed these ideas, obtaining simple, multiple and partial coefficients for the situation of proportional hazards regression. Xu and O'Quigley developed a more direct approach.One purpose of this paper is to indicate that, under an independent censoring assumption, the two population coefficients coincide. Our second purpose is to point out that a sample-based coefficient in common use can be interpreted as an estimate of explained randomness when there is no censoring.
COBISS.SI-ID: 18493657
Subtypes have been reported to exist across several breast cancer microarray studies. In this paper we identified a number of factors that can influence the accuracy of assignment of patient samples to previously identified cancer subtypes and showed that careful consideration must be given to the comparability of patient populations and datasets in assigning samples to previously identified subtypes. We also showed that a robust classification rule for assigning new samples that are not part of the original dataset from which the clusters were derived remains elusive.
COBISS.SI-ID: 23695577
In relative survival, most often an additive excess hazard model is used. The existing methods of parameter estimation postulates assumptions about the baseline excess hazard. In this paper, the authors propose a new approach to estimation of the model parameters that avoids these assumptions and rather estimates the baseline excess hazard from the data. The methods is a generalization of the Cox model, meaning that all the wealth of options in existing software for the Cox model can be used in relative survival.
COBISS.SI-ID: 24416217