Page 147 - Special Topic Session (STS) - Volume 3

STS535 Edsel A. P. et al.

Joint modeling of discrete-valued marker
process, competing recurrent events process, and
discrete-valued health status process
1
Edsel A. Pen˜a , Piaomu Liu 2
1 Department of Statistics, University of South Carolina Columbia, SC 29208
USA
2 Department of Mathematics and Statistics, Bentley University
Waltham, MA 02452

Abstract
This talk will describe an integrated class of joint stochastic models for a
discrete-valued longitudinal marker process, a competing recurrent event
process, and a health status process in studies where subjects or units are
dynamically observed over possibly random monitoring periods. This class of
models is potentially of high relevance and importance in the context of
precision or personalized medicine which aims to utilize complex,
unstructured, and big data for the purpose of implementing personalized
interventions in this futuristic and fast-developing precision medicine
approach.

Keywords
Competing risks; Continuous-time Markov chain; Counting process; Precision
medicine; Stochastic process models

1. Motivation
The potential and promise of precision or personalized medicine hinges on
the availability of appropriate mathematical and stochastic models, together
with the proper statistical inference procedures, of the complex, most possibly
unstructured, and big data underpinning the decision-making process in this
modern and futuristic approach. In this talk we describe an approach to the
joint modeling of three major components that could be synergistically
interacting for the subjects or units in studies in biomedical setting,
engineering settings, and even in socio-economic settings. These three
components pertain to a marker process, recurrent event process, and a
‘health’ status process, together with the impact of covariates and
interventions that are dynamically performed for these units. The class of joint
models will be described here. Due to space limitations, statistical inference
procedures, in particular, the estimation of model parameters, as well the
potential relevance and applications in precision medicine, together with the
assessment of the ‘quality-oflife’ (QOL) of units, will be deferred for the talk in

136 | I S I W S C 2 0 1 9