Generalized Zero-Adjusted Models to Predict Medical Expenditures

In healthcare research, medical expenditure data for the elderly are typically semicontinuous and right-skewed, which involve a point mass at zero and may exhibit heteroscedasticity. The problem of a substantial proportion of zero values prevents traditional regression techniques based on the Gaussian, gamma, or inverse Gaussian distribution, which may lead to understanding the standard errors of the parameters and overestimating their significance. A common way to counter the problem is using zero-adjusted models. However, due to the right-skewness in the nonzeros’ response, conventional zero-adjusted models such as zero-adjusted gamma, zero-adjusted Inverse Gaussian, and classic Tobit may not perform well. Here, we firstly generalize those three types of the conventional zero-adjusted model to solve the problem of right-skewness in health care. The generalized zero-adjusted models are very flexible and include the zero-adjusted Weibull, zero-adjusted gamma, zero-adjusted inverse Gaussian, and classic Tobit models as their special cases. Using the Chinese Longitudinal Healthy Longevity Survey, we find that, according to the AIC, SBC, and deviance criteria, the zero-adjusted generalized gamma model is the best one of these generalized models to predict the odds of zero cost accurately. In order to depict the predictors affecting the amount expenditure, we further discuss the situations where the mean, dispersion of a nonzero amount expenditure and model the probability of a zero amount of ZAGG in terms of predictor variables using suitable link functions, respectively. Our analysis shows that age, health, chronic diseases, household income, and residence are the main factors influencing the medical expenditure for the elderly, but the insurance is not significant. To the best of our knowledge, little study focused on these situations, and this is the first time.

Retrospective analysis of cohort risk-factors and feeding phase timing associated with non-infectious heart disease deaths in US fed cattle

Heart disease, specifically, congestive heart failure has become of increased interest to geneticists and cattle feeders. Data on cohort associations of risk factors related to heart disease and when heart disease deaths occur in US feedlot cattle are limited. The study objectives were to 1) determine potential associations between feedlot cohort demographics and the risk of at least one non-infectious heart disease (NIHD) death occurrence and 2) determine potential association between feedlot cohort demographics and the timing of NIHD deaths during the feeding phase. Data were downloaded from commercial feedyard software and analyzed by constructing a generalized linear mixed model for both analyses. A binomial and Gaussian distribution for risk of NIHD death and timing of NIHD were utilized as link functions for their respective models. Our study population consisted of 28,950 cohorts (representing 4,596,205 cattle) that were placed in 22 US commercial feedlots from January 01, 2016, to January 01, 2019. There were 3,282 cases of NIHD deaths from a population of 75,963 cattle that died during the three-year study period. Average cohort arrival weight’s effect on NIHD probability was influenced by arrival quarter and arrival year of placement (P < 0.01). Cohorts with steers were associated with a greater probability of at least one NIHD death (2.38%) compared to heifers (1.95%; P < 0.01). Increasing cohort size was associated with an increased probability of a cohort having at least one NIHD death (P < 0.01). The probability of at least one NIHD death in a cohort increased from 1.51%, to 2.12%, and 2.87% in d on feed categories 100-175, 176-250, and 251-326 respectively. Cattle > 326 d on feed were no different in the probability of a NIHD death compared to the other feeding categories. Timing of a NIHD death had a mean and median occurrence of 110 d on feed with an interquartile range of 64 to 153 d on feed. The effect of arrival weight on d at death was influenced by year placed with heavier cattle generally decreasing the model adjusted means of d on feed at NIHD death. Arrival quarter was influenced by year placed on model adjusted means on the timing of a NIHD death. Steers with NIHD died later compared to heifers (P < 0.01) diagnosed with NIHD. In conclusion, multiple factors are associated with probability and timing of a NIHD death. Probability of having at least one NIHD death within a cohort was low and half of the deaths occurred before 110 d on feed.

Phenoloxidases in Plants—How Structural Diversity Enables Functional Specificity

The metabolism of polyphenolic polymers is essential to the development and response to environmental changes of organisms from all kingdoms of life, but shows particular diversity in plants. In contrast to other biopolymers, whose polymerisation is catalysed by homologous gene families, polyphenolic metabolism depends on phenoloxidases, a group of heterogeneous oxidases that share little beyond the eponymous common substrate. In this review, we provide an overview of the differences and similarities between phenoloxidases in their protein structure, reaction mechanism, substrate specificity, and functional roles. Using the example of laccases (LACs), we also performed a meta-analysis of enzyme kinetics, a comprehensive phylogenetic analysis and machine-learning based protein structure modelling to link functions, evolution, and structures in this group of phenoloxidases. With these approaches, we generated a framework to explain the reported functional differences between paralogs, while also hinting at the likely diversity of yet undescribed LAC functions. Altogether, this review provides a basis to better understand the functional overlaps and specificities between and within the three major families of phenoloxidases, their evolutionary trajectories, and their importance for plant primary and secondary metabolism.

The HMCES DNA-protein cross-link functions as a constitutive DNA repair intermediate

The HMCES protein forms a covalent DNA-protein cross-link (DPC) with abasic (AP) sites in ssDNA, and the resulting HMCES-DPC is thought to suppress double-strand break formation in S phase. However, the dynamics of HMCES cross-linking and whether any DNA repair pathways normally include an HMCES-DPC intermediate remain unknown. Here, we show that an HMCES-DPC forms efficiently on the AP site generated during replication-coupled DNA interstrand cross-link (ICL) repair. We use this system to show that HMCES cross-links form on DNA after the replicative CMG helicase has passed over the AP site, and that HMCES is subsequently removed by the SPRTN protease. The HMCES-DPC suppresses DSB formation, slows translesion synthesis (TLS) past the AP site, and introduces a bias for insertion of deoxyguanosine opposite the AP site. These data show that HMCES-DPCs can form as constitutive intermediates in replication-coupled repair, and they suggest a general model of how HMCES protects AP sites during DNA replication.

Modelling winter storm impacts on insured claim ratios of residential buildings in German administrative districts

<p>Severe winter storms are one of the most damaging natural hazards for European residential buildings. Previous studies mainly focused on the loss ratio (loss value / total insured sum) as a monetary value for damages. In this study the focus is on the claim ratio (number of claims / number of contracts), which is derived from a storm loss dataset provided by the German Insurance Association. Due to its magnitude, the claim ratio might be a more intuitive parameter for the use in impact-based warnings than the loss ratio.</p><p>In a first step, loss ratios and claim ratios in German administrative districts are compared to investigate differences and similarities between the two variables. While there is no significant change in the ratio between claim ratio and loss ratio with increasing wind speeds, a tendency for lower loss ratios in urban areas can be confirmed.</p><p>In a second step, a generalized linear model for daily claim ratios is developed using daily maximum wind gust (ERA5) and different non-meteorological indicators for vulnerability and exposure as predictor variables. The non-meteorological predictors are derived from the Census 2011. They include information about the district-average construction years, the number of apartments per buildings and others to get a better understanding of these factors concerning the number of buildings affected by windstorms. The modelling procedure is divided into two steps. First, a logistic regression model is used to model the probabilty claim ratios larger than zero. Second, generalized linear models with different link functions are compared regarding their ability to predict claim ratios larger than zero. In a cross-validation setting a criteria for model selection is implemented and the models of both steps are verified. Both steps show an improvement over the climatological forecast and in both cases the addition of data for vulnerability and exposure leads to in decrease of the mean squared error. </p>

Interactively visualizing distributional regression models with distreg.vis

A newly emerging field in statistics is distributional regression, where not only the mean but each parameter of a parametric response distribution can be modelled using a set of predictors. As an extension of generalized additive models, distributional regression utilizes the known link functions (log, logit, etc.), model terms (fixed, random, spatial, smooth, etc.) and available types of distributions but allows us to go well beyond the exponential family and to model potentially all distributional parameters. Due to this increase in model flexibility, the interpretation of covariate effects on the shape of the conditional response distribution, its moments and other features derived from this distribution is more challenging than with traditional mean-based methods. In particular, such quantities of interest often do not directly equate the modelled parameters but are rather a (potentially complex) combination of them. To ease the post-estimation model analysis, we propose a framework and subsequently feature an implementation in R for the visualization of Bayesian and frequentist distributional regression models fitted using the bamlss, gamlss and betareg R packages.

Crash- and Simulation-Based Safety Performance Evaluation of Freeway Rest Area

The main objective of this study was to evaluate the safety effects caused by altering the lengths of deceleration and acceleration lanes at rest areas on expressways in Korea. Although general conclusions can be found through crash-based safety analysis, to examine more specific optimal conditions considering various traffic conditions, this study proposes a novel framework to explore and evaluate crash-based and simulation-based safety performances. For this purpose, the safety performance function (SPF) and crash modification factor (CMF) were developed to reflect real-world safety impacts. To consider nonlinear trends of the parameters, nonlinearizing link functions were introduced into the analysis. Two types of simulation analyses were conducted to (1) find the combination of surrogate safety measures (SSMs) that best fit with the crash-based safety performance results and (2) determine the optimal lengths of deceleration lane and acceleration lanes for different traffic conditions. The results showed that the best length of deceleration lane of a rest area is between 240 and 260 m, depending on the traffic conditions. The results also indicated that the optimal length of acceleration lane of a rest area is between 385 and 400 m, depending on the traffic parameters. The findings of this study could be used to determine the safety solutions with a micro-traffic simulator.

Regression models for binary response applied to data on neonatal deaths in newborns

The present study presents binary data modeling regarding 1.6% of neonatal deaths in 3,448 newborns from an epidemiological and observational study with a cross-sectional design, involving the retrospective analysis of 4,293 medical records of high-risk pregnant women followed in a gestational outpatient clinic from September 2012 to September 2017. Different symmetric and asymmetric link functions were considered by means of Bayesian inference. The support of more accurate inferences regarding the parameters of the model will provide biological interpretations that are more reliable and consistent with the reality. The model that presented, significantly, the lowest value for the deviance information criterion (DIC = 398.8), was the binomial with power logit (PL) link function, whose median posterior value estimated and significant for the parameter asymmetry was l = 0.25 (0.14;1.17). This significance is observed in all other models of the power family, however with very different values ​​and significantly higher DIC values, indicating less parsimonious models. The Bayesian methodology proved to be flexible. Additionally, the results show that such model shows an accuracy = 97.4% and area under the ROC curve AUC = 89.4% in the prediction of neonatal deaths based on the weight of children at birth. Specifically, for 2.500g, a value predicted in the medical literature for low weight, the model predicts a probability of 1.43%.