Article
Computer Science, Theory & Methods
Mohammad W. Hattab, David Ruppert
Summary: The paper proposes an approach to handle measurement errors in predictors when modeling flexible regression functions by directly modeling the mean and variance of the response variable after integrating out the true unobserved predictors in a penalized splines model. Simulation studies show that this approach provides satisfactory prediction accuracy, outperforming local polynomial estimators and being competitive with the Bayesian estimator even when the model is incorrectly specified.
STATISTICS AND COMPUTING
(2021)
Article
Pharmacology & Pharmacy
Lu Yang, Hui Song, Yingwei Peng, Dongsheng Tu
Summary: A joint model was proposed for analyzing quality of life scores and survival times in a clinical trial on early breast cancer, featuring a mixed effect model and a semiparametric mixture cure model linked by shared random effects. An EM algorithm was used for parameter estimation and performance evaluation was conducted through simulation studies and application to real data.
PHARMACEUTICAL STATISTICS
(2021)
Article
Mathematical & Computational Biology
An-Ming Tang, Cheng Peng, Niansheng Tang
Summary: This paper proposes a novel JMLS method for handling multivariate longitudinal and bivariate correlated survival data. Nonparametric marginal survival hazard functions are transformed to bivariate normal random variables, and Bayesian penalized splines are used to approximate unknown baseline hazard functions. By incorporating the Metropolis-Hastings algorithm into the Gibbs sampler, a Bayesian adaptive Lasso method is developed to simultaneously estimate parameters and baseline hazard functions, and select important predictors in the considered JMLS.
STATISTICS IN MEDICINE
(2023)
Article
Computer Science, Interdisciplinary Applications
Xiaoyu Liu, Liming Xiang
Summary: This study proposes a more flexible class of generalized accelerated hazards mixture cure models for analyzing interval-censored failure times. A sieve maximum likelihood estimation method is used to approximate the unknown cumulative baseline hazard function with B-splines. Simulation results demonstrate satisfactory performance of the proposed method in finite samples.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2021)
Article
Biology
Yuyan Wang, Akhgar Ghassabian, Bo Gu, Yelena Afanasyeva, Yiwei Li, Leonardo Trasande, Mengling Liu
Summary: This paper introduces a new method, semiparametric partial-linear single-index DL quantile regression, to study the effects of time-dependent exposure mixtures on different quantiles of outcome. Through simulations and application studies, the performance and value of this method are demonstrated.
Article
Mathematical & Computational Biology
Erjia Cui, E. Christi Thompson, Raymond J. Carroll, David Ruppert
Summary: The study develops a generalized partially additive model for assessing physical activity across multiple populations, tackling challenges posed by the nonlinear relationship between physical behaviors and health outcomes. By modeling each score component as a smooth term and using penalized splines, two inferential methods are proposed to address computational problems, with both exhibiting accurate performance in simulations. Applied to a national survey, the models quantify nonlinear and interpretable shapes of score components for all-cause mortality.
STATISTICS IN MEDICINE
(2022)
Article
Biology
Weibin Zhong, Guoqing Diao
Summary: This paper proposes a joint semiparametric modeling approach for efficient regression analysis of two-phase study data. The approach combines semiparametric models for the survival outcome and the expensive exposures, and exhibits good performance and robustness.
Article
Geriatrics & Gerontology
M. J. Kim, S. Y. Jang, H. -K. Cheong, In-Hwan Oh
Summary: The study revealed that frailty and each of its eight symptoms were associated with increased healthcare costs among South Korean older adults, advocating for the need to identify and manage frailty to reduce healthcare costs.
JOURNAL OF NUTRITION HEALTH & AGING
(2021)
Article
Statistics & Probability
Matthieu Marbac, Mohammed Sedki, Christophe Biernacki, Vincent Vandewalle
Summary: This study focuses on parameter estimation in regression models with missing group variables, proposing a simultaneous estimation approach for clustering and regression. Numerical experiments and real data analysis illustrate the effectiveness of the new method.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2022)
Article
Mathematical & Computational Biology
Byung-Jun Kim, Inyoung Kim
Summary: Variable selection and graphical modeling are crucial in analyzing highly correlated and high-dimensional data. Gaussian graphical models have limitations in handling nonadditive, nonparametric regression with high-dimensional variables. This paper proposes a joint semiparametric kernel network regression method to address this limitation and provide a connection between variable selection and graphical modeling.
STATISTICS IN MEDICINE
(2023)
Review
Psychology, Multidisciplinary
Debasmita Dey, Pradeep Kumar
Summary: This study fills an important research gap by exploring and validating the role of four psychological attributes in review rating prediction. Tobit regression is used to investigate the relationship between psychological attributes and review rating, and coefficients and p-value statistics are used for experimental validation. Amazon datasets of twenty categories are chosen for establishing relationships and predicting the performance of review rating prediction. The study identifies U-shaped and Inverted U-shaped relationships between attributes and review rating, using Yerkes-Dodson law and diminishing marginal utility theory to explain these curvilinear relationships.
COMPUTERS IN HUMAN BEHAVIOR
(2023)
Article
Economics
Fang Fang, Jialiang Li, Xiaochao Xia
Summary: This paper proposes a semiparametric model averaging prediction (SMAP) method for a dichotomous response, which approximates the unknown score function using a linear combination of one-dimensional marginal score functions. The weight parameters are obtained by smoothing the nonparametric marginal scores and applying parametric model averaging. SMAP provides greater flexibility than parametric models and stability compared to fully nonparametric approaches. Theoretical properties are investigated in two practical scenarios, and empirical evidences support the effectiveness of the proposed method.
JOURNAL OF ECONOMETRICS
(2022)
Article
Biology
Bo Han, Ingrid Van Keilegom, Xiaoguang Wang
Summary: This article introduces a statistical method that utilizes external information to efficiently synthesize auxiliary survival information and propose a semiparametric estimation method for the combined empirical likelihood in the nonmixture cure model framework to enhance inference about the associations between exposures and disease outcomes. By incorporating auxiliary survival information, the method improves estimation efficiency and prediction accuracy, reduces computational burden, and maintains shape constraints on the baseline distribution function, resulting in strongly consistent and asymptotically normal estimations. Simulation studies demonstrate large gains in efficiency, contributing to biomarker evaluation and treatment effect analysis within smaller studies.
Article
Mathematical & Computational Biology
Frederico M. Almeida, Enrico A. Colosimo, Vinicius D. Mayrink
Summary: The paper focuses on addressing parameter estimation issues in cure fraction models by introducing the Firth correction method to avoid nonfinite estimates resulting from the monotonicity of likelihood functions. Through simulation studies, it is found that the performance of coefficients related to binary covariates is strongly influenced by the degree of imbalance.
BIOMETRICAL JOURNAL
(2022)
Article
Mathematical & Computational Biology
Elaheh Talebi-Ghane, AhmadReza Baghestani, Farid Zayeri, Virgine Rondeau, Ali Akhavan
Summary: The study proposed a joint frailty model with cure fraction for analyzing recurrent and terminal events, estimating the effect of covariates on the cure rate and the two events concurrently.
BIOMETRICAL JOURNAL
(2021)
Article
Biology
Dong Liu, Changwei Zhao, Yong He, Lei Liu, Ying Guo, Xinsheng Zhang
Summary: This paper proposes a method for clustering and estimating heterogeneous graphs in fMRI data, achieving good results by fully exploiting the group differences of conditional dependence relationships among brain regions. The method constructs individual-level between-region network measures and uses a modified difference of convex programming with the alternating direction method of multipliers (DC-ADMM) algorithm to solve the optimization problem.
Editorial Material
Oncology
Cathy J. Bradley, Ya-Chen Tina Shih, K. Robin Yabroff
JOURNAL OF CLINICAL ONCOLOGY
(2023)
Article
Oncology
Meng Li, Kaiping Liao, Alice J. Chen, Tina Cascone, Yu Shen, Qian Lu, Ya-Chen Tina Shih
Summary: Nationwide, there is evidence to suggest that metastatic lung cancer patients residing in counties with a higher percentage of racialized population experience slower initiation of immune checkpoint inhibitor (ICI) therapy despite having a higher density of medical oncologists in their neighborhood.
JNCI-JOURNAL OF THE NATIONAL CANCER INSTITUTE
(2023)
Article
Mathematical & Computational Biology
Xiaogang Su, Youngjoo Cho, Liqiang Ni, Lei Liu, Elise Dusseldorp
Summary: Moderation analysis is crucial in precision medicine research. By exchanging the roles of outcome and treatment variable, equivalent estimation of heterogeneous treatment effects can be achieved in logistic regression models. This study establishes the joint asymptotic normality for the two estimators, enabling refined inference in moderation analysis.
STATISTICS IN MEDICINE
(2023)
Article
Oncology
Jennifer I. Vidrine, Steven K. Sutton, David W. Wetter, Ya-Chen Tina Shih, Lois M. Ramondetta, Linda S. Elting, Joan L. Walker, Katie M. Smith, Summer G. Frank-Pearce, Yisheng Li, Sarah R. Jones, Darla E. Kendzor, Vani N. Simmons, Damon J. Vidrine
Summary: The purpose of this study was to evaluate the long-term efficacy of Motivation And Problem Solving (MAPS), a novel treatment well-suited to meeting the smoking cessation needs of women who smoke and have a history of cervical intraepithelial neoplasia (CIN) or cervical cancer. It was found that MAPS led to a greater than two-fold increase in smoking abstinence among survivors of CIN and cervical cancer at 12 months, but the effect was no longer significant at 18 months.
JOURNAL OF CLINICAL ONCOLOGY
(2023)
Editorial Material
Oncology
Ya-Chen Tina Shih, Cathy Bradley, K. Robin Yabroff
JNCI-JOURNAL OF THE NATIONAL CANCER INSTITUTE
(2023)
Article
Oncology
Mariana Chavez-MacGregor, Xiudong Lei, Catalina Malinowski, Hui Zhao, Ya-Chen Shih, Sharon H. Giordano
Summary: This study used the National Cancer Database to investigate the impact of Medicaid expansion on the timing and delays of adjuvant chemotherapy among early-stage breast cancer patients. The results showed that after Medicaid expansion, the proportion of Black and Hispanic patients experiencing delays in chemotherapy initiation decreased, narrowing the racial disparities.
JNCI-JOURNAL OF THE NATIONAL CANCER INSTITUTE
(2023)
Article
Oncology
Vishal R. Patel, Thomas B. Cwalina, Arjun Gupta, Nico Nortje, Samyukta Mullangi, Ravi B. Parikh, Ya-Chen Tina Shih, S. M. Qasim Hussaini
Summary: In this cross-sectional study, oncologist participation and performance in the 2019 MIPS were examined. Oncologist participation rate was found to be low (86%), compared to the overall participation rate (97%). It was also observed that oncologists using alternative payment models (APMs) as their filing source had higher MIPS scores, indicating the importance of organizational resources for participants.
Review
Health Care Sciences & Services
Jingxia Liu, Lei Liu, Aimee S. James, Graham A. Colditz
Summary: Cluster randomized trial design is expensive and it is important to develop an optimal design to minimize costs. Local optimal designs aim to minimize the variance of the treatment effect under a fixed budget. This requires the input of an association parameter and involves the consideration of enrollment feasibility. Our contributions include summarizing available local optimal designs and proposing new designs for different scenarios, as well as developing Statistical Analysis System (SAS) macros for all optimal designs.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2023)
Article
Mathematical & Computational Biology
Lili Liu, Kevin He, Di Wang, Shujie Ma, Annie Qu, Lu Lin, J. Philip Miller, Lei Liu
Summary: There is a growing research interest in evaluating the performance of healthcare centers based on patient outcomes. Conventional assessments can be done using fixed or random effects models, like provider profiling. A new method is proposed using fusion penalty to cluster healthcare centers based on survival outcomes, providing a data-driven approach for grouping without prior knowledge. An efficient algorithm is developed to implement the proposed method, and its validity is demonstrated through simulation studies and application to kidney transplant registry data.
STATISTICS IN MEDICINE
(2023)
Article
Oncology
Vishal R. Patel, Thomas B. Cwalina, Nico Nortje, Samyukta Mullangi, Ravi B. Parikh, Ya-Chen Tina Shih, Arjun Gupta, S. M. Qasim Hussaini
Summary: The Merit-Based Incentive Payment System (MIPS) is the only federally mandated value-based payment model for oncologists. The inclusion of cost measures in MIPS may disproportionately affect oncologists, who have higher costs of care compared to other specialties. This study examines the implications of incorporating cost measures on physician reimbursements and highlights the need for specialty-specific recalibration to ensure fairness and preserve healthcare quality.
JCO ONCOLOGY PRACTICE
(2023)
Article
Oncology
Andrew M. D. Wolf, Kevin C. Oeffinger, Tina Ya-Chen Shih, Louise C. Walter, Timothy R. Church, Elizabeth T. H. Fontham, Elena B. Elkin, Ruth D. Etzioni, Carmen E. Guerra, Rebecca B. Perkins, Karli K. Kondo, Tyler B. Kratzer, Deana Manassaram-Baptiste, William L. Dahut, Robert A. Smith
Summary: Lung cancer is the leading cause of cancer-related deaths and years of life lost in the US. Early detection through screening has been shown to reduce mortality. The American Cancer Society has updated its guidelines for lung cancer screening, recommending annual low-dose CT screening for individuals aged 50-80 who currently smoke or formerly smoked and have a significant smoking history.
CA-A CANCER JOURNAL FOR CLINICIANS
(2023)
Editorial Material
Oncology
Marcelo Coca Perraillon, Ya-Chen Tina Shih
JOURNAL OF CLINICAL ONCOLOGY
(2023)
Article
Mathematical & Computational Biology
Yiming Shi, Huilin Li, Chan Wang, Jun Chen, Hongmei Jiang, Ya-Chen T. Shih, Haixiang Zhang, Yizhe Song, Yang Feng, Lei Liu
Summary: In this article, a flexible model for microbiome count data is proposed. The model is based on a quasi-likelihood framework, which does not assume any specific distribution for the microbiome count but assumes the variance as an unknown but smooth function of the mean. Simulation studies demonstrate that the flexible quasi-likelihood method provides valid inferential results compared to the negative binomial generalized linear model (GLM) and Poisson GLM. The utility of the method is further demonstrated using a real microbiome study on the relationship between adenomas and microbiota. An R package, fql, is provided for applying the method.
STATISTICS IN MEDICINE
(2023)
Article
Mathematical & Computational Biology
Fang Niu, Cheng Zheng, Lei Liu
Summary: Recurrent events and terminal events are often related in biomedical studies. Although joint models have been proposed to analyze their correlation, there is a lack of suitable methods to investigate the causal mechanisms between specific exposures, recurrent events, and terminal events.
STATISTICS IN MEDICINE
(2023)