Volume no :16, Issue no: 2, December (2016)

Journal of Statistics: Advances in Theory and Applications

JOINT MODELLING OF THE RELATIONSHIP BETWEEN SLEEP, DISEASE AND MORTALITY, EXCLUSIVELY IN A COHORT OF OLDER AUSTRALIAN WOMEN (AGED 70-75 YEARS AT BASELINE)

Author's: Lucy Leigh, Irene Lena Hudson and Julie Ellen Byles
Pages: [185] - [254]
Received Date: October 29, 2016
Submitted by:
DOI: http://dx.doi.org/10.18642/jsata_7100121735

Abstract

This paper explores the relationship between sleep, disease, and survival, utilising shared random effects joint modelling to account for informative dropout, in a cohort of very old women. Joint modelling simultaneously models the longitudinal sleep trajectory and its effect on survival. A series of nested joint models are implemented to separate out the effects on survival of sleep and disease, the latter of which is known to have a causal role in mortality, and other factors of interest, such as self-rated health, physical functioning, mental health, vitality, baseline age, BMI, area of residence, marital status, and education. Dynamic survival predictions are also created and compared for a small sample of subjects.

Keywords

ALSWH, joint models, longitudinal sleep trajectory, shared parameter models, sleep difficulty, survival.