Search

Peer review is the gold standard for scientific communication, but its ability to guarantee the quality of published research remains difficult to verify. Recent modeling studies suggest that peer review is sensitive to reviewer misbehavior, and it has been claimed that referees who sabotage work they perceive as competition may severely undermine the quality of publications. Here we examine which aspects of suboptimal reviewing practices most strongly impact quality, and test different mitigating strategies that editors may employ to counter them. We find that the biggest hazard to the quality of published literature is not selfish rejection of high-quality manuscripts but indifferent acceptance of low-quality ones. Bypassing or blacklisting bad reviewers and consulting additional reviewers to settle disagreements can reduce but not eliminate the impact. The other editorial strategies we tested do not significantly improve quality, but pairing manuscripts to reviewers unlikely to selfishly reject them and allowing revision of rejected manuscripts minimize rejection of above-average manuscripts. In its current form, peer review offers few incentives for impartial reviewing efforts. Editors can help, but structural changes are more likely to have a stronger impact. [ABSTRACT FROM AUTHOR]

Prophylactic interventions such as vaccine allocation are some of the most effective public health policy planning tools. The supply of vaccines, however, is limited and an important challenge is to optimally allocate the vaccines to minimize epidemic impact. This resource allocation question (which we refer to as VID) has multiple dimensions: when, where, to whom, etc. Most of the existing literature in this topic deals with the latter (to whom), proposing policies that prioritize individuals by age and disease risk. However, since seasonal influenza spread has a typical spatial trend, and due to the temporal constraints enforced by the availability schedule, the when and where problems become equally, if not more, relevant. In this paper, we study the VID problem in the context of seasonal influenza spread in the United States. We develop a national scale metapopulation model for influenza that integrates both short and long distance human mobility, along with realistic data on vaccine uptake. We also design GA, a greedy algorithm for allocating the vaccine supply at the state level under temporal constraints and show that such a strategy improves over the current baseline of pro-rata allocation, and the improvement is more pronounced for higher vaccine efficacy and moderate flu season intensity. Further, the resulting strategy resembles a ring vaccination applied spatiallyacross the US. [ABSTRACT FROM AUTHOR]

The social functions of urbanized areas are highly dependent on and supported by the convenient access to public transportation systems, particularly for the less privileged populations who have restrained auto ownership. To accurately evaluate the public transit accessibility, it is critical to capture the spatiotemporal variation of transit services. This can be achieved by measuring the shortest paths or minimum travel time between origin-destination (OD) pairs at each time-of-day (e.g. every minute). In recent years, General Transit Feed Specification (GTFS) data has been gaining popularity for between-station travel time estimation due to its interoperability in spatiotemporal analytics. Many software packages, such as ArcGIS, have developed toolbox to enable the travel time estimation with GTFS. They perform reasonably well in calculating travel time between OD pairs for a specific time-of-day (e.g. 8:00 AM), yet can become computational inefficient and unpractical with the increase of data dimensions (e.g. all times-of-day and large network). In this paper, we introduce a new algorithm that is computationally elegant and mathematically efficient to address this issue. An open-source toolbox written in C++ is developed to implement the algorithm. We implemented the algorithm on City of St. George’s transit network to showcase the accessibility analysis enabled by the toolbox. The experimental evidence shows significant reduction on computational time. The proposed algorithm and toolbox presented is easily transferable to other transit networks to allow transit agencies and researchers perform high resolution transit performance analysis. [ABSTRACT FROM AUTHOR]

Objective: The Centers for Disease Control and Prevention (CDC) coordinates a labor-intensive process to measure the prevalence of autism spectrum disorder (ASD) among children in the United States. Random forests methods have shown promise in speeding up this process, but they lag behind human classification accuracy by about 5%. We explore whether more recently available document classification algorithms can close this gap. Materials and methods: Using data gathered from a single surveillance site, we applied 8 supervised learning algorithms to predict whether children meet the case definition for ASD based solely on the words in their evaluations. We compared the algorithms’ performance across 10 random train-test splits of the data, using classification accuracy, F1 score, and number of positive calls to evaluate their potential use for surveillance. Results: Across the 10 train-test cycles, the random forest and support vector machine with Naive Bayes features (NB-SVM) each achieved slightly more than 87% mean accuracy. The NB-SVM produced significantly more false negatives than false positives (P = 0.027), but the random forest did not, making its prevalence estimates very close to the true prevalence in the data. The best-performing neural network performed similarly to the random forest on both measures. Discussion: The random forest performed as well as more recently available models like the NB-SVM and the neural network, and it also produced good prevalence estimates. NB-SVM may not be a good candidate for use in a fully-automated surveillance workflow due to increased false negatives. More sophisticated algorithms, like hierarchical convolutional neural networks, may not be feasible to train due to characteristics of the data. Current algorithms might perform better if the data are abstracted and processed differently and if they take into account information about the children in addition to their evaluations. Conclusion: Deep learning models performed similarly to traditional machine learning methods at predicting the clinician-assigned case status for CDC’s autism surveillance system. While deep learning methods had limited benefit in this task, they may have applications in other surveillance systems. [ABSTRACT FROM AUTHOR]

Studies of rescuing early-career scientists often take narrow approaches and focus on PhD students or postdoc populations. In a multi-method systems approach, we examine the inter-relations between the two ends of the pipeline and ask: what are the effects of late retirement on aging and hiring in academia? With a simulation model, we postulate that the decline in the retirement rate in academia contributes to the aging pattern through two mechanisms: (a) direct effect: longer stay of established professors, and (b) indirect effect: a hiring decline in tenure-track positions. Late retirement explains more than half of the growth in average age and brings about 20% decline in hiring. We provide empirical evidence based on the natural experimental set-up of the removal of mandatory retirement in the 1990s. [ABSTRACT FROM AUTHOR]

Climate change is affecting the growth, phenology, and distribution of species across northeastern United States. In response to these changes, some species have been adversely impacted while others have benefited. One species that has benefited from climate change, historically and in response to experimental treatments, is common ragweed (Ambrosia artemisiifolia), a widely distributed annual weed and a leading cause of hay fever in North America. To better understand how climate change may affect the distribution of common ragweed, we built a maximum entropy (Maxent) predictive model using climate and bioclimatic data and over 700 observations across the eastern U.S. Our model performed well with an AUC score of 0.765 using four uncorrelated variables, including precipitation seasonality, mean diurnal temperature range, August precipitation, and January maximum temperature. After building and testing our model, we then projected potential future common ragweed distribution using a suite of 13 global climate models (GCMs) under two future greenhouse gas scenarios for mid and late-century. In addition to providing georeferenced hot spots of potential future expansion, we also provide a metric of confidence by evaluating the number of GCMs that agree. We show a substantial contraction of common ragweed in central Florida, southern Appalachian Mountains, and northeastern Virginia and areas of potential expansion at the northern margins of its current distribution, notably in northeastern U.S. However, the vast majority of this increase is projected to occur by mid-century and may be moderated somewhat by the 2070s, implying that common ragweed may be sensitive to climatic variability. Although other factors and modeling approaches should be explored, we offer preliminary insight into where common ragweed might be a new concern in the future. Due to the health impacts of ragweed, local weed control boards may be well advised to monitor areas of expansion and potentially increase eradication efforts. [ABSTRACT FROM AUTHOR]

Background: Risk prediction is crucial in many areas of medical practice, such as cardiac transplantation, but existing clinical risk-scoring methods have suboptimal performance. We develop a novel risk prediction algorithm and test its performance on the database of all patients who were registered for cardiac transplantation in the United States during 1985-2015. Methods and findings: We develop a new, interpretable, methodology (ToPs: Trees of Predictors) built on the principle that specific predictive (survival) models should be used for specific clusters within the patient population. ToPs discovers these specific clusters and the specific predictive model that performs best for each cluster. In comparison with existing clinical risk scoring methods and state-of-the-art machine learning methods, our method provides significant improvements in survival predictions, both post- and pre-cardiac transplantation. For instance: in terms of 3-month survival post-transplantation, our method achieves AUC of 0.660; the best clinical risk scoring method (RSS) achieves 0.587. In terms of 3-year survival/mortality predictions post-transplantation (in comparison to RSS), holding specificity at 80.0%, our algorithm correctly predicts survival for 2,442 (14.0%) more patients (of 17,441 who actually survived); holding sensitivity at 80.0%, our algorithm correctly predicts mortality for 694 (13.0%) more patients (of 5,339 who did not survive). ToPs achieves similar improvements for other time horizons and for predictions pre-transplantation. ToPs discovers the most relevant features (covariates), uses available features to best advantage, and can adapt to changes in clinical practice. Conclusions: We show that, in comparison with existing clinical risk-scoring methods and other machine learning methods, ToPs significantly improves survival predictions both post- and pre-cardiac transplantation. ToPs provides a more accurate, personalized approach to survival prediction that can benefit patients, clinicians, and policymakers in making clinical decisions and setting clinical policy. Because survival prediction is widely used in clinical decision-making across diseases and clinical specialties, the implications of our methods are far-reaching. [ABSTRACT FROM AUTHOR]

Recent research has produced a number of methods for forecasting seasonal influenza outbreaks. However, differences among the predicted outcomes of competing forecast methods can limit their use in decision-making. Here, we present a method for reconciling these differences using Bayesian model averaging. We generated retrospective forecasts of peak timing, peak incidence, and total incidence for seasonal influenza outbreaks in 48 states and 95 cities using 21 distinct forecast methods, and combined these individual forecasts to create weighted-average superensemble forecasts. We compared the relative performance of these individual and superensemble forecast methods by geographic location, timing of forecast, and influenza season. We find that, overall, the superensemble forecasts are more accurate than any individual forecast method and less prone to producing a poor forecast. Furthermore, we find that these advantages increase when the superensemble weights are stratified according to the characteristics of the forecast or geographic location. These findings indicate that different competing influenza prediction systems can be combined into a single more accurate forecast product for operational delivery in real time. [ABSTRACT FROM AUTHOR]

Time spent fishing is the effort metric often studied in fisheries but it may under-represent the effort actually expended by fishers. Entire fishing trips, from the time vessels leave port until they return, may prove more useful for examining trends in fleet dynamics, fisher behavior, and fishing costs. However, such trip information is often difficult to resolve. We identified ~30,000 trips made by vessels that targeted walleye pollock (Gadus chalcogrammus) in the Eastern Bering Sea from 2008–2014 by using vessel monitoring system (VMS) and landings data. We compared estimated trip durations to observer data, which were available for approximately half of trips. Total days at sea were estimated with < 1.5% error and 96.4% of trip durations were either estimated with < 5% error or they were within expected measurement error. With 99% accuracy, we classified trips as fishing for pollock, for another target species, or not fishing. This accuracy lends strong support to the use of our method with unobserved trips across North Pacific fisheries. With individual trips resolved, we examined potential errors in datasets which are often viewed as “the truth.” Despite having > 5 million VMS records (timestamps and vessel locations), this study was as much about understanding and managing data errors as it was about characterizing trips. Missing VMS records were pervasive and they strongly influenced our approach. To understand implications of missing data on inference, we simulated removal of VMS records from trips. Removal of records straightened (i.e., shortened) vessel trajectories, and travel distances were underestimated, on average, by 1.5–13.4% per trip. Despite this bias, VMS proved robust for trip characterization and for improved quality control of human-recorded data. Our scrutiny of human-reported and VMS data advanced our understanding of the potential utility and challenges facing VMS users globally. [ABSTRACT FROM AUTHOR]

Rift Valley fever (RVF) is a zoonotic disease endemic in sub-Saharan Africa with periodic outbreaks in human and animal populations. Mosquitoes are the primary disease vectors; however, Rift Valley fever virus (RVFV) can also spread by direct contact with infected tissues. The transmission cycle is complex, involving humans, livestock, and multiple species of mosquitoes. The epidemiology of RVFV in endemic areas is strongly affected by climatic conditions and environmental variables. In this research, we adapt and use a network-based modeling framework to simulate the transmission of RVFV among hypothetical cattle operations in Kansas, US. Our model considers geo-located livestock populations at the individual level while incorporating the role of mosquito populations and the environment at a coarse resolution. Extensive simulations show the flexibility of our modeling framework when applied to specific scenarios to quantitatively evaluate the efficacy of mosquito control and livestock movement regulations in reducing the extent and intensity of RVF outbreaks in the United States. [ABSTRACT FROM AUTHOR]

The rapid increase of wealth inequality in the past few decades is one of the most disturbing social and economic issues of our time. Studying its origin and underlying mechanisms is essential for policy aiming to control and even reverse this trend. In that context, controlling the distribution of income, using income tax or other macroeconomic policy instruments, is generally perceived as effective for regulating the wealth distribution. We provide a theoretical tool, based on the realistic modeling of wealth inequality dynamics, to describe the effects of personal savings and income distribution on wealth inequality. Our theoretical approach incorporates coupled equations, solved using iterated maps to model the dynamics of wealth and income inequality. Notably, using the appropriate historical parameter values we were able to capture the historical dynamics of wealth inequality in the United States during the course of the 20th century. It is found that the effect of personal savings on wealth inequality is substantial, and its major decrease in the past 30 years can be associated with the current wealth inequality surge. In addition, the effect of increasing income tax, though naturally contributing to lowering income inequality, might contribute to a mild increase in wealth inequality and vice versa. Plausible changes in income tax are found to have an insignificant effect on wealth inequality, in practice. In addition, controlling the income inequality, by progressive taxation, for example, is found to have a very small effect on wealth inequality in the short run. The results imply, therefore, that controlling income inequality is an impractical tool for regulating wealth inequality. [ABSTRACT FROM AUTHOR]

The Prius, a power-split hybrid electric vehicle developed by Toyota, has been the top-selling vehicle in the United States hybrid electric vehicle market for the last decade. The transmission system of the vehicle is a frequent theme of study for hybrid electric vehicles. However, the control concept of the vehicle is not well known, since analyzing control behaviors requires well-designed facilities to obtain testing results and well-defined processes to analyze the obtained results. Argonne National Laboratory has these resources and capabilities. In addition, Argonne has produced a reliable simulation tool, Autonomie, by which a vehicle model for the 2010 Prius is developed on the basis of the analyzed results, and it is validated with the results of testing. The developed model demonstrates that results of vehicle performance from simulation are close to those of from real-world tests—within 5%. The main focus of this study is to provide information about the supervisory control for the 2010 Prius, so that researchers can reproduce the real-world behavior of the vehicle through simulations. The analyzed control ideas based on the testing results will be very helpful in terms of understanding the control behavior of the vehicle, and the information resulting from this study is useful to develop the controller for the vehicle at a simulation level. [ABSTRACT FROM AUTHOR]