Multistate analysis of multitype recurrent event and failure time data with event feedbacks in biomarkers

Semiparametric Modeling of Biomarker Trajectory and Variability With Correlated Measurement Errors

The prognostic significance of biomarker variability in predicting associated disease risk is well-established. However, prevailing methods that assess the relationship between biomarker variability and time to event often overlook within-subject correlation in longitudinal measurement errors, resulting in biased parameter estimates and erroneous statistical inference. Additionally, these methods typically assume that biomarker trajectory can be represented as a linear combination of spline basis functions with normally distributed random effects. This not only leads to significant computational demands due to the necessity of high-dimensional integration over the random effects but also limits the applicability because of the normality restriction imposed on the random effects. This paper addresses these limitations by incorporating correlated longitudinal measurement errors and proposing a novel semiparametric multiplicative random effects model. This model does not assume normality for the random effects and eliminates the need for integration with respect to them. The biomarker variability is incorporated as a covariate within a Cox model for time-to-event data, thus facilitating a joint modeling strategy. We demonstrate the asymptotic properties of the proposed estimators and validate their performance through simulation studies. The methodology is applied to assess the impact of systolic blood pressure variability on cardiovascular mortality using data from the Atherosclerosis Risk in Communities study.

Multistate modeling and structure selection for multitype recurrent events and terminal event data

In cardiovascular disease studies, a large number of risk factors are measured but it often remains unknown whether all of them are relevant variables and whether the impact of these variables is changing with time or remains constant. In addition, more than one kind of cardiovascular disease events can be observed in the same patient and events of different types are possibly correlated. It is expected that different kinds of events are associated with different covariates and the forms of covariate effects also vary between event types. To tackle these problems, we proposed a multistate modeling framework for the joint analysis of multitype recurrent events and terminal event. Model structure selection is performed to identify covariates with time-varying coefficients, time-independent coefficients, and null effects. This helps in understanding the disease process as it can detect relevant covariates and identify the temporal dynamics of the covariate effects. It also provides a more parsimonious model to achieve better risk prediction. The performance of the proposed model and selection method is evaluated in numerical studies and illustrated on a real dataset from the Atherosclerosis Risk in Communities study.