A dynamic Mover–Stayer model for recurrent event processes subject to resolution

Abstract

SummaryIn studies of affective disorder, individuals are often observed to experience recurrent symptomatic exacerbations warranting hospitalization. Interest may lie in modeling the occurrence of such exacerbations over time and identifying associated risk factors. In some patients, recurrent exacerbations are temporally clustered following disease onset, but cease to occur after a period of time.We develop a dynamic Mover-Stayer model in which a canonical binary variable associated with each event indicates whether the underlying disease has resolved. An individual whose disease process has not resolved will experience events following a standard point process model governed by a latent intensity. When the disease process resolves, the complete data intensity becomes zero and no further event will occur. An expectation-maximization algorithm is described for parametric and semiparametric model fitting based on a discrete time dynamic Mover-Stayer model and a latent intensity-based model of the underlying point process.

Mover‐stayer model with covariate effects on stayer's probability and mover's transitions

“…For a mover's transition matrix, we found only four covariates to be statistically significant for estimation of the first row transition probabilities (see Table 2A) and three for the estimation of the second row transition probabilities (see Table 2B). The covariates were selected using the methods described after (11) and (12). We note that the first three covariates in Table 1 have odds ratio greater than one and the remaining four less than 1.…”

“…The MS model in a discrete and continuous time and its generalizations have been widely applied in diverse fields including sociology, labor economics, finance, large data sets, agriculture, and medicine . Most of this works has not considered covariates.…”

“…Most of this works has not considered covariates. The exception is the work of Cook et al, who incorporated covariate effects on the transition intensities and MS probability in the context of their generalized continuous time MS model, and the work of Shen and Cook, who estimated covariate effects on the probability of becoming a stayer in their generalized continuous time dynamic MS model. In contrast, our development of the MS‐C model takes place in a discrete time.…”

“…The MS model in a discrete and continuous time and its generalizations have been widely applied in diverse fields including sociology, 7 labor economics, 8 finance, 9 large data sets, 10 agriculture, 11 and medicine. 12,13 Most of this works has not considered covariates. The exception is the work of Cook et al, 13 who incorporated covariate effects on the transition intensities and MS probability in the context of their generalized continuous time MS model, and the work of Shen and Cook, 12 who estimated covariate effects on the probability of becoming a stayer in their generalized continuous time dynamic MS model.…”

A joint modelling approach for multistate processes subject to resolution and under intermittent observations

“…Instead, if the observable non‐disability state is partitioned into a latent absorbing resolved state and partially latent transient temporary non‐disability state, patients could be characterised as resolved (by the resolved state) at any point in time regardless of their observed history. We also note that Shen and Cook recently extended their model in for recurrent event processes to allow for an intermittent observation scheme.…”