Estimation and inference for the mediation effect in a time-varying mediation model.

BACKGROUND: Traditional mediation analysis typically examines the relations among an intervention, a time-invariant mediator, and a time-invariant outcome variable. Although there may be a total effect of the intervention on the outcome, there is a need to understand the process by which the intervention affects the outcome (i.e., the indirect effect through the mediator). This indirect effect is frequently assumed to be time-invariant. With improvements in data collection technology, it is possible to obtain repeated assessments over time resulting in intensive longitudinal data. This calls for an extension of traditional mediation analysis to incorporate time-varying variables as well as time-varying effects. METHODS: We focus on estimation and inference for the time-varying mediation model, which allows mediation effects to vary as a function of time. We propose a two-step approach to estimate the time-varying mediation effect. Moreover, we use a simulation-based approach to derive the corresponding point-wise confidence band for the time-varying mediation effect. RESULTS: Simulation studies show that the proposed procedures perform well when comparing the confidence band and the true underlying model. We further apply the proposed model and the statistical inference procedure to data collected from a smoking cessation study. CONCLUSIONS: We present a model for estimating time-varying mediation effects that allows both time-varying outcomes and mediators. Simulation-based inference is also proposed and implemented in a user-friendly R package.

Variable selection for partially linear models via Bayesian subset modeling with diffusing prior.

Most existing methods of variable selection in partially linear models (PLM) with ultrahigh dimensional covariates are based on partial residuals, which involve a two-step estimation procedure. While the estimation error produced in the first step may have an impact on the second step, multicollinearity among predictors adds additional challenges in the model selection procedure. In this paper, we propose a new Bayesian variable selection approach for PLM. This new proposal addresses those two issues simultaneously as (1) it is a one-step method which selects variables in PLM, even when the dimension of covariates increases at an exponential rate with the sample size, and (2) the method retains model selection consistency, and outperforms existing ones in the setting of highly correlated predictors. Distinguished from existing ones, our proposed procedure employs the difference-based method to reduce the impact from the estimation of the nonparametric component, and incorporates Bayesian subset modeling with diffusing prior (BSM-DP) to shrink the corresponding estimator in the linear component. The estimation is implemented by Gibbs sampling, and we prove that the posterior probability of the true model being selected converges to one asymptotically. Simulation studies support the theory and the efficiency of our methods as compared to other existing ones, followed by an application in a study of supermarket data.

Comparison of methods for handling covariate missingness in propensity score estimation with a binary exposure.

BACKGROUND: Causal effect estimation with observational data is subject to bias due to confounding, which is often controlled for using propensity scores. One unresolved issue in propensity score estimation is how to handle missing values in covariates. METHOD: Several approaches have been proposed for handling covariate missingness, including multiple imputation (MI), multiple imputation with missingness pattern (MIMP), and treatment mean imputation. However, there are other potentially useful approaches that have not been evaluated, including single imputation (SI) + prediction error (PE), SI + PE + parameter uncertainty (PU), and Generalized Boosted Modeling (GBM), which is a nonparametric approach for estimating propensity scores in which missing values are automatically handled in the estimation using a surrogate split method. To evaluate the performance of these approaches, a simulation study was conducted. RESULTS: Results suggested that SI + PE, SI + PE + PU, MI, and MIMP perform almost equally well and better than treatment mean imputation and GBM in terms of bias; however, MI and MIMP account for the additional uncertainty of imputing the missingness. CONCLUSIONS: Applying GBM to the incomplete data and relying on the surrogate split approach resulted in substantial bias. Imputation prior to implementing GBM is recommended.

Challenges and Opportunities in Collecting and Modeling Ambulatory Electrodermal Activity Data.

BACKGROUND: Ambulatory assessment of electrodermal activity (EDA) is an emerging technique for capturing individuals' autonomic responses to real-life events. There is currently little guidance available for processing and analyzing such data in an ambulatory setting. OBJECTIVE: This study aimed to describe and implement several methods for preprocessing and constructing features for use in modeling ambulatory EDA data, particularly for measuring stress. METHODS: We used data from a study examining the effects of stressful tasks on EDA of adolescent mothers (AMs). A biosensor band recorded EDA 4 times per second and was worn during an approximately 2-hour assessment that included a 10-min mother-child videotaped interaction. The initial processing included filtering noise and motion artifacts. RESULTS: We constructed the features of the EDA data, including the number of peaks and their amplitude as well as EDA reactivity, quantified as the rate at which AMs returned to baseline EDA following an EDA peak. Although the pattern of EDA varied substantially across individuals, various features of EDA may be computed for all individuals enabling within- and between-individual analyses and comparisons. CONCLUSIONS: The algorithms we developed can be used to construct features for dry-electrode ambulatory EDA, which can be used by other researchers to study stress and anxiety.

Automated Proposition Density Analysis for Discourse in Aphasia.

Purpose: This study evaluates how proposition density can differentiate between persons with aphasia (PWA) and individuals in a control group, as well as among subtypes of aphasia, on the basis of procedural discourse and personal narratives collected from large samples of participants. Method: Participants were 195 PWA and 168 individuals in a control group from the AphasiaBank database. PWA represented 6 aphasia types on the basis of the Western Aphasia Battery-Revised (Kertesz, 2006). Narrative samples were stroke stories for PWA and illness or injury stories for individuals in the control group. Procedural samples were from the peanut-butter-and-jelly-sandwich task. Language samples were transcribed using Codes for the Human Analysis of Transcripts (MacWhinney, 2000) and analyzed using Computerized Language Analysis (MacWhinney, 2000), which automatically computes proposition density (PD) using rules developed for automatic PD measurement by the Computerized Propositional Idea Density Rater program (Brown, Snodgrass, & Covington, 2007; Covington, 2007). Results: Participants in the control group scored significantly higher than PWA on both tasks. PD scores were significantly different among the aphasia types for both tasks. Pairwise comparisons for both discourse tasks revealed that PD scores for the Broca's group were significantly lower than those for all groups except Transcortical Motor. No significant quadratic or linear association between PD and severity was found. Conclusion: Proposition density is differentially sensitive to aphasia type and most clearly differentiates individuals with Broca's aphasia from the other groups.