Semiparametric random-effects conditional density models for longitudinal analysis with concomitant intervention

Abstract

Longitudinal data in biomedical studies often involve concomitant interventions in addition to the pre-specified repeatedly measured outcome and covariate variables. Since a concomitant intervention is often initiated when a patient exhibits an undesirable health trend, adequate statistical methods should properly incorporate the starting time of a concomitant intervention in order to reduce the potential bias of the estimated intervention effects. We propose in this paper a class of semiparametric random-effects conditional density models for evaluating the distributions and concomitant intervention effects with longitudinal observations. These models simultaneously incorporate concomitant intervention effects and intra-subject longitudinal dependence structures, and quantify the change of the distribution functions through the ratio of two conditional density functions.The conditional density ratio is assumed to have a parametric form, while the baseline density function is nonparametric. We develop a likelihood-based method for estimating the parameters and a goodness-of-fit test for testing the validity of the Statistica Sinica: Newly accepted Paper (accepted version subject to English editing) Tianqing Liu, Colin O. Wu, Zhaohai Li and Yuanzhang Li models. Finite sample properties of our estimation and testing procedures are illustrated through a simulation study and an application to a longitudinal clinical trial in depression and heart disease.

Modeling the Population Mean Outcome Trajectory in Observational Studies With Varying Time to Intervention

“…However, an intervention that has an impact on the changes in the outcome can occur at different times during the course of longitudinal studies. When the impact of the intervention depends on time to intervention (TTI), it is crucial to adjust for the TTI in modeling the longitudinal outcome trajectory; see Wu and Tian (2008), Xing and Ying (2012), Liu et al (2018), and Cho et al (2020).…”

“…Intervention can occur at different time across individuals in many studies where an impact of the TTI on the repeated outcome remains uncertain. In recent years,Wu and Tian (2008),Xing and Ying (2012),Liu et al (2018), andCho et al (2020) have proposed longitudinal models that account for the varying TTI impact on the repeated outcome in cases where none of factors that confound associations between TTI and the outcome exist. In observational studies, the intervention is rather initiated on the basis of other factors that may confound the TTI impact.…”