Biometrics最新文献

Covariate-dependent graph learning has gained increasing interest in the graphical modeling literature for the analysis of heterogeneous data. This task, however, poses challenges to modeling, computational efficiency, and interpretability. The parameter of interest can be naturally represented as a 3-dimensional array with elements that can be grouped according to 2 directions, corresponding to node level and covariate level, respectively. In this article, we propose a novel dual group spike-and-slab prior that enables multi-level selection at covariate-level and node-level, as well as individual (local) level sparsity. We introduce a nested strategy with specific choices to address distinct challenges posed by the various grouping directions. For posterior inference, we develop a full Gibbs sampler for all parameters, which mitigates the difficulties of parameter tuning often encountered in high-dimensional graphical models and facilitates routine implementation. Through simulation studies, we demonstrate that the proposed model outperforms existing methods in its accuracy of graph recovery. We show the practical utility of our model via an application to microbiome data where we seek to better understand the interactions among microbes as well as how these are affected by relevant covariates.

In cancer treatment, the development of combination therapies requires demonstrating the contribution of each individual drug and optimizing the dose during early-phase trials. This necessitates a large sample size, presenting formidable obstacles for drug developers. To address this issue, we propose a 2-stage randomized phase II design that seamlessly integrates combination dose optimization with component contribution assessment. In stage 1, the optimal combination dose is determined by maximizing the risk-benefit tradeoff across multiple candidate combination doses. In stage 2, a multi-arm randomized phase is initiated to evaluate the contribution of each component within the combination therapy. To increase trial efficiency and reduce the sample size, efficacy data from both stages are adaptively combined using a Bayesian logistic regression model with a spike-and-slab prior. The sample size and decision cutoffs of the proposed design are systematically determined based on a novel calibration procedure to achieve desired operating characteristics. Extensive simulation studies show that the proposed design achieves the dual goals of dose optimization and contribution assessment, while yielding substantial sample size savings compared to competing designs.

Occupancy models are used to infer species distributions over large spatial extents while accounting for imperfect detection. Current approaches, however, are unable to model species occurrence over continuous spatial domains while accounting for the discrete spatial domain of the observed data. We develop a new class of spatial occupancy models that embeds a change of spatial support between the observed data and occurrence process. We use a clipped Gaussian process to represent species occurrence in continuous space, which can provide inferences at a finer resolution than the observed occupancy data. Our approach is beneficial because it allows for more realistic models of species occurrence, can account for species occurring in only a portion of a surveyed site, and can relate detection probabilities to these within-site occurrence proportions. We show how our model can be fit using Bayesian methods and develop a computationally efficient MCMC algorithm. In particular, we rely on a Vecchia approximation to implement the spatial Gaussian process describing species occurrence and develop a surrogate data approach for jointly updating the spatial terms and spatial covariance parameters. We demonstrate our model using simulated data and compare our approach to alternative spatial occupancy models. We also use our model to analyze ovenbird occurrence data collected in New Hampshire, USA.

In the studies of time-to-event outcomes, it often happens that a fraction of subjects will never experience the event of interest, and these subjects are said to be cured. The studies with a cure fraction often yield a low event rate. To reduce cost and enhance study power, 2-phase sampling designs are often adopted, especially when the covariates of interest are expensive to measure or obtain. In this paper, we consider the generalized case-cohort design for studies with a cure fraction. Under this design, the expensive covariates are measured for a subset of the study cohort, called subcohort, and for all or a subset of the remaining subjects outside the subcohort who have experienced the event during the study, called cases. We propose a 2-step estimation procedure under a class of semiparametric transformation mixture cure models. We first develop a sieve maximum weighted likelihood method based only on the complete data and also devise an Expectation-Maximization (EM) algorithm for implementation. We then update the resulting estimator via a working model between the outcome and cheap covariates or auxiliary variables using the full data. We show that the proposed update estimator is consistent and asymptotically at least as efficient as the complete-data estimator, regardless of whether the working model is correctly specified or not. We also propose a weighted bootstrap procedure for variance estimation. Extensive simulation studies demonstrate the superior performance of the proposed method in finite-sample. An application to the National Wilms' Tumor Study is provided for illustration.

Pathogens usually exist in heterogeneous variants, like subtypes and strains. Quantifying treatment effects on the different variants is important for guiding prevention policies and vaccine development. Here, we ground analyses of variant-specific effects on a formal framework for causal inference. This allows us to clarify the interpretation of existing methods and define new estimands. Unlike most of the existing literature, we explicitly consider the (realistic) setting with interference in the target population: even if individuals can be sensibly perceived as iid in randomized trial data, there will often be interference in the target population where treatments, such as vaccines, are rolled out. Thus, one of our contributions is to derive explicit conditions guaranteeing that commonly reported vaccine efficacy parameters quantify well-defined causal effects, also in the presence of interference. Furthermore, our results give alternative justifications for reporting estimands on the relative, rather than absolute, scale. We illustrate the findings with an analysis of a large HIV1 vaccine trial, where there is interest in distinguishing vaccine effects on viruses with different genome sequences.

Causal weighted quantile treatment effects (WQTEs) complement standard mean-focused causal contrasts when interest lies at the tails of the counterfactual distribution. However, existing methods for estimating and inferring causal WQTEs assume complete data on all relevant factors, which is often not the case in practice, particularly when the data are not collected for research purposes, such as electronic health records (EHRs) and disease registries. Furthermore, these data may be particularly susceptible to the outcome data being missing-not-at-random (MNAR). This paper proposes to use double sampling, through which the otherwise missing data are ascertained on a sub-sample of study units, as a strategy to mitigate bias due to MNAR data in estimating causal WQTEs. With the additional data, we present identifying conditions that do not require missingness assumptions in the original data. We then propose a novel inverse-probability weighted estimator and derive its asymptotic properties, both pointwise at specific quantiles and uniformly across quantiles over some compact subset of (0,1), allowing the propensity score and double-sampling probabilities to be estimated. For practical inference, we develop a bootstrap method that can be used for both pointwise and uniform inference. A simulation study is conducted to examine the finite sample performance of the proposed estimators. We illustrate the proposed method using EHR data examining the relative effects of 2 bariatric surgery procedures on BMI loss 3 years post-surgery.

Mediation analysis aims to decipher the underlying causal mechanisms between an exposure, an outcome, and intermediate variables called mediators. Initially developed for fixed-time mediator and outcome, it has been extended to the framework of longitudinal data by discretizing the assessment times of mediator and outcome. Yet, processes in play in longitudinal studies are usually defined in continuous time and measured at irregular and subject-specific visits. This is the case in dementia research when cerebral and cognitive changes measured at planned visits in cohorts are of interest. We thus propose a methodology to estimate the causal mechanisms between a time-fixed exposure ($X$), a mediator process ($mathcal {M}_t$), and an outcome process ($mathcal {Y}_t$) both measured repeatedly over time in the presence of a time-dependent confounding process ($mathcal {L}_t$). We consider 2 types of causal estimands, the natural effects and path-specific effects. We provide identifiability assumptions, and we employ a multivariate mixed model based on differential equations for their estimation. The performances of the method are assessed in simulations, and the method is illustrated in 2 real-world examples motivated by the 3C cerebral aging study to assess (1) the effect of educational level on functional dependency through depressive symptomatology and cognitive functioning and (2) the effect of a genetic factor on cognitive functioning potentially mediated by vascular brain lesions and confounded by neurodegeneration.

We propose non-parametric estimators of the hazard ratio for comparing two survival curves using estimating equations defined in terms of group-specific cumulative hazard functions. We first describe the methods and their asymptotic properties in the case of a constant hazard ratio. We then extend the methods and the asymptotic results when the hazard ratio is time dependent and well approximated by a locally constant function. We propose a method to select the change points in the local hazard ratios. We extend the methods to stratified estimators and propose tests for heterogeneity of constant and time-dependent hazard ratios across strata. In a simulation study, we describe the finite sample properties of the proposed estimators and compare their performance with the Cox partial maximum likelihood estimator (MLE) in terms of efficiency and accuracy of coverage rates. An example is provided to illustrate an application of the proposed methods in practice.

Understanding how biomarkers change in relation to disease pathogenesis is a key area in biomedical research. We propose a semiparametric joint model to analyze the temporal evolution of biomarkers prior to the onset of disease. The model allows for a flexible biomarker trajectory that depends on two time scales: a natural time scale such as age and time to disease onset. In practice, the natural time scale often differs from time-on-study, leading to analytical challenges such as left-truncation bias. We introduce a profile kernel estimating equation approach to estimate regression coefficients and unspecified baseline mean trajectory functions. We establish the large-sample properties of the proposed estimators and conduct simulation studies to evaluate their finite-sample performance. Our method is applied to investigate brain biomarker trajectories before the onset of preclinical Alzheimer's disease. We observed a decline in cortical thickness prior to disease onset across brain regions, with APOE4 carriers showing lower levels compared to non-carriers.

Confounding and exposure measurement error can introduce bias when drawing inference about the marginal effect of an exposure on an outcome of interest. While there are broad methodologies for addressing each source of bias individually, confounding and exposure measurement error frequently co-occur, and there is a need for methods that address them simultaneously. In this paper, corrected score methods are derived under classical additive measurement error to draw inference about marginal exposure effects using only measured variables. Three estimators are proposed based on g-formula, inverse probability weighting, and doubly-robust estimation techniques. The estimators are shown to be consistent and asymptotically normal, and the doubly-robust estimator is shown to exhibit its namesake property. The methods, which are implemented in the R package mismex, perform well in finite samples under both confounding and measurement error as demonstrated by simulation studies. The proposed doubly-robust estimator is applied to study the effects of two biomarkers on HIV-1 infection using data from the HVTN 505 preventative vaccine trial.