IDENTIFICATION OF REGRESSION MODELS WITH A MISCLASSIFIED AND ENDOGENOUS BINARY REGRESSOR

We study identification in nonparametric regression models with a misclassified and endogenous binary regressor when an instrument is correlated with misclassification error. We show that the regression function is nonparametrically identified if one binary instrument variable and one binary covariate satisfy the following conditions. The instrumental variable corrects endogeneity; the instrumental variable must be correlated with the unobserved true underlying binary variable, must be uncorrelated with the error term in the outcome equation, but is allowed to be correlated with the misclassification error. The covariate corrects misclassification; this variable can be one of the regressors in the outcome equation, must be correlated with the unobserved true underlying binary variable, and must be uncorrelated with the misclassification error. We also propose a mixture-based framework for modeling unobserved heterogeneous treatment effects with a misclassified and endogenous binary regressor and show that treatment effects can be identified if the true treatment effect is related to an observed regressor and another observable variable.

Children and adolescents with disabilities in Brazil: effects of the COVID-19 pandemic

The present research aimed to perform data survey and analysis of children and teenagers among 0- to 19-year-olds related to SARS (severe acute respiratory syndrome) and COVID-19 in populations with or without disabilities during the year of 2020. The database used for the evaluations was Sivep-Gripe, made available by the Ministry of Health. The database did not present variables regarding the type of disability and a proxy was created by the binary variable of Down syndrome and by qualitative analysis of clinical data of descriptive morbidity. This limitation hindered the consideration of the experience of disability as an interaction between bodily impairments and the environment, as well as the generalization regarding cases in this population. The analysis variables included individual, regional and progression characteristics of the cases, such as the need for hospitalization, admission to the ICU, use of ventilatory supports and evolution of the cases for the recovery or death. 83,491 cases of children up to 19 years old were considered. Of this total, 2,370 (3.27%) were categorized with the disability proxy. The analyzes showed the differences between cases and progressions between children and young people without disabilities and by type of disability, with the highest proportions of COVID-19 cases found in those with physical, intellectual or psychosocial disabilities. Considering the age groups, we found higher frequencies of these cases among children up to 4 years old in general; with intellectual or psychosocial disabilities between 5 and 9 years; and those with physical disabilities between 14 and 19 years. The progression of the children's cases demonstrates the relevance of considering their vulnerability and its effects on hospital establishments, since they are more susceptible to being hospitalized, requiring ICU admissions and respiratory support. Even with the use of these resources for the maintenance of life, the proportion of children with disabilities who evolve to death is equivalent to more than double (in the cases of COVID-19) and triple (SARS) of those without disabilities. Based on these verifications, we emphasize the need to further investigate, plan and execute public policies that target this population, especially in relation to health services in the current context of increasing cases of COVID-19 and its variants across the country. In addition, we seek to contribute to academic discussions that address disability as a relevant social marker that permeates the different layers of inequality and social exclusion, exposed and deepened in the pandemic context.

Geoeffectiveness Prediction of CMEs

Coronal mass ejections (CMEs), the most important pieces of the puzzle that drive space weather, are continuously studied for their geomagnetic impact. We present here an update of a logistic regression method model, that attempts to forecast if a CME will arrive at the Earth and it will be associated with a geomagnetic storm defined by a minimum Dst value smaller than −30 nT. The model is run for a selection of CMEs listed in the LASCO catalogue during the solar cycle 24. It is trained on three fourths of these events and validated for the remaining one fourth. Based on five CME properties (the speed at 20 solar radii, the angular width, the acceleration, the measured position angle and the source position – binary variable) the model successfully predicted 98% of the events from the training set, and 98% of the events from the validation one.

Association between neutrophil-to-eosinophil ratio (NER) and efficacy outcomes in the JAVELIN Renal 101 study.

4549 Background: Baseline NER has been reported to be associated with outcomes of immuno-oncology based combination treatment in advanced renal cell carcinoma (aRCC). We report outcomes by baseline NER of patients with aRCC in the JAVELIN Renal 101 trial who received avelumab + axitinib (A + Ax) or sunitinib (S). Methods: We calculated the median NER (mNER) for patients in the A + Ax and the S arms at the data cutoff (April 20, 2020) for the 3rd interim analysis (IA3). Progression-free survival (PFS), overall survival (OS), and objective response (OR) by NER are reported. Multivariate Cox regression analyses of PFS and OS were also conducted. Results: At the IA3 cutoff date, the mNERs for the A + Ax arm (n = 383) and S arm (n = 396) were 29.2 and 27.0, respectively. OR, PFS and OS for both arms are summarized in the table below. Better observed treatment outcomes in OR (63.9% vs 55.2%) and median PFS (15.5 vs, 11.1 months) were observed for patients with a NER < median vs. NER ≥ median in the A + Ax arm, while there were not major differences in outcome based on NER in the S arm. The stratified hazard ratio (HR) for PFS in patients with a NER < median compared with those with a NER ≥ median in the A + Ax arm was 0.81 (95% CI, 0.630-1.035) and 0.93 (95% CI, 0.728-1.181) in the S arm. Patients with a NER < median had improved OS compared with those with a NER ≥ median in the A + Ax arm (stratified HR, 0.67; 95% CI, 0.481-0.940) and the S arm (stratified HR, 0.57; 95% CI, 0.424-0.779). Multivariate analysis showed that a low NER was associated with longer PFS and OS by treating baseline NER as either a continuous variable or a binary variable (dichotomized by median). Conclusions: Baseline NER may be predictive of OR and PFS in aRCC patients treated with A + Ax, and prognostic for overall survival regardless of therapy. Clinical trial information: NCT02684006. [Table: see text]

A population-level analysis of the protective effects of androgen deprivation therapy against COVID-19 disease incidence and severity

ABSTRACT Importance: The incidence and severity of coronavirus disease 19 (COVID-19) is higher in men. Sex hormones potentially offer one explanation for differences by sex. Objective: To determine whether men exposed to androgen deprivation therapy (ADT) have lower incidence and severity of COVID-19. Design: We conducted an observational study of male Veterans treated in the Veterans Health Administration from February 15th to July 15th, 2020. We developed a propensity score model to predict the likelihood to undergo Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) testing. We performed multivariable logistic regression modeling adjusted with inverse probability weighting to examine the relationship between ADT and COVID-19 incidence. We conducted logistic regression analysis among COVID-19 patients to test the association between ADT and COVID-19 severity. Setting: The U.S. Department of Veterans Affairs Participants: The study sample consisted of 6,250,417 male Veterans who were alive as of February 15, 2020. Exposure: Exposure to ADT was defined as having any prescription for a luteinizing hormone releasing hormone analogue or an antiandrogen in the six months prior to the index date. Main Outcomes and Measures: To assess incidence, we used a binary variable indicating any positive reverse transcriptase polymerase chain reaction SARS-CoV-2 test result through July 15, 2020. To measure severity, we constructed a binary variable indicating whether a patient was admitted to the intensive care unit, placed on mechanical ventilation, or dead in the 60 days following a positive test up to July 15, 2020. Results: We identified 246,087 patients who had been tested for SARS-CoV-2, of whom 3,057 were exposed to ADT, and 36,096 patients with cancer and no ADT exposure. Of these, 295 ADT patients and 2,427 other cancer patients had COVID-19 illness. In the primary, propensity-weighted comparison of ADT patients to cancer patients not on ADT, ADT was associated with decreased likelihood of testing positive for SARS-CoV-2 (adjusted OR, 0.88 [95% CI, 0.81-0.95]; p=0.001). ADT was associated with fewer severe COVID-19 outcomes (OR 0.72 [95% CI 0.53-0.96]; p=0.03). Conclusions and Relevance: ADT is associated with reduced incidence and severity of COVID-19 amongst male Veterans. Repurposing of drugs that modulate androgen production and/or action may represent viable potential treatments for COVID-19.

Beyond ‘Migrant Background’: How to Select Relevant, Social Justice Oriented, and Feasible Social Categories in Educational Research

Across continental Europe, educational research samples are often divided by ‘migrant background,’ a binary variable criticized for masking participant heterogeneity and reinforcing exclusionary norms of belonging. This study endorses more meaningful, representative, and precise research by offering four guiding questions for selecting relevant, social justice oriented, and feasible social categories. Using a preregistered empirical example, we compare selected social categories (‘migrant background,’ family heritage, religion, citizenship, cultural identification, generation status) in their potential to reveal participant heterogeneity and differences in means and relations between variables (discrimination experiences, perceived societal Islamophobia, national identity) and academic motivation among 1335 adolescents in Germany (48% female, Mage = 14.69). Regression analyses and multigroup SEM revealed differential experiences with and implications of discrimination for academic motivation. Results highlight the need for a deliberate, transparent use of social categories to make discrimination visible and centre participants’ subjective experiences.

A Survey on Nonstrategic Models of Opinion Dynamics

The paper presents a survey on selected models of opinion dynamics. Both discrete (more precisely, binary) opinion models as well as continuous opinion models are discussed. We focus on frameworks that assume non-Bayesian updating of opinions. In the survey, a special attention is paid to modeling nonconformity (in particular, anticonformity) behavior. For the case of opinions represented by a binary variable, we recall the threshold model, the voter and q-voter models, the majority rule model, and the aggregation framework. For the case of continuous opinions, we present the DeGroot model and some of its variations, time-varying models, and bounded confidence models.

Bread production technology modelling using hop starter in ovens of various types

The article studies the processes of structure formation of baked dough pieces using wheat and hop starters and changes in the properties of finished products and semi-finished products depending on the baking methods. A whole complex of physical, chemical and biochemical processes takes place in the dough piece during baking under the influence of heat and moisture, which causes considerable changes in the bread dough. These processes cause changes in the baked dough piece, that cause turning the dough into bread. Duration and intensity of the processes occurring on the surface and in the inner layers of the dough piece during baking depend on the temperature. Therefore, creation of optimal modes of heating the baked dough piece at various stages allows you to get products of the required quality. To simulate the processes of crumb formation, changes in temperature inside the dough piece of baking, and specific volume over time with different methods of dough and baking, regression models were used, that take into account the influence of qualitative factors. Each qualitative factor having two grades was replaced by one binary variable. The solution of a multicriteria optimization problem showed that the studied indicators reach the optimal values when baking in an air-o-steam and preparing a dough using hop starter.

Mobile Banking Adoption in Organization: Review of Empirical Literature

This study reviews literature on mobile banking adoption in organizations to identify its influential factors and its operationalization in prior literature. We classify the factors that influence mobile banking adoption using the three contexts suggested by the Technology, Organization and Environment (TOE) framework, namely, technology, organization, and environment. The finding suggests that the influences of these factors vary across studies and most of the studies have operationalized mobile banking adoption using intention to adopt mobile banking or binary variable, rather than the actual use of the technology.

MODELING AND MANAGING MEDICAL DATA

The article describes a statistical method for analyzing medical data based on the comparison of binary samples. Processing data that is accumulated in transactional medical information systems, based on the analysis of binary samples, allows you to determine the indicators of laboratory research and diagnoses that are characteristic of harmful production factors. This will contribute to the development of digital technologies in healthcare, which will improve both diagnostics and treatment methods, as well as facilitate the adoption of competent management decisions. The research results were converted to binary form by comparing them with the statistical norm interval. Diagnoses were considered initially as a binary variable. The samples obtained as a result of binarization for two groups, the first group includes people whose production activities contain harmful factors, and the second – those who do not have these factors, were compared with each other. The initial group turned out to be heterogeneous in relation to the other group, so it was decided to conduct a further study based on the development and testing of methods for adjusting samples in order to achieve uniformity while maximizing the preservation of medical data used for analysis.