B0477
Title: The multivariate Bernoulli detector: Change point detection in discrete survival analysis
Authors: Willem van den Boom - National University of Singapore (Singapore) [presenting]
Maria De Iorio - National University of Singapore (Singapore)
Fang Qian - National University of Singapore (Singapore)
Alessandra Guglielmi - Politecnico di Milano (Italy)
Abstract: Time-to-event data are often recorded on a discrete scale with multiple, competing risks as potential causes for the event. In this context, the application of continuous survival analysis methods with a single risk suffers from biased estimation. Therefore, the multivariate Bernoulli detector is proposed for competing risks with discrete times involving a multivariate change point model on the cause-specific baseline hazards. Through the prior on the number of change points and their locations, dependence between change points across risks is imposed, as well as allowing for data-driven learning of their number. Then, conditionally on these change points, a multivariate Bernoulli prior is used to infer which risks are involved. The focus of posterior inference is cause-specific hazard rates and dependence across risks. Such dependence is often present due to subject-specific changes across time that affect all risks. Full posterior inference is performed through a tailored local-global Markov chain Monte Carlo (MCMC) algorithm, which exploits a data augmentation trick and MCMC updates from non-conjugate Bayesian nonparametric methods. The model in simulations and on prostate cancer data is illustrated, comparing its performance with existing approaches.
