Article
Health Care Sciences & Services
Angelika Geroldinger, Rok Blagus, Helen Ogden, Georg Heinze
Summary: In binary logistic regression, separable data refers to the existence of a linear combination of explanatory variables that perfectly predicts the outcome. Firth's logistic regression (FL) is a popular solution to obtain finite estimates in such cases. When analyzing clustered data, like in clinical research, using generalized estimating equations (GEE), convergence becomes more complicated. This article investigates extensions of FL to GEE and compares their convergence behavior and performance using simulated and real data.
BMC MEDICAL RESEARCH METHODOLOGY
(2022)
Article
Medicine, General & Internal
William J. Hulme, Elizabeth Williamson, Elsie M. F. Horne, Amelia Green, Helen I. I. McDonald, Alex J. J. Walker, Helen J. J. Curtis, Caroline E. Morton, Brian MacKenna, Richard Croker, Amir Mehrkar, Seb Bacon, David Evans, Peter Inglesby, Simon Davy, Krishnan Bhaskaran, Anna Schultze, Christopher T. Rentsch, Laurie Tomlinson, Ian J. Douglas, Stephen J. W. Evans, Liam Smeeth, Tom Palmer, Ben Goldacre, Miguel A. Hernan, Jonathan A. C. Sterne
Summary: The COVID-19 vaccines were developed and evaluated through randomized trials, but important questions remain unanswered. Observational studies and target trial emulation can provide valuable insights, although potential biases need to be managed. This article presents two approaches to emulate target trials using observational data.
ANNALS OF INTERNAL MEDICINE
(2023)
Article
Nutrition & Dietetics
Francesco Campa, Catarina Matias, Filipe Teixeira, Joana Reis, Maria Valamatos, Giuseppe Coratella, Cristina Monteiro
Summary: This study aimed to compare the accuracy of athlete-specific equations and generalized equations in estimating fat-free mass (FFM) in resistance-trained exercisers. The results showed that the athlete-specific equation was more accurate in assessing FFM in resistance-trained exercisers, while it underestimated FFM in the general population.
Article
Mathematical & Computational Biology
Linda J. Harrison, Rui Wang
Summary: The study proposes power calculation methods for stepped-wedge cluster randomized trials using a logistic model fitted by generalized estimating equations. It shows that power based on a logistic model is lower than that from assuming a linear model in the presence of period effects, and evaluates the impact of background prevalence changes over time on power. Additionally, the methods are generalized to complex correlation structures to allow for changing correlations over time and with treatment status.
STATISTICS IN MEDICINE
(2021)
Article
Environmental Sciences
Ronald E. McRoberts, Erik Naesset, Zhengyang Hou, Göran Stahl, Svetlana Saarela, Jessica Esteban, Davide Travaglini, Jahangir Mohammadi, Gherardo Chirici
Summary: When probability samples are not available, the model-based framework and resampling methods like the bootstrap become essential options for constructing prediction intervals and obtaining standard errors. This study aims to develop a procedure for terminating resampling that ensures the stability of standard error estimation using bootstrap with one million replications. The primary contribution of this study is the development and demonstration of criteria for terminating resampling to stabilize the bootstrap estimate of the standard error.
REMOTE SENSING OF ENVIRONMENT
(2023)
Article
Environmental Sciences
Karen L. Neely, Colin P. Shea, Kevin A. Macaulay, Emily K. Hower, Michelle A. Dobler
Summary: The topical application of amoxicillin was found to be highly effective at halting SCTLD lesions, with significantly greater effectiveness compared to chlorine treatment. While reinfection rates varied by species and geographic regions, all sites and species showed a decreased probability of reinfection with time since initial treatment.
FRONTIERS IN MARINE SCIENCE
(2021)
Article
Green & Sustainable Science & Technology
Hitalo Tobias Lobo Lopes, Luis Rodrigo Fernandes Baumann, Paulo Sergio Scalize
Summary: In Brazilian rural communities, the contamination of dug shallow well (DSW) water is predicted using a generalized linear model. The model considers factors such as water quality, distance to contamination sources, structural conditions, and local geology. The final model includes variables such as well diameter, contour paving width, and the existence of poultry and swine husbandry. The model has an accuracy of 82.61%, with a true positive predictor of 82.18% and a negative predictor of 85.71%.
Article
Chemistry, Analytical
Svjatoslavs Kistkins, Timurs Mihailovs, Sergejs Lobanovs, Valdis Pirags, Harald Sourij, Othmar Moser, Dmitrijs Bliznuks
Summary: This study investigated the effectiveness of three different predictive models for glucose level classification and found that logistic regression model performs best for short-term glucose prediction, while LSTM model performs best for long-term glucose prediction.
Article
Management
Anna Crisci
Summary: This paper describes and illustrates the application of generalized estimation equations and several diagnostic measures. The generalized estimating equations generalize and extend the likelihood score equation for a generalized linear model by including the covariance matrix of the clustered responses. The paper investigates various measures for the strength of association between a response variable and covariates, and considers diagnostic measures for checking the adequacy of the generalized estimating equations method.
JOURNAL OF THE OPERATIONAL RESEARCH SOCIETY
(2023)
Article
Thermodynamics
Shahadat Hosan, Md Matiar Rahman, Shamal Chandra Karmaker, Andrew J. Chapman, Bidyut Baran Saha
Summary: This paper investigates the impact of remittances on multidimensional energy poverty in Bangladesh using a nationally representative Household Income and Expenditure Survey. The findings suggest that households with remittance income have a significantly lower level of energy poverty. The authors argue that increasing remittance inflow can contribute to alleviating energy poverty in Bangladesh and other developing nations, and therefore, national programs should be established to promote migrant workers, reduce household energy costs, and invest in modern energy technologies.
Article
Medicine, General & Internal
Claudio Fanconi, Anne de Hond, Dylan Peterson, Angelo Capodici, Tina Hernandez-Boussard
Summary: Machine learning predictions are being integrated into medical practice, and Bayesian logistic regression models offer better uncertainty estimation for risk predictions in cancer patients. This study found that these models perform similarly to standard logistic regression models and can identify patient subgroups with higher uncertainty.
Article
Physics, Multidisciplinary
Ayman Alzaatreh, Mohammad Aljarrah, Ayanna Almagambetova, Nazgul Zakiyeva
Summary: The paper proposes regression models based on generalizations of the normal distribution, suitable for modeling highly skewed response data. The maximum likelihood method is used for estimating parameters, and experimental results demonstrate the flexibility and usefulness of the models.
Article
Medicine, General & Internal
Yuanquan Zheng, Yingli Nie, Jingjing Lu, Hong Yi, Guili Fu
Summary: Through weighted gene co-expression network analysis and functional annotation analysis, we identified key genes associated with the severity of AA. Serum levels of proteins coded by these genes were quantitatively detected and found to be significantly increased (CD8A, PRF1, and XCL1) or decreased (BMP2) in AA tissues, especially in the subtypes of AT and AU. The serum levels of these markers were also found to be closely correlated with the Severity of Alopecia Tool (SALT) score. A prediction model combining multiple markers was established and showed high accuracy in forecasting the recurrence of AA patients.
FRONTIERS IN MEDICINE
(2023)
Article
Public, Environmental & Occupational Health
Lulu Hou, Lele Chen, Wenpei Zhang
Summary: A large sample study found a high comorbidity between premenstrual syndrome (PMS) and depression, and demonstrated that PMS can predict the development of depression.
FRONTIERS IN PUBLIC HEALTH
(2023)
Article
Geosciences, Multidisciplinary
Milton Pifano S. Ferreira, Vinicius D. Mayrink, Antonio Luiz P. Ribeiro
Summary: This study develops factor analysis to explore areal data collected in space and time, incorporating nonlinear interactions to handle spatiotemporal random effects in a mixed generalized linear regression. The presence of nonlinear interactions aims at improving cluster detection. A comprehensive simulation study is conducted to investigate the performance of the proposed approach, focusing on the analysis of ECG data related to patients affected by acute myocardial infarction.
SPATIAL STATISTICS
(2021)