Year: 2006
Type: article
Abstract: Hazard functions are an important component of survival analysis as they quantify the instantaneous risk of failure at a given time point.Increasing attention has been given to testing the assumption of a constant failure rate against a failure rate that changes at a single point in time.We expand the set of alternatives to allow for the consideration of multiple change-points, and propose a model selection approach using sequential testing that considers two types of models: the piecewise constant hazard model and piecewise linear hazard model.The latter model will easily accommodate the addition of covariates.These methods are data driven and allow us to estimate not only the trend in the hazard function but where those changes in trend occur.Such an analysis is valuable as it has implications in health care policy decisions.Methods, like the ones proposed in this paper, that estimate the overall survival trends for an entire population allow researchers and clinicians a better understanding of how changing medical practice affects the survival experience for a patient population.We illustrate our methods by applying them to the NIH SEER prostate and breast cancer data sets, and the NCHS birth cohort data.
Authors: Melody S. Goodman, Yi Li, Ram C. Tiwari
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W203320639