Article
Mathematical & Computational Biology
Peijin Wang, Shein-Chung Chow
Summary: This article discusses methods for sample size re-estimation in clinical trials, including the adjusted effect size (AES) approach and the iterated expectation/variance (IEV) approach, which take into account the variability of observed responses. Results show that the IEV approach generally performs best in controlling type I error inflation, but may lead to a larger increase in sample size when detecting smaller effect sizes.
STATISTICS IN MEDICINE
(2021)
Article
Pharmacology & Pharmacy
Chengxue Zhong, Qing Li, Liwen Wu, Jianchang Lin
Summary: Platform design for exploring multiple drugs concurrently is crucial for drug development efficiency. A major challenge is the lack of data for interim decisions, particularly in treatment arm selection and sample size determination. The modified conditional power method utilizes data from both primary and surrogate endpoints for interim analysis.
JOURNAL OF BIOPHARMACEUTICAL STATISTICS
(2022)
Article
Medicine, Research & Experimental
Ruitao Lin, Zhao Yang, Ying Yuan, Guosheng Yin
Summary: The heterogeneity of clinical trial participants poses a fundamental challenge in the field of precision medicine, but adaptive enrichment designs offer a flexible and intuitive solution. By enriching the subgroup of trial participants with a higher likelihood of benefit from a new treatment, these designs can control type I error rate and improve statistical power and expected sample size.
CONTEMPORARY CLINICAL TRIALS
(2021)
Article
Mathematical & Computational Biology
Liwen Wu, Qing Li, Mengya Liu, Jianchang Lin
Summary: Adaptive subgroup enrichment design is an efficient framework that allows accelerated development for investigational treatments and offers flexibility in population selection. This article improves the interim decision making process by incorporating information from surrogate endpoints and prior knowledge, aiming to overcome the challenge of immature data in adaptive designs.
STATISTICS IN BIOPHARMACEUTICAL RESEARCH
(2022)
Article
Multidisciplinary Sciences
Amrit Sudershan, Kanak Mahajan, Rakesh K. Panjaliya, Manoj K. Dhar, Parvinder Kumar
Summary: Sampling methods for studying population behavior are uncertain and require representative samples for generalization. Sample size greatly affects the detection of research effects, with smaller samples having lower statistical power and higher risk of missing underlying differences. This study provides a calculation method for determining sample availability during research.
SCIENTIFIC REPORTS
(2022)
Article
Multidisciplinary Sciences
Neil K. Chada, Ajay Jasra, Fangyuan Yu
Summary: This article considers the development of unbiased estimators for the Hessian of the log-likelihood function in partially observed diffusion processes. It provides an unbiased estimator based on Girsanov's Theorem and randomization schemes. The developed estimator is tested and verified against a newly proposed particle filtering methodology.
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2022)
Article
Engineering, Civil
Shuoyu Liu, Ying Luo, Liuliu Peng, Yan Jiang, Ercong Meng, Bo Li
Summary: This study develops a nonparametric wind pressure interpolation method based on wind tunnel test data, and predicts the probability density function of wind pressure coefficient features at target locations. The results can be presented through deterministic single values and probabilistic prediction intervals.
JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS
(2022)
Article
Engineering, Electrical & Electronic
Francesco Grassi, Angelo Coluccia
Summary: This study addresses the challenging problem of distribution-agnostic linear (weighted) unbiased estimation of a global parameter from heterogeneous and unbalanced data. It proposes a family of estimators based on trimmed weights and analyzes different criteria for the cut-off threshold. It also derives and analyzes estimators for hyperparameters.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2023)
Article
Ecology
Beth E. Ross, Mitch D. Weegman
Summary: Understanding mechanistic causes of population change is critical for managing and conserving species. Integrated population models (IPMs) allow for quantifying population changes while directly relating environmental drivers to vital rates. The study found that the temporal duration of a study and effect size had the greatest influence on the power to identify trends in adult survival and fecundity. IPMs had greater power to identify trends and environmental effects on vital rates compared to traditional analysis methods.
ECOLOGICAL APPLICATIONS
(2022)
Article
Social Sciences, Mathematical Methods
Marcos Cruz, Javier Gonzalez-Villa
Summary: Population sizing is crucial in various fields, and gigapixel cameras offer high-resolution images of entire populations. A method based on geometric sampling has been proposed to reduce manual counting and achieve relative standard errors of 5-10%. By projecting the sampling grid onto gigapixel images using camera projection, perspective effects can be neutralized, restoring relative standard errors to the 5-10% range.
SOCIOLOGICAL METHODS & RESEARCH
(2021)
Article
Mathematical & Computational Biology
Kevin Kunzmann, Michael J. Grayling, Kim May Lee, David S. Robertson, Kaspar Rufibach, James M. S. Wason
Summary: Adapting the final sample size of a trial based on accrued information is a natural way to address planning uncertainty. This study reviews and compares common approaches to estimating conditional power and discusses the connection between heuristic sample size recalculation and optimal two-stage designs. It proposes alternative methods to react to newly emerging trial-external evidence in a way that is consistent with the originally planned design, aiming to avoid design inconsistencies.
STATISTICS IN MEDICINE
(2022)
Article
Medicine, Research & Experimental
Yi Liu, Heng Xu
Summary: The paper discusses three main approaches to address the issue of inflated Type I error rates in sample size re-estimation designs, namely combination test, conditional error, and conventional test with sample size increase in the allowable region. Although traditional test guarantees Type I error rate control, it often results in lower power compared to the corresponding combination test approach.
CONTEMPORARY CLINICAL TRIALS
(2021)
Article
Health Care Sciences & Services
Faris F. F. Gulamali, Ashwin S. S. Sawant, Patricia Kovatch, Benjamin Glicksberg, Alexander Charney, Girish N. N. Nadkarni, Eric Oermann
Summary: Sample size estimation is a crucial step in experimental design, but it is understudied in the context of deep learning. This study focuses on estimating the minimum sample size of labeled training data necessary for training computer vision models. The researchers used autoencoder loss to estimate the minimum converging sample (MCS) for fully connected neural network classifiers. The findings suggest that MCS and convergence estimation are promising methods to guide sample size estimates for data collection and labeling prior to training deep learning models in computer vision.
NPJ DIGITAL MEDICINE
(2022)
Article
Mathematical & Computational Biology
Masataka Igeta, Shigeyuki Matsui
Summary: Blinded sample size re-estimation (BSSR) is an adaptive design that prevents power reduction caused by misspecifications of nuisance parameters in sample size calculation for comparative clinical trials. This article proposes a robust BSSR method that can handle misspecifications of the working variance function and shows through simulation studies that it has relatively stable power.
STATISTICS IN MEDICINE
(2022)
Article
Mathematical & Computational Biology
Jianrong Wu, Yimei Li, Liang Zhu
Summary: A multi-arm trial allows simultaneous comparison of multiple experimental treatments with a common control, improving efficiency compared to traditional randomized controlled trials. This paper presents a group sequential multi-arm multi-stage (MAMS) trial design based on the sequential conditional probability ratio test, providing analytical solutions for futility and efficacy boundaries for an arbitrary number of stages and arms.
STATISTICS IN MEDICINE
(2023)
