研究园简易版 · 经典书籍 · 期刊文章 · 有原文 · 完整版(可多选筛选、搜索) · 首页
2026 年的期刊文章 · 第 26 页
共 1756 篇,按发表时间排,新的在前,每页 40 篇。这一页是 2026-04-14 到 2026-04-02。
按期刊、按主题看 →
JASA Causal Inference Machine Learning Networks & Graphs · 发表于 2026-04-14
Gemma Moran, Bryon Aragam
摘要
Recent developments in generative artificial intelligence (AI) rely on machine learning techniques such as deep learning and generative modeling to achieve state-of-the-art performance across wide-ranging domains. These methods’ surprising performance is due in part to their ability to learn implicit “representations” of complex, multi-modal data. Unfortunately, deep neural networks are notoriously black boxes that obscure these representations, making them difficult to interpret or analyze. To resolve these difficulties, one approach is to build new interpretable neural network models from the ground up. This is the goal of the emerging field of causal representation learning (CRL) that uses causality as a vector for building flexible, interpretable, and transferable generative AI. CRL can be seen as a synthesis of three intrinsically statistical ideas: (i) latent variable models such as factor analysis; (ii) causal graphical models with latent variables; and (iii) nonparametric statistics and deep learning. This article introduces CRL from a statistical perspective, focusing on connections to classical models as well as statistical and causal identifiability results. We also highlight key application areas, implementation strategies, and open statistical questions. Supplementary materials for this article are available online, including a standardized description of the materials available for reproducing the work. *(abstract via OpenAlex)*
SiM Nonparametric & Semiparametric · 发表于 2026-04-14,收录于 2026-04-17
Xuezhixing Zhang, Paul Milligan, Yin Bun Cheung
摘要
The self‐controlled case series (SCCS) method is frequently employed to explore the relationship between transient exposures and subsequent health events, utilizing data from individuals who have experienced the event of interest. Conventional spline‐based SCCS models typically do not account for overlapping exposure periods and fail to accommodate complex interactive effects among multiple exposures. In this paper, we introduce a novel semiparametric SCCS method that employs a functional partial‐linear single index (PLSI) link function, allowing for the estimation of overlapping exposure risks. Our approach offers greater interpretability and flexibility compared with existing methods by consolidating multiple exposures into a single index and modeling complex interactions through a nonparametric link function. We validate our model through simulation studies comparing its performance with standard methods under various practical exposure settings. Furthermore, we apply our method to two real‐world datasets involving MMR vaccination and malaria chemoprevention, demonstrating its practical utility and enhanced capability to handle multiple, overlapping exposures effectively. Our findings suggest that the PLSI‐SCCS model is a robust tool for modern epidemiological and pharmaceutical research, providing a nuanced understanding of exposure effects, particularly in complex multi‐exposure scenarios.
JASA Other · 发表于 2026-04-14
Xiao-Li Meng
摘要
1. Upon receiving Editor Hongtu Zhu’s invitation to comment on this refreshing article (for statistical journals), two of my inner voices immediately quarreled for a full 24 hr. The wiser one remin... *(abstract via OpenAlex)*
JASA Other · 发表于 2026-04-14
James Zou, Mert Yuksekgonul
摘要
1. An AI agent is an autonomous system that can interact with its environment, reason about tasks, take actions-often by invoking external tools or APIs-and adapt its behavior based on feedback. In... *(abstract via OpenAlex)*
Biometrika Nonparametric & Semiparametric Causal Inference · 发表于 2026-04-13,收录于 2026-10-17
Aditya Ghosh and others
摘要
Summary: Mean-based estimators of causal effects in randomized experiments may behave poorly if the potential outcomes have a heavy tail or contain outliers. An alternative estimator proposed by Rosenbaum (1993) estimates a constant additive treatment effect by inverting a randomization test using ranks. We develop a design-based asymptotic theory for this rank-based estimator and study its robustness and efficiency properties. We show that Rosenbaum’s estimator is robust against outliers with a breakdown point that uniformly dominates that of any weighted quantile estimator. When pretreatment covariates are available, a regression-adjusted version of Rosenbaum’s estimator uses an agnostic linear regression on the covariates and bases inference on the ranks of residuals. Under mild integrability conditions, we show that this estimator is at most 13.6% less efficient, in the worst case, than the commonly used mean-based regression adjustment method proposed by Lin (2013), and often outperforms it when the residuals have heavy tails. Moreover, under suitable assumptions, Rosenbaum’s regression-adjusted estimator is at least as efficient as the unadjusted one. Finally, we initiate the study of Rosenbaum’s estimator when the constant treatment effect assumption may be violated. To analyse the regression-adjusted estimator, we develop local asymptotics of rank statistics under the design-based framework, which may be of independent interest.
Biometrika Networks & Graphs · 发表于 2026-04-13
Pei Heng, Shiyuan He, Yi Sun, Jianhua Guo
摘要
Collapsibility provides a principled approach to dimension reduction in contingency tables and graphical models. Madigan & Mosurski (1990) pioneered the study of minimal collapsible sets in decomposable models, but existing algorithms for general graphs remain computationally demanding. We show that a model is collapsible on to a target set precisely when that set contains at least one minimal separator between its nonadjacent vertices. This insight motivates the close minimal separator absorption algorithm, which constructs minimal collapsible sets using only local separator searches at very low costs. Simulations confirm substantial efficiency gains, making collapsibility analysis practical in high-dimensional settings.
SiM Diagnostics & Prediction Clinical Trials · 发表于 2026-04-13,收录于 2026-05-01
Abdollah Safari, Paul Gustafson, Mohsen Sadatsafavi
摘要
(verbatim from arXiv:2410.03096v2 — Wiley returned 403, so this is the preprint abstract; exact online date not confirmed)*: "Risk prediction models are often advertised as deterministic functions that map covariates to predicted risks. However, they are typically trained using finite samples, and as such, their predictions are inherently uncertain. This uncertainty has been addressed in terms of uncertainty around metrics of model performance (e.g., confidence intervals around c-statistic), as well as uncertainty or instability of predictions. Correspondingly, sample size calculations for model development studies target the precision of estimates of summary statistics and the stability of predictions. However, when evaluating the clinical utility of a model (as in Net Benefit (NB) calculations in decision curve analysis), statistical inference is less relevant. From a decision-theoretic perspective, the finite size of the sample results in utility loss due to the discrepancy between the fitted model and the correct model. From this perspective, procuring more development data is associated with an expected gain in the utility of using the model. In this work, we define the Expected Value of Sample Information (EVSI) as the expected gain in clinical utility, defined in NB terms, by procuring an additional development sample of a given size. We propose a bootstrap-based algorithm for EVSI computations and demonstrate its feasibility and face validity in a case study. We conclude that decision-theoretic metrics can complement classical inferential methods when designing studies aimed at developing risk prediction models.
SMMR Causal Inference Nonparametric & Semiparametric · 发表于 2026-04-12
Yunlong Cao, Yuquan Wang, Dapeng Shi, Dong Chen, Yue-Qing Hu
摘要
Inferring causal effects with unmeasured confounder is a main challenge in causal inference. Many researchers impose parametric assumptions on the distribution of unmeasured confounder. However, due to the unobservable nature of the unmeasured confounder, it is more reasonable to leave its distribution unrestricted. Another key challenge in causal inference is the involvement of invalid instrumental variables, which may lead to biased inference and potentially deceptive results. To this end, we employ a flexible semiparametric model that allows for possibly invalid instruments without specifying the distribution of unmeasured confounder in this work. A penalized semiparametric estimator for causal effects is constructed and its oracle and asymptotic properties are well established for statistical inference. We evaluate the performance of the estimator through simulation studies, revealing that our proposed estimator exhibits asymptotic unbiasedness and robustness in estimating causal effects, along with consistent selection of invalid instruments. We also demonstrate its application using Atherosclerosis Risk in Communities Study data set, which further validates its robustness in the presence of invalid instruments.
SiM Causal Inference Bayesian Methods · 发表于 2026-04-11,收录于 2026-05-01
Zijin Liu, Zhihui Liu, Jennifer Dang, Charles Catton, Himanshu R. Lukka, Peter Chung, Olli Saarela
摘要
In radiotherapy for cancer, organs surrounding the target tumor, known as organs‐at‐risk (OARs), should be protected from excessive radiation to avoid toxicity. Radiation exposure to multiple OARs can be summarized using matrix‐valued dose‐volume histograms (DVH), and understanding the causal relationship between DVHs and toxicity outcomes can improve treatment planning. Conventional causal models are not tailored to high‐dimensional, highly correlated matrix‐valued data. In this paper, we propose a Bayesian three‐component joint model for a matrix‐valued DVH exposure with a causal interpretation. Dimension reduction is achieved via multilinear principal component analysis (MPCA), which extracts information from matrices more efficiently than conventional PCA. A Hamiltonian Monte Carlo algorithm is adapted for estimation. We demonstrate the model's performance in estimating average causal effects through simulations. For interpretation, we map dose effects back to the original DVH matrix, illustrating that our model can correctly identify relevant effects in both simulation and application studies.
SiM Longitudinal Data · 发表于 2026-04-10,收录于 2026-05-01
Yongxi Long, Bart C. Jacobs, Ewout W. Steyerberg, Erik W. van Zwet
摘要
Initially proposed for analyzing composite endpoints, the win odds have recently received increasing interest for the analysis of ordinal outcomes. When comparing an ordinal outcome between two groups, the win odds are the odds that a randomly selected subject from the first group has a better outcome than a randomly selected subject from the second group. As such, the win odds are an effect size that is closely related to the Mann–Whitney U test. The win odds can be adjusted for covariates by the probabilistic index model. Here, we aim to extend this model for repeated measurements. We modify the estimation equations of the probabilistic index model to account for within‐subject correlation. Parameter estimation can be conveniently done via some data re‐structuring and the R package geepack . We implement a sandwich‐type estimator to estimate the variance‐covariance matrix. Simulations show that the estimation of the win odds is consistent and the coverage of confidence intervals is close to nominal. We provide an application by reanalyzing a neurological trial for the treatment of Guillain–Barré syndrome (SID‐GBS trial). This extension establishes the win odds as a promising summary measure to compare longitudinal ordinal outcomes. R package lwo is available on GitHub for implementing the proposed method.
SiM Other · 发表于 2026-04-10,收录于 2026-05-01
Guoqing Diao, Arvind Shah, Jianxin Lin, Joseph G. Ibrahim
摘要
Meta‐analysis is a popular statistical technique in biomedical research. In particular, meta‐analysis can assist clinicians in determining whether an intervention is effective or which intervention is most effective. Conventional meta‐analysis combines aggregated information from multiple published studies and addresses a common research question. In recent years, there has been a growing interest in combining studies with individual participant data and studies with only aggregated data, with the aim of potentially improve statistical efficiency of the treatment effect estimators. We develop a novel information‐based method using composite likelihood for hybrid meta‐analysis integrating individual participant data and aggregated data. The proposed method utilizes all available information from aggregated data, including the descriptive statistics (if any) of the outcome variable and covariates reported in the published papers, and accounts for potential between‐study variability in terms of the underlying distribution of the outcome. The unknown parameters are estimated by maximizing the composite likelihood function. The resulting estimators are shown to be consistent and asymptotically normal by using composite likelihood theory. Simulation studies demonstrate that the proposed method tends to be more efficient than the existing meta‐analysis methods. An application to multiple clinical trials comparing LDL‐C‐lowering treatments is provided.
SiM High-Dimensional Longitudinal Data · 发表于 2026-04-09,收录于 2026-05-01
Mingtao Zhao, Jingxiang Cao, Jun Sun, Yan Fan, Sanying Feng, Fanqun Li
摘要
In this article, we propose a bias‐corrected double penalized quadratic inference functions method to simultaneously identify model structure, estimate parameters, and perform variable selection for varying coefficient errors‐in‐variables (EV) models with longitudinal data. Unlike the linear models or the partial linear varying coefficient models, the proposed method does not assume in advance whether each regression coefficient is constant or varying. Instead, it represents each coefficient as a nonparametric function and identifies whether it is constant or varying using the proposed method. By employing a B‐spline basis to approximate the unknown coefficient functions, the proposed method integrates a bias‐corrected quadratic inference function with two penalized terms to achieve structure identification, estimation, and variable selection. Under certain regularized conditions, the consistency and sparsity properties of the estimator are established. Moreover, a three‐step iterative algorithm is developed to implement the proposed method in practice. Simulation studies and a real data analysis demonstrate the superior finite‐sample performance of the method.
Biometrika Nonparametric & Semiparametric · 发表于 2026-04-08
A Martin, M Santacatterina, I Díaz
摘要
Summary: Marginal structural models are a popular method for estimating causal effects in the presence of time-varying exposures. In spite of their popularity, no scalable nonparametric estimator exists for marginal structural models with multi-valued or continuous time-varying treatments. In this paper, we combine flexible, data-adaptive regression methods, including ensemble learning techniques, with recent developments in semiparametric efficiency theory for longitudinal studies to propose such an estimator. The proposed estimator is based on a study of the nonparametric identifying functional, including first-order von Mises expansions, as well as the efficient influence function and the efficiency bound. We show conditions under which the proposed estimators are efficient, asymptotically normal and sequentially doubly robust. We perform a simulation study to illustrate the properties of the estimators, and present the results of our motivating study on a COVID-19 dataset, studying the impact of mobility on the cumulative number of observed cases.
SiM Causal Inference Survey & Sampling · 发表于 2026-04-08,收录于 2026-05-01
Yukang Zeng, Fan Li, Guangyu Tong
摘要
Propensity score weighting is a common method for estimating treatment effects with observational data, by addressing confounding due to measured baseline covariates. However, when the observational data sample is drawn based on a survey, the existing literature does not reach a consensus on the optimal use of survey weights for population‐level causal inference. Under the balancing weights framework, we provide a unified solution for incorporating survey weights and derive a set of weighting and augmented weighting estimators for different target populations, including the combined, treated, controlled, and overlap populations. We also develop closed‐form sandwich variance estimators for each estimator via the theory of M‐estimators. Through an extensive series of simulation studies, we examined the performance of our estimators and compared the results to those of alternative methods. We carried out two case studies to illustrate the application of the different propensity score methods with complex survey data. We concluded with a discussion of our findings and provided some practical recommendations for propensity score weighting analysis of survey observational data.
SMMR Latent Variables & Mixtures Longitudinal Data Functional Data · 发表于 2026-04-08
Ying Cui, Jeong Hoon Jang, Robert G Mannino, Amita K Manatunga
摘要
In this paper, we develop a novel method for clustering multiple images while adjusting for the effect of available covariates on cluster membership. The key strategy is to represent each image as two-dimensional functional data and formulate a functional latent class mixed model, which fully leverages the structural information of images while effectively addressing their high dimensionality and accounting for covariate effects. We apply the proposed method to color intensity matrices extracted from patient-sourced smartphone fingernail photos to identify distinct subgroups, while adjusting for the effect of image metadata, which may act as an effect modifier on cluster membership. Information on these subgroups can assist public health officials in low-resource settings by enabling rapid and non-invasive identification of high-risk subpopulations for anemia, thereby facilitating the timely delivery of targeted interventions. The results suggest that the three clusters identified by the proposed method correspond to varying levels of anemia risk, with 0%, 79%, and 86% of subjects in each cluster classified as anemic. These findings demonstrate the utility of the proposed method and highlight the potential of a smartphone application leveraging fingernail images for non-invasive and cost-effective anemia screening.
SMMR Causal Inference Nonparametric & Semiparametric · 发表于 2026-04-08
Sima Sharghi, Kevin E Stoll, Sally W Thurston, Emily Barrett, Brent Johnson
摘要
To study whether air pollution is detrimental to reproductive development is imperative. In the absence of randomized trials to study the effects of air pollution on human health, data from observational studies have been utilized in which the researchers attempted to capture the causal associations between air pollution and the health outcomes. Many of these studies rely on parametric assumptions which may be limiting. In this tutorial, we explain and implement the nonparametric empirical likelihood (EL) Algorithm within the causal inference framework of a classic methodology and a newer technique based on machine learning tools. We show the competitive results of the assumption free EL in simulations. We also apply the developed methods to study the causal association between PM2.5 and NO2 exposure and anogenital distance at birth, a marker of androgen activity.
SiM Missing Data & Measurement Error · 发表于 2026-04-08,收录于 2026-05-01
Yuanyuan Luan, Roger S. Zoh, Sneha Jadhav, Lan Xue, Carmen D. Tekwe
摘要
Most methods for adjusting for biases due to measurement errors in covariates in generalized linear regression models focus on scalar covariates. Less work exists to correct for biases due to measurement error in a mixture of functional and scalar covariates. To address this limitation, we develop joint functional simulation extrapolation (FSIMEX) and mixed effects model‐based (MEM) approaches in a mixture of functional and scalar covariates prone to classical measurement errors in generalized functional linear regression. The proposed methods were compared with the Oracle estimator, which served as a benchmark, principal components analysis through the conditional expectation (PACE), Naive_ave, and Naive_one methods through extensive simulations to assess their finite sample performance. Across a range of conditions, the joint FSIMEX estimator generally had low biases, followed by the MEM estimator, similar to those of the Oracle estimator. The SIMEX and MEM methods for measurement error adjustment performed notably better than the Naive_one estimators that did not adjust for measurement error. Although the Naive_ave and PACE estimators adjusted for some of the measurement errors, they did not account for the heteroscedastic covariance structures associated with the functional covariate and the PACE approach was not designed for modeling scalar predictors. We applied our methods to data from the 2011–2014 cycles of the National Health and Examination Survey (NHANES) to assess the relationships physical activity, total caloric intake and demographic characteristics have with type 2 diabetes status of community dwelling adults living in the United States. We treated the device‐based measure of physical activity as a functional covariate prone to complex arbitrary heteroscedastic errors and total caloric intake as a scalar covariate prone to error. Estimation with the joint FSIMEX and MEM estimators reduced the associations for some covariates but increased them for others, particularly physical activity, in comparison to estimation with the estimators that did not adjust for biases due to measurement error. Overall, we observed that failing to account for measurement error due to biases can lead to biased estimations of the functional and scalar covariates, which are prone to errors in the generalized functional linear regression models.
SMMR Causal Inference Nonparametric & Semiparametric · 发表于 2026-04-08
Rong J.B. Zhu
摘要
Propensity score weighting is a common method in causal inference methods. However, this approach faces two well-known challenges: (i) high variance due to small probability values in the denominator, and (ii) sensitivity to model specification errors when estimating propensity scores in observational studies. In this article, we establish that the expected potential outcomes, conditional on the propensity score, can be identified as a function of the propensity score. Based on this identification result, we utilize spline regression to estimate the function. Treatment effect estimation and inference are derived from the asymptotic normality of the spline regression. Our approach preserves the low bias of inverse probability weighting (IPW), benefiting from the flexibility of nonparametric models, while achieving significantly lower variance due to its model-based stability. Furthermore, we extend this method to regression-based adjustment for improved efficiency in causal inference. Extensive simulations show that our approach achieves lower variance than IPW-based methods while maintaining low bias and robustness to propensity score misspecification. A real-data application further demonstrates its reliability in inference.
SMMR Causal Inference Bayesian Methods Clinical Trials · 发表于 2026-04-08
Annan Deng, Carole Siegel, Hyung G Park
摘要
An important goal of precision medicine is to personalize medical treatment by identifying individuals who are most likely to benefit from a specific treatment. The likely responder (LR) framework, which identifies a subpopulation where treatment response is expected to exceed a certain clinical threshold, plays a role in this effort. However, the LR framework, and more generally, data-driven subgroup analyses, often fail to account for uncertainty in the estimation of model-based data-driven subgrouping. We propose a simple two-stage approach that integrates subgroup identification with subsequent subgroup-specific inference on treatment effects. We incorporate model estimation uncertainty from the first stage into subgroup-specific treatment effect estimation in the second stage, by utilizing Bayesian posterior distributions from the first stage. We evaluate our method through simulations, demonstrating that the proposed Bayesian two-stage model produces better calibrated confidence intervals than naïve approaches. We apply our method to an international COVID-19 treatment trial, which shows substantial variation in treatment effects across data-driven subgroups.
SiM Causal Inference Survival Analysis · 发表于 2026-04-07,收录于 2026-04-24
Yi Guo, Dan Chen, Xinming Xu, Zhicheng Zhang, Yu Wen, Xueying Zheng, Zhenyu Wu, Xu Qin
摘要
The validity of mediation analysis critically depends on the assumption of no unmeasured confounding, which is typically evaluated through sensitivity analysis. However, existing sensitivity analysis methods for mediation analysis with survival outcomes often rely on the rare outcome assumption and neglect the relationship between exposure and unmeasured confounding. To address this challenge, we developed a sensitivity analysis approach to assess the robustness of mediation results. Specifically, it can assess sensitivity to both mediator‐outcome confounding and confounding involving the exposure in observational studies. Our method innovatively generates a simulated unmeasured confounder from its conditional distribution, constructed through sensitivity parameters such as regression coefficients linking the outcome, mediator, and exposure to the unmeasured confounder. Then the sensitivity of mediation analysis can be evaluated by comparing the results before and after adjusting for the simulated unmeasured confounder. A three‐dimensional visualization tool was developed to visualize the sensitivity of mediation analysis results. We validated our methodology on simulated datasets and further applied it to a real dataset from the China Health and Nutrition Survey (CHNS) to investigate the relationship between obesity and stroke mediated by hypertension. An R package “medsenssurv” has also been developed to facilitate implementation of our proposed method ( https://github.com/Guo‐yi‐y/medsenssurv ).
SiM Genetics & Omics · 发表于 2026-04-07,收录于 2026-05-01
Johannes Ostner, Hongzhe Li, Christian L. Müller
摘要
The class of a‐b power interaction models, proposed by [1], provides a general framework for modeling sparse compositional data with pairwise feature interactions. This class includes many distributions as special cases and enables modeling of zero entries through power transformations, making it particularly suitable for modern high‐throughput sequencing data with excess zeros, including single‐cell RNA‐Seq and microbial amplicon data. Here, we present an extension of this class of models that allows inclusion of covariate information, thus enabling accurate characterization of covariate dependencies in heterogeneous populations. Combining this model with a tailored differential abundance (DA) test leads to a novel DA testing scheme, cosmoDA , that can reduce the false positive detection rate caused by correlated features. cosmoDA uses penalized generalized score matching for parsimonious model fitting. We show on simulated benchmarks that cosmoDA can accurately estimate feature interactions in the presence of population heterogeneity and significantly reduces the false discovery rate when testing for differential abundance of correlated features. Using single‐cell and amplicon data, we illustrate cosmoDA 's ability to estimate data‐adaptive Box–Cox‐type data transformations and assess the impact of zero replacement and power transformations on downstream differential abundance results. cosmoDA is available at https://github.com/bio‐datascience/cosmoDA .
SiM Survival Analysis Nonparametric & Semiparametric · 发表于 2026-04-07,收录于 2026-05-01
Seungwoo Lee, Laurent Briollais, Yun‐Hee Choi
摘要
Modeling of medical interventions, such as preventive surgeries, on a survival outcome necessitates an accurate and flexible representation of the time‐dependent effect of the intervention. We propose using B‐splines to model the time‐dependent effect of a binary time‐dependent covariate on a survival outcome in the presence of correlated competing risks. The proposed B‐splines can also help choose the best functional form for the time‐dependent effect. Our simulation studies demonstrate that the proposed competing risks model with B‐splines performs well in estimating the time‐dependent effect of the intervention and the penetrance function in all scenarios considered. We applied our proposed method to families carrying a pathogenic variant in BRCA1 recruited through the Breast Cancer Family Registry and evaluated the time‐dependent effects of risk‐reducing salpingo‐oophorectomy on breast cancer risk in the presence of ovarian cancer and death from other causes as competing events. We found that the B‐splines model better captured the changes in the association between the intervention and breast cancer over time and provided better predictive ability than alternative parametric models, such as the permanent exposure model and the Cox and Oakes model. Our results indicate that risk‐reducing salpingo‐oophorectomy has a significant protective effect on breast cancer incidence over the follow‐up period.
SiM Bayesian Methods Testing & Uncertainty Nonparametric & Semiparametric · 发表于 2026-04-07,收录于 2026-05-01
Xin‐Wei Huang, Saptarshi Chakraborty
摘要
Spontaneous reporting system databases are key resources for post‐marketing surveillance, providing real‐world evidence (RWE) on the adverse events (AEs) of regulated drugs or other medical products. Various statistical methods have been proposed for AE signal detection in these databases, flagging drug‐specific AEs with disproportionately high observed counts compared to expected counts under independence. However, signal detection remains challenging for rare AEs or newer drugs, which receive small observed and expected counts and thus suffer from reduced statistical power. Principled information sharing on signal strengths across drugs/AEs is crucial in such cases to enhance signal detection. However, existing methods typically ignore complex between‐drug associations on AE signal strengths, limiting their ability to detect signals. We propose novel local–global mixture Dirichlet process (DP) prior‐based nonparametric Bayesian models to capture these associations, enabling principled information sharing between drugs while balancing flexibility and shrinkage for each drug, thereby enhancing statistical power. We develop efficient Markov chain Monte Carlo algorithms for implementation and employ a false discovery rate (FDR)‐controlled, false negative rate (FNR)‐optimized hypothesis testing framework for AE signal detection. Extensive simulations demonstrate our methods' superior sensitivity—often surpassing existing approaches by a twofold or greater margin—while strictly controlling the FDR. An application to FDA FAERS data on statin drugs further highlights our methods' effectiveness in real‐world AE signal detection. Software implementing our methods is provided as supplementary material.
SiM Bayesian Methods Missing Data & Measurement Error Latent Variables & Mixtures · 发表于 2026-04-07,收录于 2026-05-01
Thomas Klausch, Birgit I. Lissenberg‐Witte, Veerle M. H. Coupé
摘要
Screening and surveillance programs for cancer, such as colorectal cancer (CRC), often yield electronic health records (EHR) of screening time, test results, and covariates. We consider EHR from CRC surveillance of individuals who have a high cancer risk due to their family history. These individuals, therefore, receive regular colonoscopies with the goal of finding and removing adenomas, precursor lesions to CRC. Our objective is to estimate time to adenoma incidence and explore associations with covariates. However, in doing so, several challenges of the CRC surveillance EHR have to be addressed. Importantly, the adenoma events are interval‐censored, meaning the exact event times are unknown and only fall within intervals defined by colonoscopy visits. Furthermore, colonoscopies can miss adenomas due to human or technical error, leading to misclassification of individuals with adenomas as adenoma‐free. Finally, the EHR data include individuals with adenomas at baseline, termed prevalent cases. This prevalence status may be unobserved if the baseline colonoscopy is missing or fails to detect existing adenomas. To address these challenges in the CRC EHR, and screening data in general, we develop a new prevalence‐incidence mixture model (PIM) with a Bayesian estimation back‐end through data augmentation and regularization priors. We show how to fit the model, estimate cumulative incidence functions, and evaluate model fit using information criteria as well as a non‐parametric estimator. In extensive simulations, we show good performance of the model when informative priors on the test sensitivity are provided, which is usually possible. An implementation in the R package BayesPIM is provided.
SiM Bayesian Methods High-Dimensional Genetics & Omics · 发表于 2026-04-06,收录于 2026-05-01
Wenting Liu, Lu Luo, Huiqiong Li, Niansheng Tang
摘要
High‐dimensional data is becoming increasingly prevalent in scientific fields such as genomics, economics, and medicine. However, recent studies have indicated that such data is often heterogeneous. Most existing research focuses either on high‐dimensional multi‐source data or on isolated heterogeneous datasets, leaving a significant gap for joint modeling and inference of high‐dimensional multi‐source heterogeneous data. In this paper, we employ a Bayesian method to address the estimation problem associated with high‐dimensional multi‐source heterogeneous data, aiming to extract shared features across all subpopulations while also examining the unique heterogeneity within each subpopulation. We introduce a scalable and interpretable Bayesian model for multi‐source heterogeneous linear data that employs a sparsity‐inducing spike‐and‐slab prior, featuring a Laplace slab and a Dirac spike. To address the computational challenges associated with the posterior, we implement a mean‐field variational approximation that utilizes a factorizable family of spike‐and‐slab distributions. Our method overcomes the high computational cost of Gibbs sampling while preserving valuable features, providing a posterior distribution for the parameters and offering a natural mechanism for variable selection via posterior inclusion probabilities. Through simulation studies and an application to five cancer datasets in real‐world the Cancer Genome Atlas (TCGA), we demonstrate the effectiveness of our variational Bayesian approach. Our results show the advantages of our method in terms of computational efficiency and scalability compared to Gibbs sampling method and penalized frequentist methods for integrative analysis, making it well‐suited for analyzing high‐dimensional multi‐source heterogeneous data. The proposed variational Bayesian algorithms have been implemented in the R package VBMS, which is publicly available on CRAN.
SiM Survival Analysis · 发表于 2026-04-06,收录于 2026-05-01
Antoniya Dineva, Oliver Kuss, Annika Hoyer
摘要
In cohort studies, the occurrence of a specific disease is often main interest. For proper modeling, death as a competing risk should be accounted for. Especially, the 'semi-competing' character of data has to be acknowledged: Disease occurrence (the non-terminal event) can be observed before death (terminal event), but not vice versa. Hence, the terminal event might censor the non-terminal event but the non-terminal remains observable if the terminal event occurs first. The underlying setting can be described by an illness-death model incorporating transitions between three states 'healthy,' 'diseased,' and 'dead.' This article aims to introduce a new method employing accelerated failure time (AFT) models for each of the transitions in the model. The major advantage of this approach is its intuitive and straightforward interpretation based on the survival time instead of the hazard function, facilitating communication of results. We propose a parametric model assuming Weibull distributions for ages in different states, modeling (1) age at disease onset, (2) age at death in patients without diagnosis, (3) age at death after diagnosis. We add random effects to adjust for intra-individual correlations, yielding a trivariate parametric model. Model parameters are estimated by maximum likelihood principle. The likelihood function incorporates left truncation for both events and the terminal event, reflecting the delayed entry into the study cohort. As in cohort studies diagnosis is usually made at intermittent follow-up visits and exact age at disease onset is only known to lie in interval between last two visits, we finally consider interval censoring for the occurrence of the non-terminal event. The method is illustrated using data from the PAQUID study, focusing on dementia as the non-terminal event. Our approach leads to plausible results, confirming the findings by other existing methods. The overall performance is assessed by simulation study, yielding promising results concerning the accuracy and numerical robustness of our model." (Reconstructed from the OpenAlex abstract index, since the Wiley page itself returned a 403 error.)
Biometrika Testing & Uncertainty · 发表于 2026-04-04
Yonghoon Lee, Edgar Dobriban, Eric J Tchetgen Tchetgen
摘要
Summary: We consider the problem of comparing a reference distribution with several other distributions. Given a sample from both the reference and the comparison groups, we aim to identify the comparison groups whose distributions differ from that of the reference group. Viewing this as a multiple-testing problem, we introduce a methodology that provides exact, distribution-free control of the false discovery rate. To do so, we introduce the concept of batch conformal p -values and demonstrate that they satisfy positive regression dependence across the groups Benjamini & Yekutieli (2001), thereby enabling control of the false discovery rate through the Benjamini–Hochberg procedure. The proof of positive regression dependence introduces a novel technique for the inductive construction of rank vectors with almost-sure dominance under exchangeability. We evaluate the performance of the proposed procedure through simulations. Despite being distribution-free, in some cases it shows performance comparable to methods with knowledge of the data-generating normal distribution, and it further has more power than direct approaches based on conformal out-of-distribution detection. Furthermore, we illustrate our methods on a hepatitis C treatment dataset, where they identify patient groups with large treatment effects, and on the Current Population Survey dataset, where they identify subpopulations with long working hours.
SiM Diagnostics & Prediction Missing Data & Measurement Error · 发表于 2026-04-03,收录于 2026-05-01
Runjia Zou, Brian D Williamson, Susan M Shortreed, R Yates Coley
摘要
Electronic health records (EHRs) provide a rich data source for building prediction models to improve the quality of care. However, EHR data are prone to measurement error, and outcomes used to evaluate prediction model performance may be misclassified. Comparing risk predictions to misclassified outcomes will result in unreliable estimates of a prediction model's performance. We propose a method that leverages a smaller chart review sample with gold‐standard outcome measurements to adjust validation for outcome misclassification and provide more accurate prediction model evaluation. We derive formulae to estimate true positive rate (TPR), false positive rate (FPR), positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve (AUC) in the presence of outcome misclassification. Different scenarios of misclassification are explored, including when the misclassification is independent or dependent on features, and when misclassification is unidirectional (e.g., only missed diagnoses) or bidirectional. In simulation studies, we compare the bias and 95% confidence interval coverage of performance estimates obtained using our proposed method to those estimated with misclassified outcomes (without accounting for misclassification) or in the smaller chart review sample. Simulation results indicate that, across all misclassification scenarios examined, our proposed estimates have good accuracy and improved precision. Outcome misclassification should be considered when evaluating a prediction model's performance in order to accurately inform decision‐making about whether and how to use a clinical prediction model.
SiM Optimization & Learning Theory · 发表于 2026-04-02,收录于 2026-05-01
Puyuan Liu, Donatello Telesca, Abigail Dickinson, Shafali Jeste, Damla Şentürk
摘要
Functional boxplots are an informative exploratory tool for visualizing functional data and providing a concise summary of a dataset's central tendency, spread, and potential outliers. Motivated by within‐ and across‐sample visualization of electroencephalogram (EEG) power spectral densities in autism studies, we propose Wasserstein boxplots for summarizing a sample of densities. Densities take on nonnegative values and integrate to 1. In order to take into account the constraints of the signal and to avoid metric distortions associated with transformations, the proposed Wasserstein boxplots utilize the 2‐Wasserstein metric for quantifying distances between densities. In addition, the proposed boxplots extend the traditional approach of summarizing central tendency within a single dataset to comparative settings where central tendency of data within a target sample is quantified with respect to a reference group. Cross‐sample Wasserstein boxplots are motivated by quantification of deviations in power spectra of autistic children (target group) from neurotypical development (reference group). Finally, covariate‐adjusted boxplots are proposed for quantifying deviations in the target group from the Fréchet mean in the reference group, conditional on covariates. A unique feature of EEG power spectra is the peak alpha frequency (PAF), which shifts to higher frequencies as children age. Hence, covariate‐adjusted boxplots are used to quantify deviations in the autistic sample from neurotypical spectra, conditional on age. The proposed exploratory tools, especially the comparative analyses are applicable more broadly, beyond the motivating autism context, to studies involving both a target and a reference group.
SMMR Survival Analysis Clinical Trials Causal Inference · 发表于 2026-04-02
Andrea Callegaro, Nathan W. Bean
摘要
The ICH E9(R1) addendum stresses the importance of clearly pre-specifying clinically interpretable treatment effect measures (estimands) and proposes different strategies to deal with intercurrent events. In this paper, we consider different estimands of the vaccine efficacy that are interpretable even when the proportional hazard assumption is not met, including estimands based on the average hazard ratio, the cumulative incidence ratio, and the restricted mean time lost ratio. Under the assumption that different estimands target relevant clinical questions, power becomes a crucial factor in choosing the preferred estimand. We focus on the power of these different estimands in vaccine efficacy trials, and we illustrate them in a human papillomavirus phase III vaccine trial. In classical settings of prophylactic vaccine efficacy trials, our results show that the average hazard ratio and cumulative incidence ratio give similar results, while the restricted mean time lost ratio is expected to be less powerful.
SiM Diagnostics & Prediction · 发表于 2026-04-02,收录于 2026-05-01
Kehao Zhu, Kwun Chuen Gary Chan, Ying‐Qi Zhao, Yingye Zheng
摘要
Accurate classifiers that use novel biomarkers and readily available predictors significantly enhance decision‐making in various clinical scenarios, such as assessing the need for biopsies in cancer diagnosis. When classification performance is limited, a decision framework can be applied to rule in or rule out invasive diagnostic procedures while incorporating a neutral zone for indeterminate classifications. Building on this framework, we propose a new family of two‐step classifiers that selectively use costly biomarker testing for a targeted subset of individuals undergoing multiple evaluations. The optimal solution expands upon the Neyman‐Pearson Lemma, highlighting a vital trade‐off between the costs of expensive biomarker measurements and improving classification performance while minimizing uncertainty in the decision process. We demonstrate the practical utility of our approach through a biomarker study focused on prostate cancer diagnosis.
JASA Other · 发表于 2026-04-02
Dohyeong Ki, Adityanand Guntuboyina
摘要
Shape constraints in nonparametric regression provide a powerful framework for estimating regression functions under realistic assumptions without tuning parameters. However, most existing methods—except additive models—impose too weak restrictions, often leading to overfitting in high dimensions. Conversely, additive models can be too rigid, failing to capture covariate interactions. This paper introduces a novel multivariate shape-constrained regression approach based on total concavity, originally studied by T. Popoviciu. Our method allows interactions while mitigating the curse of dimensionality, with convergence rates that depend only logarithmically on the number of covariates. We characterize and compute the least squares estimator over totally concave functions, derive theoretical guarantees, and demonstrate its practical effectiveness through empirical studies on real-world datasets. *(abstract via OpenAlex)*
JASA Econometrics & Economics · 发表于 2026-04-02
Wenhao Cui
摘要
High-frequency financial data are often contaminated by market microstructure effects. In this study, we consider a setting where a portion of the microstructure noise can be explained by observable trading information, referred to as the explicative noise component. To formally analyze this component, we first develop a model-free variable importance measure in the high-frequency setting that quantifies the price impact of subsets of trading variables. Based on the identified significant variables, we then introduce a nonparametric estimator for the explicative noise and establish its asymptotic properties. The finite-sample performance of the proposed methods is assessed through Monte Carlo simulations calibrated to real data. Finally, an empirical application shows that the explicative noise component plays a key role in explaining return variation, and that accounting for it substantially smooths the volatility signature curve. Supplementary materials for this article are available online. *(abstract via OpenAlex)*
SMMR Survival Analysis Causal Inference · 发表于 2026-04-02
Yan-Lin Chen, Hsiang-Hsi Hung, Sheng-Hsuan Lin
摘要
The Cox proportional hazards model has popularized the conventional hazard ratio as a standard measure for assessing the effect of exposure on time-to-event outcomes. However, as noted in Hernán's influential critique, interpreting the hazard ratio as a causal effect can introduce substantial selection bias. While Hernán's work discusses this issue qualitatively, it lacks a method to quantify the extent of bias or the strength an unmeasured confounder would need to account for it. Our study fills this gap by proposing a nonparametric sensitivity analysis framework, inspired by Ding and VanderWeele, which quantifies “how hazardous the hazard ratio is.” Using the causal hazard ratio proposed by Aalen as the “true hazard ratio,” we derive an upper bound for the bias in the conventional hazard ratio (termed HR bias) in both discrete and continuous time settings, demonstrating that this bias accumulates over time. Additionally, we derive the E -value for HR bias, allowing researchers to evaluate whether unmeasured confounding alone could fully explain the conventional hazard ratio under a certain level of the causal hazard ratio. This approach provides researchers with a robust sensitivity analysis framework, reducing reliance on restrictive assumptions and enhancing the reliability of causal interpretations in survival analysis.
JASA Testing & Uncertainty · 发表于 2026-04-02
Paula Gablenz, Matteo Sesia, Tianshu Sun, Chiara Sabatti
摘要
We introduce local conditional hypotheses that express how the relation between explanatory variables and outcomes changes across different contexts, described by covariates. By expanding upon the model-X knockoff filter, we show how to adaptively discover these local associations, all while controlling the false discovery rate. Our enhanced inferences can help explain sample heterogeneity and uncover interactions, making better use of the capabilities offered by modern machine learning models. Specifically, our method is able to leverage any model for the identification of data-driven hypotheses pertaining to different contexts. Then, it rigorously test these hypotheses without succumbing to selection bias. Importantly, our approach is efficient and does not require sample splitting. We demonstrate the effectiveness of our method through numerical experiments and by studying the genetic architecture of Waist-Hip-Ratio across different sexes in the UKBiobank. *(abstract via OpenAlex)*
SiM Nonparametric & Semiparametric Survival Analysis · 发表于 2026-04-02,收录于 2026-05-01
Spencer Matthews, Bin Nan
摘要
Cohort studies of the onset of a disease often encounter left‐truncation on the event time of interest in addition to right‐censoring due to variable enrollment times of study participants. Analysis of such event time data can be biased if left‐truncation is not handled properly. We propose a semiparametric sieve likelihood approach for fitting a linear regression model to data where the response variable is subject to both left‐truncation and right‐censoring. We show that the estimators of regression coefficients are consistent, asymptotically normal and semiparametrically efficient. Extensive simulation studies show the effectiveness of the method across a wide variety of error distributions. We further illustrate the method by analyzing datasets from the Canadian Study of Health and Aging and The 90+ Study for aging and dementia.
SiM Survival Analysis Testing & Uncertainty · 发表于 2026-04-02,收录于 2026-05-01
Taylor Krajewski, Gary Koch
摘要
This paper introduces a randomization‐based method for covariate‐adjusted comparisons of restricted mean survival time (RMST) between treatment arms in randomized controlled trials. Existing approaches for covariate‐adjusted RMST analysis have model‐based assumptions that may not be compatible with the complexity of survival data. We estimate the treatment difference in RMST using randomization‐based analysis of covariance (RB‐ANCOVA) for categorized time‐to‐event data by constraining the covariate mean differences between treatment groups to zero. Accordingly, we provide corresponding confidence intervals that offer greater precision than those based on unadjusted RMST differences. The methodology is detailed for comparing two treatment groups for a single or multiple time intervals, as well as for multiple treatment groups over a single interval. We demonstrate the application of these methods using data from a randomized, double‐blind, placebo‐controlled clinical trial evaluating three doses of a test treatment for ALS.