arXiv subjects

Zhenke Wu

Publications and source records attributed to Zhenke Wu.

At least 37 records · Page 2Linked to original sources

arXiv2025

Random Forest for Dynamic Risk Prediction or Recurrent Events: A Pseudo-Observation Approach

Recurrent events are common in clinical, healthcare, social and behavioral studies. A recent analysis framework for potentially censored recurrent event data is to construct a censored longitudinal data set consisting of times to the first recurrent event in multiple prespecified follow-up windows of length $τ$. With the staggering number of potential predictors being generated from genetic, -omic, and electronic health records sources, machine learning approaches such as the random forest are growing in popularity, as they can incorporate information from highly correlated predictors with non-standard relationships. In this paper, we bridge this gap by developing a random forest approach for dynamically predicting probabilities of remaining event-free during a subsequent $τ$-duration follow-up period from a reconstructed censored longitudinal data set. We demonstrate the increased ability of our random forest algorithm for predicting the probability of remaining event-free over a $τ$-duration follow-up period when compared to the recurrent event modeling framework of Xia et al. (2020) in settings where association between predictors and recurrent event outcomes is complex in nature. The proposed random forest algorithm is demonstrated using recurrent exacerbation data from the Azithromycin for the Prevention of Exacerbations of Chronic Obstructive Pulmonary Disease (Albert et al., 2011).

stat.AP↗

arXiv2025

Multivariate Dynamic Mediation Analysis under a Reinforcement Learning Framework

Mediation analysis is an important analytic tool commonly used in a broad range of scientific applications. In this article, we study the problem of mediation analysis when there are multivariate and conditionally dependent mediators, and when the variables are observed over multiple time points. The problem is challenging, because the effect of a mediator involves not only the path from the treatment to this mediator itself at the current time point, but also all possible paths pointed to this mediator from its upstream mediators, as well as the carryover effects from all previous time points. We propose a novel multivariate dynamic mediation analysis approach. Drawing inspiration from the Markov decision process model that is frequently employed in reinforcement learning, we introduce a Markov mediation process paired with a system of time-varying linear structural equation models to formulate the problem. We then formally define the individual mediation effect, built upon the idea of simultaneous interventions and intervention calculus. We next derive the closed-form expression and propose an iterative estimation procedure under the Markov mediation process model. We study both the asymptotic property and the empirical performance of the proposed estimator, and further illustrate our method with a mobile health application.

stat.ME↗

arXiv2025

Counterfactually Fair Reinforcement Learning via Sequential Data Preprocessing

When applied in healthcare, reinforcement learning (RL) seeks to dynamically match the right interventions to subjects to maximize population benefit. However, the learned policy may disproportionately allocate efficacious actions to one subpopulation, creating or exacerbating disparities in other socioeconomically-disadvantaged subgroups. These biases tend to occur in multi-stage decision making and can be self-perpetuating, which if unaccounted for could cause serious unintended consequences that limit access to care or treatment benefit. Counterfactual fairness (CF) offers a promising statistical tool grounded in causal inference to formulate and study fairness. In this paper, we propose a general framework for fair sequential decision making. We theoretically characterize the optimal CF policy and prove its stationarity, which greatly simplifies the search for optimal CF policies by leveraging existing RL algorithms. The theory also motivates a sequential data preprocessing algorithm to achieve CF decision making under an additive noise assumption. We prove and then validate our policy learning approach in controlling unfairness and attaining optimal value through simulations. Analysis of a digital health dataset designed to reduce opioid misuse shows that our proposal greatly enhances fair access to counseling.

stat.ML↗

arXiv2025

Testing Stationarity and Change Point Detection in Reinforcement Learning

We consider offline reinforcement learning (RL) methods in possibly nonstationary environments. Many existing RL algorithms in the literature rely on the stationarity assumption that requires the system transition and the reward function to be constant over time. However, the stationarity assumption is restrictive in practice and is likely to be violated in a number of applications, including traffic signal control, robotics and mobile health. In this paper, we develop a consistent procedure to test the nonstationarity of the optimal Q-function based on pre-collected historical data, without additional online data collection. Based on the proposed test, we further develop a sequential change point detection method that can be naturally coupled with existing state-of-the-art RL methods for policy optimization in nonstationary environments. The usefulness of our method is illustrated by theoretical results, simulation studies, and a real data example from the 2018 Intern Health Study. A Python implementation of the proposed procedure is available at https://github.com/limengbinggz/CUSUM-RL.

stat.ML↗

arXiv2024

A Bayesian Approach to Modeling Variance of Intensive Longitudinal Biomarker Data as a Predictor of Health Outcomes

Intensive longitudinal biomarker data are increasingly common in scientific studies that seek temporally granular understanding of the role of behavioral and physiological factors in relation to outcomes of interest. Intensive longitudinal biomarker data, such as those obtained from wearable devices, are often obtained at a high frequency typically resulting in several hundred to thousand observations per individual measured over minutes, hours, or days. Often in longitudinal studies, the primary focus is on relating the means of biomarker trajectories to an outcome, and the variances are treated as nuisance parameters, although they may also be informative for the outcomes. In this paper, we propose a Bayesian hierarchical model to jointly model a cross-sectional outcome and the intensive longitudinal biomarkers. To model the variability of biomarkers and deal with the high intensity of data, we develop subject-level cubic B-splines and allow the sharing of information across individuals for both the residual variability and the random effects variability. Then different levels of variability are extracted and incorporated into an outcome submodel for inferential and predictive purposes. We demonstrate the utility of the proposed model via an application involving bio-monitoring of hertz-level heart rate information from a study on social stress.

stat.ME↗

arXiv2023

ddtlcm: An R package for overcoming weak separation in Bayesian latent class analysis via tree-regularization

Traditional applications of latent class models (LCMs) often focus on scenarios where a set of unobserved classes are well-defined and easily distinguishable. However, in numerous real-world applications, these classes are weakly separated and difficult to distinguish, creating significant numerical challenges. To address these issues, we have developed an R package ddtlcm that provides comprehensive analysis and visualization tools designed to enhance the robustness and interpretability of LCMs in the presence of weak class separation, particularly useful for small sample sizes. This package implements a tree-regularized Bayesian LCM that leverages statistical strength between latent classes to make better estimates using limited data. A Shiny app has also been developed to improve user interactivity. In this paper, we showcase a typical analysis pipeline with simulated data using ddtlcm. All software has been made publicly available on CRAN and GitHub.

stat.CO↗

arXiv2023

A Reinforcement Learning Framework for Dynamic Mediation Analysis

Mediation analysis learns the causal effect transmitted via mediator variables between treatments and outcomes and receives increasing attention in various scientific domains to elucidate causal relations. Most existing works focus on point-exposure studies where each subject only receives one treatment at a single time point. However, there are a number of applications (e.g., mobile health) where the treatments are sequentially assigned over time and the dynamic mediation effects are of primary interest. Proposing a reinforcement learning (RL) framework, we are the first to evaluate dynamic mediation effects in settings with infinite horizons. We decompose the average treatment effect into an immediate direct effect, an immediate mediation effect, a delayed direct effect, and a delayed mediation effect. Upon the identification of each effect component, we further develop robust and semi-parametrically efficient estimators under the RL framework to infer these causal effects. The superior performance of the proposed method is demonstrated through extensive numerical studies, theoretical results, and an analysis of a mobile health dataset.

stat.ML↗

arXiv2023

Bayesian Nested Latent Class Models for Cause-of-Death Assignment using Verbal Autopsies Across Multiple Domains

Understanding cause-specific mortality rates is crucial for monitoring population health and designing public health interventions. Worldwide, two-thirds of deaths do not have a cause assigned. Verbal autopsy (VA) is a well-established tool to collect information describing deaths outside of hospitals by conducting surveys to caregivers of a deceased person. It is routinely implemented in many low- and middle-income countries. Statistical algorithms to assign cause of death using VAs are typically vulnerable to the distribution shift between the data used to train the model and the target population. This presents a major challenge for analyzing VAs as labeled data are usually unavailable in the target population. This article proposes a Latent Class model framework for VA data (LCVA) that jointly models VAs collected over multiple heterogeneous domains, assign cause of death for out-of-domain observations, and estimate cause-specific mortality fractions for a new domain. We introduce a parsimonious representation of the joint distribution of the collected symptoms using nested latent class models and develop an efficient algorithm for posterior inference. We demonstrate that LCVA outperforms existing methods in predictive performance and scalability. Supplementary materials for this article and the R package to implement the model are available online.

stat.ME↗

arXiv2023

Bayesian Active Questionnaire Design for Cause-of-Death Assignment Using Verbal Autopsies

Only about one-third of the deaths worldwide are assigned a medically-certified cause, and understanding the causes of deaths occurring outside of medical facilities is logistically and financially challenging. Verbal autopsy (VA) is a routinely used tool to collect information on cause of death in such settings. VA is a survey-based method where a structured questionnaire is conducted to family members or caregivers of a recently deceased person, and the collected information is used to infer the cause of death. As VA becomes an increasingly routine tool for cause-of-death data collection, the lengthy questionnaire has become a major challenge to the implementation and scale-up of VAs. In this paper, we propose a novel active questionnaire design approach that optimizes the order of the questions dynamically to achieve accurate cause-of-death assignment with the smallest number of questions. We propose a fully Bayesian strategy for adaptive question selection that is compatible with any existing probabilistic cause-of-death assignment methods. We also develop an early stopping criterion that fully accounts for the uncertainty in the model parameters. We also propose a penalized score to account for constraints and preferences of existing question structures. We evaluate the performance of our active designs using both synthetic and real data, demonstrating that the proposed strategy achieves accurate cause-of-death assignment using considerably fewer questions than the traditional static VA survey instruments.

stat.AP↗

arXiv2022

Estimating Time-Varying Direct and Indirect Causal Excursion Effects with Longitudinal Binary Outcomes

Construction of just-in-time adaptive interventions, such as prompts delivered by mobile apps to promote and maintain behavioral change, requires knowledge about time-varying moderated effects to inform when and how we deliver intervention options. Micro-randomized trials (MRT) have emerged as a sequentially randomized design to gather requisite data for effect estimation. The existing literature (Qian et al., 2020; Boruvka et al., 2018; Dempsey et al., 2020) has defined a general class of causal estimands, referred to as "causal excursion effects", to assess the time-varying moderated effect. However, there is limited statistical literature on how to address potential between-cluster treatment effect heterogeneity and within-cluster interference in a sequential treatment setting for longitudinal binary outcomes. In this paper, based on a cluster conceptualization of the potential outcomes, we define a larger class of direct and indirect causal excursion effects for proximal and lagged binary outcomes, and propose a new inferential procedure that addresses effect heterogeneity and interference. We provide theoretical guarantees of consistency and asymptotic normality of the estimator. Extensive simulation studies confirm our theory empirically and show the proposed procedure provides consistent point estimator and interval estimates with valid coverage. Finally, we analyze a data set from a multi-institution MRT study to assess the time-varying moderated effects of mobile prompts upon binary study engagement outcomes.

stat.ME↗

arXiv2022

Dynamic Survival Transformers for Causal Inference with Electronic Health Records

In medicine, researchers often seek to infer the effects of a given treatment on patients' outcomes. However, the standard methods for causal survival analysis make simplistic assumptions about the data-generating process and cannot capture complex interactions among patient covariates. We introduce the Dynamic Survival Transformer (DynST), a deep survival model that trains on electronic health records (EHRs). Unlike previous transformers used in survival analysis, DynST can make use of time-varying information to predict evolving survival probabilities. We derive a semi-synthetic EHR dataset from MIMIC-III to show that DynST can accurately estimate the causal effect of a treatment intervention on restricted mean survival time (RMST). We demonstrate that DynST achieves better predictive and causal estimation than two alternative models.

cs.LG↗

arXiv2022

Outcome Adaptive Propensity Score Methods for Handling Censoring and High-Dimensionality: Application to Insurance Claims

Propensity scores are commonly used to reduce the confounding bias in non-randomized observational studies for estimating the average treatment effect. An important assumption underlying this approach is that all confounders that are associated with both the treatment and the outcome of interest are measured and included in the propensity score model. In the absence of strong prior knowledge about potential confounders, researchers may agnostically want to adjust for a high-dimensional set of pre-treatment variables. As such, variable selection procedure is needed for propensity score estimation. In addition, recent studies show that including variables related to treatment only in the propensity score model may inflate the variance of the treatment effect estimates, while including variables that are predictive of only the outcome can improve efficiency. In this paper, we propose a flexible approach to incorporating outcome-covariate relationship in the propensity score model by including the predicted binary outcome probability (OP) as a covariate. Our approach can be easily adapted to an ensemble of variable selection methods, including regularization methods and modern machine learning tools based on classification and regression trees. We evaluate our method to estimate the treatment effects on a binary outcome, which is possibly censored, among multiple treatment groups. Simulation studies indicate that incorporating OP for estimating the propensity scores can improve statistical efficiency and protect against model misspecification. The proposed methods are applied to a cohort of advanced stage prostate cancer patients identified from a private insurance claims database for comparing the adverse effects of four commonly used drugs for treating castration-resistant prostate cancer.

stat.ME↗

arXiv2022

Variance as a predictor of health outcomes: Subject-level trajectories and variability of sex hormones to predict body fat changes in peri- and post-menopausal women

Longitudinal biomarker data and cross-sectional outcomes are routinely collected in modern epidemiology studies, often with the goal of informing tailored early intervention decisions. For example, hormones such as estradiol and follicle-stimulating hormone may predict changes in womens' health during the midlife. Most existing methods focus on constructing predictors from mean marker trajectories. However, subject-level biomarker variability may also provide critical information about disease risks and health outcomes. In this paper, we develop a joint model that estimates subject-level means and variances of longitudinal biomarkers to predict a cross-sectional health outcome. Simulations demonstrate excellent recovery of true model parameters. The proposed method provides less biased and more efficient estimates, relative to alternative approaches that either ignore subject-level differences in variances or perform two-stage estimation where estimated marker variances are treated as observed. Analyses of women's health data reveal larger variability of E2 or larger variability of FSH were associated with higher levels of fat mass change and higher levels of lean mass change across the menopausal transition.

stat.AP↗

arXiv2022

Weakly-supervised Multi-output Regression via Correlated Gaussian Processes

Multi-output regression seeks to borrow strength and leverage commonalities across different but related outputs in order to enhance learning and prediction accuracy. A fundamental assumption is that the output/group membership labels for all observations are known. This assumption is often violated in real applications. For instance, in healthcare datasets, sensitive attributes such as ethnicity are often missing or unreported. To this end, we introduce a weakly-supervised multi-output model based on dependent Gaussian processes. Our approach is able to leverage data without complete group labels or possibly only prior belief on group memberships to enhance accuracy across all outputs. Through intensive simulations and case studies on an Insulin, Testosterone and Bodyfat dataset, we show that our model excels in multi-output settings with missing labels, while being competitive in traditional fully labeled settings. We end by highlighting the possible use of our approach in fair inference and sequential decision-making.

stat.ML↗

arXiv2022

baker: An R package for Nested Partially-Latent Class Models

This paper describes and illustrates the functionality of the baker R package. The package estimates a suite of nested partially-latent class models (NPLCM) for multivariate binary responses that are observed under a case-control design. The baker package allows researchers to flexibly estimate population-level class prevalences and posterior probabilities of class membership for individual cases. Estimation is accomplished by calling a cross-platform automatic Bayesian inference software JAGS through a wrapper R function that parses model specifications and data inputs. The baker package provides many useful features, including data ingestion, exploratory data analyses, model diagnostics, extensive plotting and visualization options, catalyzing communications between practitioners and domain scientists. Package features and workflows are illustrated using simulated and real data sets. Package URL: https://github.com/zhenkewu/baker

stat.ME↗

arXiv2022

Probabilistic Learning of Treatment Trees in Cancer

Accurate identification of synergistic treatment combinations and their underlying biological mechanisms is critical across many disease domains, especially cancer. In translational oncology research, preclinical systems such as patient-derived xenografts (PDX) have emerged as a unique study design evaluating multiple treatments administered to samples from the same human tumor implanted into genetically identical mice. In this paper, we propose a novel Bayesian probabilistic tree-based framework for PDX data to investigate the hierarchical relationships between treatments by inferring treatment cluster trees, referred to as treatment trees (Rx-tree). The framework motivates a new metric of mechanistic similarity between two or more treatments accounting for inherent uncertainty in tree estimation; treatments with a high estimated similarity have potentially high mechanistic synergy. Building upon Dirichlet Diffusion Trees, we derive a closed-form marginal likelihood encoding the tree structure, which facilitates computationally efficient posterior inference via a new two-stage algorithm. Simulation studies demonstrate superior performance of the proposed method in recovering the tree structure and treatment similarities. Our analyses of a recently collated PDX dataset produce treatment similarity estimates that show a high degree of concordance with known biological mechanisms across treatments in five different cancers. More importantly, we uncover new and potentially effective combination therapies that confer synergistic regulation of specific downstream biological pathways for future clinical investigations. Our accompanying code, data, and shiny application for visualization of results are available at: https://github.com/bayesrx/RxTree.

stat.AP↗

arXiv2021

Tree-informed Bayesian multi-source domain adaptation: cross-population probabilistic cause-of-death assignment using verbal autopsy

Determining causes of deaths (COD) occurred outside of civil registration and vital statistics systems is challenging. A technique called verbal autopsy (VA) is widely adopted to gather information on deaths in practice. A VA consists of interviewing relatives of a deceased person about symptoms of the deceased in the period leading to the death, often resulting in multivariate binary responses. While statistical methods have been devised for estimating the cause-specific mortality fractions (CSMFs) for a study population, continued expansion of VA to new populations (or "domains") necessitates approaches that recognize between-domain differences while capitalizing on potential similarities. In this paper, we propose such a domain-adaptive method that integrates external between-domain similarity information encoded by a pre-specified rooted weighted tree. Given a cause, we use latent class models to characterize the conditional distributions of the responses that may vary by domain. We specify a logistic stick-breaking Gaussian diffusion process prior along the tree for class mixing weights with node-specific spike-and-slab priors to pool information between the domains in a data-driven way. Posterior inference is conducted via a scalable variational Bayes algorithm. Simulation studies show that the domain adaptation enabled by the proposed method improves CSMF estimation and individual COD assignment. We also illustrate and evaluate the method using a validation data set. The paper concludes with a discussion on limitations and future directions.

stat.ME↗

arXiv2021

Assessing Time-Varying Causal Effect Moderation in the Presence of Cluster-Level Treatment Effect Heterogeneity

The micro-randomized trial (MRT) is a sequential randomized experimental design to empirically evaluate the effectiveness of mobile health (mHealth) intervention components that may be delivered at hundreds or thousands of decision points. MRTs have motivated a new class of causal estimands, termed "causal excursion effects", for which semiparametric inference can be conducted via a weighted, centered least squares criterion (Boruvka et al., 2018). Existing methods assume between-subject independence and non-interference. Deviations from these assumptions often occur. In this paper, causal excursion effects are revisited under potential cluster-level treatment effect heterogeneity and interference, where the treatment effect of interest may depend on cluster-level moderators. Utility of the proposed methods is shown by analyzing data from a multi-institution cohort of first year medical residents in the United States.

stat.ME↗