Structured Programming Research Articles

Recruitment and retention of participant and study partner dyads in two multinational Alzheimer\u2019s disease registration trials

BackgroundEarly study exit is detrimental to statistical power and increases the risk for bias in Alzheimer’s disease clinical trials. Previous analyses in early phase academic trials demonstrated associations between rates of trial incompletion and participants’ study partner type, with participants enrolling with non-spouse study partners being at greater risk.MethodsWe conducted secondary analyses of two multinational phase III trials of semagacestat, an oral gamma secretase inhibitor, for mild-to-moderate AD dementia.Cox’s proportional hazards regression model was used to estimate the relationship between study partner type and the risk of early exit from the trial after adjustment for a priori identified potential confounding factors. Additionally, we used a random forest model to identify top predictors of dropout.ResultsAmong participants with spousal, adult child, and other study partners, respectively, 35%, 38%, and 36% dropped out or died prior to protocol-defined study completion, respectively. In unadjusted models, the risk of trial incompletion differed by study partner type (unadjusted p value = 0.027 for test of differences by partner type), but in models adjusting for potential confounding factors, the differences were not statistically significant (p value = 0.928). In exploratory modeling, participant age was identified as the primary characteristic to explain the relationship between study partner type and the risk of failing to complete the trial. Participant age was also the strongest predictor of trial incompletion in the random forest model.ConclusionsAfter adjustment for age, no differences in the risk of incompletion were observed when comparing participants with different study partner types in these trials. Differences between our findings and the findings of previous studies may be explained by differences in trial phase, size, geographic regions, or the composition of academic and non-academic sites.

DeepFogSim: A Toolbox for Execution and Performance Evaluation of the Inference Phase of Conditional Deep Neural Networks with Early Exits Atop Distributed Fog Platforms

The recent introduction of the so-called Conditional Neural Networks (CDNNs) with multiple early exits, executed atop virtualized multi-tier Fog platforms, makes feasible the real-time and energy-efficient execution of analytics required by future Internet applications. However, until now, toolkits for the evaluation of energy-vs.-delay performance of the inference phase of CDNNs executed on such platforms, have not been available. Motivated by these considerations, in this contribution, we present DeepFogSim. It is a MATLAB-supported software toolbox aiming at testing the performance of virtualized technological platforms for the real-time distributed execution of the inference phase of CDNNs with early exits under IoT realms. The main peculiar features of the proposed DeepFogSim toolbox are that: (i) it allows the joint dynamic energy-aware optimization of the Fog-hosted computing-networking resources under hard constraints on the tolerated inference delays; (ii) it allows the repeatable and customizable simulation of the resulting energy-delay performance of the overall Fog execution platform; (iii) it allows the dynamic tracking of the performed resource allocation under time-varying operating conditions and/or failure events; and (iv) it is equipped with a user-friendly Graphic User Interface (GUI) that supports a number of graphic formats for data rendering. Some numerical results give evidence for about the actual capabilities of the proposed DeepFogSim toolbox.

Learning-in-the-Fog (LiFo): Deep Learning Meets Fog Computing for the Minimum-Energy Distributed Early-Exit of Inference in Delay-Critical IoT Realms

Fog Computing (FC) and Conditional Deep Neural Networks (CDDNs) with early exits are two emerging paradigms which, up to now, are evolving in a standing-alone fashion. However, their integration is expected to be valuable in IoT applications in which resource-poor devices must mine large volume of sensed data in real-time. Motivated by this consideration, this article focuses on the optimized design and performance validation of <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Learning-in-the-Fog</i> (LiFo), a novel virtualized technological platform for the minimum-energy and delay-constrained execution of the inference-phase of CDDNs with early exits atop multi-tier networked computing infrastructures composed by multiple hierarchically-organized wireless Fog nodes. The main research contributions of this article are threefold, namely: (i) we design the main building blocks and supporting services of the LiFo architecture by explicitly accounting for the multiple constraints on the per-exit maximum inference delays of the supported CDNN; (ii) we develop an adaptive algorithm for the minimum-energy distributed joint allocation and reconfiguration of the available computing-plus-networking resources of the LiFo platform. Interestingly enough, the designed algorithm is capable to self-detect (typically, unpredictable) environmental changes and quickly self-react them by properly re-configuring the available computing and networking resources; and, (iii) we design the main building blocks and related virtualized functionalities of an Information Centric-based networking architecture, which enables the LiFo platform to perform the aggregation of spatially-distributed IoT sensed data. The energy-vs.-inference delay LiFo performance is numerically tested under a number of IoT scenarios and compared against the corresponding ones of some state-of-the-art benchmark solutions that do not rely on the Fog support.

The efficacy of single suture for exit site wound closure and stabilization of hemodialysis central tunneled catheter.

Introduction:The majority of the end-stage renal disease patients begin hemodialysis with the central tunneled catheter as a permanent or bridge vascular access. The procedure of central tunneled catheter insertion can be complicated by exit site bleeding, prolonged tunnel healing, and infection. The study aimed at evaluating whether the catheter exit site wound closing with a single-suture method is equally effective as the double suture method, which is most frequently applied. The assumption was that the single-suture method, which is less traumatizing for the skin, could offer an advantage for the patients with “paper” skin, that is, elderly, after long-term immunosuppressive treatment.Methods:Insertion of central tunneled catheter with antegrade tunnel formation was performed in 140 patients divided randomly into two groups of 70 patients using either single-suture method or double suture method.Results:Follow-up observations revealed a comparable number of early complications, that is, bleeding or exit site infection, in about 26% of the patients from both groups. No difference appeared in the catheter displacement, either. The percentage of patients with a prolonged central tunneled catheter exit site wound healing was not significantly lower in the single-suture method group: 5.8% versus 11.3%, p = 0.367. There was no impact of single-suture method versus double suture method on the central tunneled catheter long-term function and survival.Conclusion:The single-suture method of the central tunneled catheter exit closure is equal in efficacy to the double suture method. Its potential benefit for the patients with an increased risk of poor wound healing should be tested in further study.

Evaluation of Models for Longevity in a Closed Nucleus Breeding Programme: A Case Study Kenyan Sahiwal

Longevity and stay ability are parameters in any breeding programme. The Overstaying of any animal on a farm has got negative effects on the farmer. Short stay also means early exit of good animals that could have been used for a longer time and more productive. The intention of this review paper is to emphasize the importance of models that an upcoming breeding program can use without affecting the quality of the herd. In this paper, longevity is taken purely on the performance measure. The functional and type traits of longevity should be studied further before all models are put into play. But the case of closed breeding programmes, performance traits are easily measurable because many animals will exit the herd as they have been sold to farmers hence many functional straits will be expressed outside the herd. Genetic evaluation of the Kenyan Sahiwal should account for inbreeding. This study has also provided genetic and phenotypic parameters to enable the inclusion of longevity in the breeding objective for the Sahiwal cattle improvement programme.

Managing and Deploying Distributed and Deep Neural Models Through Kafka-ML in the Cloud-to-Things Continuum

The Internet of Things (IoT) is constantly growing, generating an uninterrupted data stream pipeline to monitor physical world information. Hence, Artificial Intelligence (AI) continuously evolves, improving life quality and business and academic activities. Kafka-ML is an open-source framework that focuses on managing Machine Learning (ML) and AI pipelines through data streams in production scenarios. Consequently, it facilitates Deep Neural Network (DNN) deployments in real-world applications. However, this framework does not consider the distribution of DNN models on the Cloud-to-Things Continuum. Distributed DNN significantly reduces latency, allocating the computational and network load between different infrastructures. In this work, we have extended our Kafka-ML framework to support the management and deployment of Distributed DNN throughout the Cloud-to-Things Continuum. Moreover, we have considered the possibility of including early exits in the Cloud-to-Things layers to provide immediate responses upon predictions. We have evaluated these new features by adapting and deploying the DNN model AlexNet in three different Cloud-to-Things scenarios. Experiments demonstrate that Kafka-ML can significantly improve response time and throughput by distributing DNN models throughout the Cloud-to-Things Continuum, compared to a Cloud-only deployment.

Моніторинг погодних умов і продуктивність елітних форм вишні (Cerasus vulgaris L.) в Західному Лісостепу України

Cherry is rather sensitive to the temperature and water regime violation that reflects itself primarily on the fruits quantity and quality. The weather conditions in the Right Bank subzone of the Ukraine’s Western Lisosteppe are characterized with the general air temperature increase, less expressed frosty periods during hibernation, shortening of the winter and spring seasons, greater summer duration. The enumerated conditions result in the early exit of the plants from the dormancy state with the considerable loss of the abilities of the acclimation to potential frosts in the springtime period. Due to this the decision was taken to carry out the monitoring researches of the winter and spring periods of 2018-2020 in order to obtain the operative information about the cherry orchards state. This enables to predict the potential of the investigated plants business productivity in the real time. As a result of analyzing the abiotic stress-factors influence on the crop the following unfavourable factors were noted, above all: heavy winds at the beginning and during the mass flowering, violation of the water and temperature regime in the impregnation and infructescense periods. Those factors cause the harvest decrease. However, the formation of the plants high productivity depends in full on their adaptivity to the environmental conditions. Therefore among the 7 explored hybrids the generative buds of 1/20, 1/21 and 2/60 were affected essentially in the form of the little drying of the pistils stigmas, decrease of the pollen qualitative indices and fertility. That indicates the lower trees adaptivity to the abiotic stress-factors as compared to the control cultivar Podbielska. The forms 2/39 and 2/54 distinguish themselves for the high winter-hardiness. This is connected with the high environment adaptivity and characterizes them as early-ripening large-fruited and productive hybrids.

Investigating Why Academically Successful Community College Students Leave College Without a Degree

Even though a postsecondary degree can offer economic, social, and civic benefits, many community college students leave without earning a degree—including some who have performed well academically and made substantial progress toward graduation. To better understand the factors contributing to early exit, we surveyed a number of former students in a large community college system. We improve the generalizability of the survey responses through multilevel regression with poststratification, which we use to reweight the responses to better represent the population in our original survey frame. We find that tuition and fees, living expenses, and no longer being eligible for financial aid are the factors contributing to early exit for the largest share of students. We also find variation in both financial and nonfinancial factors across subgroups, suggesting that targeted supports may be useful in helping students persist or return to college and complete their degree.

Understanding Home Hemodialysis Patient Attrition: A Cohort Study.

Background:Home hemodialysis (HHD) offers a flexible, patient-centered modality for patients with kidney failure. Growth in HHD is achieved by increasing the number of patients starting HHD and reducing attrition with strategies to prevent the modifiable reasons for loss.Objective:Our primary objective was to describe a Canadian HHD population in terms of technique failure and time to exit from HHD in order to understand reasons for exit. Our secondary objectives include the following: (1) determining reasons for training failure, (2) reasons for early exit from HHD, and (3) timing of program exit.Design:A retrospective cohort study of incident adult HHD patients between January 1, 2013—June 30, 2020.Setting:Alberta Kidney Care South, AKC-S HHD program.Participants:Patients who started training for HHD in AKC-S.Methods:A retrospective, cohort study of incident adult HHD patients with primary outcome time on home hemodialysis, secondary outcomes include reason for train failure, time to and reasons for technique failure. Cox-proportional hazard model to determine associations between patient characteristics and technique failure. The cumulative probability of technique failure over time was reported using a competing risks model.Results:A total of 167 patients entered HHD. Training failure occurred in 20 (12%), at 3.1 [2.0, 5.5] weeks; these patients were older (P < .001) and had 2 or more comorbidities (P < .001). Reasons for HHD exit after training included transplant (35; 21%), death (8; 4.8%), and technique failure (24; 14.4%). Overall, the median time to HHD exit, was 23 months [11, 41] and the median time of technique failure was 17 months [8.9, 36]. Reasons for technique failure included: psychosocial reasons (37%) at a median time 8.9 months [7.7, 13], safety (12.5%) at 19 months [19, 36], and medical (37.5%) at 26 months [11, 50].Limitations:Small patient population with quality of data limited by the electronic-based medical record and non-standardized definitions of reasons for exit.Conclusions:Training failure is a particularly important source of patient loss. Reasons for exit differ according to duration on HHD. Early interventions aimed at reducing train failure and increasing psychosocial supports may help program growth.

The STEM Pipeline: Do Media and Objectified Body Consciousness Create an Early Exit for Middle School Girls?

Girls and women are underrepresented in science, technology, engineering, and mathematics (STEM) fields. Research has investigated various explanatory factors (e.g., lack of role models, sexism) in these patterns. However, to date, the roles of media use and body concerns have not been examined. This study investigated associations among appearance-focused social media behaviors (e.g., posting photos), objectified body consciousness (i.e., body shame, body surveillance), and attitudes toward math and science (i.e., self-expectancy, task value, cost, future plans) in a sample of U.S. middle school girls ( N = 243). Results demonstrated that body shame partially mediated associations between appearance-focused social media behaviors and math and science attitudes. These findings suggest that media use and body concerns have implications for girls’ involvement in STEM fields.

Efficient Federated Learning for Cloud-Based AIoT Applications

As a promising method for central model training on decentralized device data without compromising user privacy, federated learning (FL) is becoming more and more popular in Internet-of-Things (IoT) design. However, due to limited computing and memory resources of devices that restrict the capabilities of hosted deep learning models, existing FL approaches for artificial intelligence IoT (AIoT) applications suffer from inaccurate prediction results. To address this problem, this article presents a collaborative <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Big.Little</i> branch architecture to enable efficient FL for AIoT applications. Inspired by the architecture of BranchyNet which has multiple prediction branches, our approach deploys deep neural network (DNN) models across both cloud and AIoT devices. Our <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Big.Little</i> branch model has two branches, where the big branch is deployed on cloud for strengthened prediction accuracy, and the little branches are used to fit for AIoT devices. When AIoT devices cannot make the prediction with high confidence using local little branches, they will resort to the big branch for further inference. To increase both prediction accuracy and early exit rate of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Big.Little</i> branch model, we propose a two-stage training and coinference scheme, which considers the local characteristics of AIoT scenarios. Comprehensive experiment results obtained from a real AIoT environment demonstrate the efficiency and effectiveness of our approach in terms of prediction accuracy and average inference time.

Support for a Lower State Pension Age for Disadvantaged Older Workers

Due to policy reforms, early exit from the labor market has decreased substantially and people are participating in the labor market until much higher ages than before. As a result, there are increasingly many people that struggle to continue working until they can comfortably retire, for example due to chronic health conditions or having to provide informal care. A potential solution would be to grant earlier access to state pension benefits (such as Social Security) for disadvantaged older workers. While it is known that many people are supportive of such a policy, the question remains how much earlier access would be granted under which circumstances. Here, using a quasi-experimental vignette design (10,350 observations nested in 2,070 respondents), we study how much earlier Dutch people would like to grant access to disadvantaged older workers. Relevant characteristics of older workers that are judged are the age at which they started working, the level of physical strain in their job, whether they have chronic health conditions, and whether they provide informal care to a loved one. The result show that, on average, people would grant older workers with chronic muscoskeletal conditions or cardiovascular disease one year earlier access to the state pension than normal, while older workers that provide daily informal care would be granted 10 months earlier access. Cumulative disadvantage could lead to a maximum of three years earlier access to pension benefits. This study provides important insights into fairness considerations surrounding state pension provisions, and implications for practice will be discussed.

EdgeKE: An On-Demand Deep Learning IoT System for Cognitive Big Data on Industrial Edge Devices

Motivated by the prospects of 5G communications and industrial Internet of Things (IoT), recent years have seen the rise of a new computing paradigm, edge computing, which shifts data analytics to network edges that are at the proximity of big data sources. Although deep neural networks (DNNs) have been extensively used in many platforms and scenarios, they are usually both compute and memory intensive, thus, difficult to be deployed on resource-limited edge devices and in performance-demanding edge applications. Hence, there is an urgent need for techniques that enable DNN models to fit into edge devices, while ensuring acceptable execution costs and inference accuracy. This article proposes an on-demand DNN model inference system for industrial edge devices, called knowledge distillation and early exit on edge (EdgeKE). It focuses on the following two design knobs: first, DNN compression based on knowledge distillation, which trains the compact edge models under the supervision of large complex models for improving accuracy and speed; second, DNN acceleration based on early exit, which provides flexible choices for satisfying distinct latency or accuracy requirements from edge applications. By extensive evaluations on the CIFAR100 dataset and across three state-of-art edge devices, experimental results demonstrate that EdgeKE significantly outperforms the baseline models in terms of inference latency and memory footprint, while maintaining competitive classification accuracy. Furthermore, EdgeKE is verified to be efficiently adaptive to the application requirements on the inference performance. The accuracy loss is within 4.84% under various latency constraints, and the speedup ratio is up to 3.30× under various accuracy requirements.

Towards a feminist political economy of time: labour circulation, social reproduction & the ‘afterlife’ of cheap labour

This article explores ‘time’ as a crucial category of analysis shaping and shaped by the dynamics of exploitation and social reproduction across the global assembly line. Focusing on the Indian garment industry, the article develops a feminist political economy of time stressing the productive and reproductive temporalities of exploitation, which give rise to multiple forms of labour circulation, including early exit from industrial work. Then, the study places this early exit under the microscope, and analyses the ‘afterlife’ available to women workers outside the factory, which often involves a transition back to informal occupations. The narrative draws both on extensive knowledge of India’s garment sweatshops, and on the detailed analysis of twenty life histories of women former factory workers in Bengaluru. The investigation of the feminist political economy of time of the global assembly line developed here suggests the presence of a revolving door between industrial and informal work in the lives of the working classes. It disproves linear global industrial development narratives constructing industrial work as ‘better work’ and contributes to feminist IPE debates by illustrating how social reproduction – its rhythms, temporalities, and everyday necessities – concretely co-constitutes the world of work across the global economy.

Work Ability among Upper-Secondary School Teachers: Examining the Role of Burnout, Sense of Coherence, and Work-Related and Lifestyle Factors

Maintaining and promoting teachers’ work ability is essential for increasing productivity and preventing early exit from the teaching profession. This study aimed to identify the predictors of work ability among upper-secondary school teachers and examine the mediating role of burnout. A large and diverse group of Czech upper-secondary school teachers was surveyed to address this goal. The sample comprised 531 upper-secondary school teachers (50.0 ± 9.94 years, 19.9 ± 10.62 in the teaching profession, 57.6% females). Relatively greater empirical support was found for the effects of burnout, sense of coherence, work–life balance, and perceived relationships in the school environment on work ability than for the impact of age, homeroom teacher duties, workload, and caring for elderly relatives. Furthermore, burnout served as an important mediator of the relationship between sense of coherence and work ability. Teachers with a higher sense of coherence are thus better able to cope with adverse work circumstances and identify and mobilize internal and external resources to prevent professional exhaustion and the subsequent decline in work ability. The study can guide interventions on the work ability of teachers.

Irreconcilable Differences: Black Teachers’ Lives and K-12 Schools

University-based teacher education programs struggle with recruitment and retention of Black teachers. While the enrollment of children of color in K-12 public schools has held steady for over a decade, Black teachers continue to represent only a small percentage of classroom teachers and leave the classroom at higher rates than their White counterparts. In this article, drawing from an in-depth interview sequence (Seidman in Interviewing as qualitative research: a guide for researchers in education and the social sciences, Teachers College Press, New York, 2013), we theorize one Black teacher’s racialized lived experience with colleagues and institutions to explore the factors that contributed to her early exit from the profession. With a Subject-Object Theory analytic framework (Kegan and Lahey, in: Lyons (ed) Handbook of reflection and reflective inquiry: mapping a way of knowing for professional reflective inquiry, Springer, Boston, 2010), our analysis underscores how a series of racially hostile encounters undermined her motivation and ability to enact change at the institutional level. Findings suggest that focusing primarily on consciousness and awareness raising in teacher education programs may fall short for Black educators.

Using a life course perspective to understand early labor market exits for people in their late 50s living in the UK

This paper explores the reasons why people exit the UK labor market early and some of the barriers working against them returning. The specific focus is a qualitative exploration of three out of work individuals, approaching the age of 60, each of whom had experienced poverty and periods of worklessness during their lifetime. The fieldwork took place in 2016 and was part of a wider mixed methods study about retirement, which used data from the 1958 birth cohort study. Researchers used narrative interviews to uncover the lived experiences and realities of these three people’s lives, and applied a life course perspective to understand how the accumulation of advantages and disadvantages during their lives shapes their attitudes and expectations. The data also show the effects of money, health, and previous employment on their decision making, and how structures such as social class, gender, and poverty are represented in and through the stories they tell.

Working conditions as risk factors for early exit from work\u2014in a cohort of 2351 employees in Germany

ObjectivesWe would assess the possible impact of a range of physical and psychosocial working conditions on early exit from paid employment (i.e., before retirement age) in a representative employee population in Germany.MethodsWe analysed a cohort from the German Study on Mental Health at Work (S-MGA) with a baseline of 2351 employees in 2011/12, sampled randomly from the register of integrated employment biographies (IEB) at the Institute for Employment Research (IAB). Follow-up ended mid-2015. Early Exit comprised episodes of either pensioning, long-term sickness absence or unemployment ≥ 18 months. Total follow-up years were 8.422. Working conditions were partly assessed by the Copenhagen Psychosocial Questionnaire (COPSOQ). Through Cox regressions, associations of baseline working conditions with time to event of exit were estimated—adjusting for baseline age, gender, poverty, fixed-term contract and socioeconomic position.ResultsIn multiple regressions, awkward body postures (HR = 1.24; 95% CI = 1.07–1.44), heavy lifting (1.17; 1.00–1.37) and high work pace (1.41; 1.16–1.72) were associated with exit. The estimated attributable fraction of exit for being exposed to less than optimal work environment was 25%. Regarding specific exit routes, repetitive movements (1.25; 1.03–1.53) increased the risk for the long-term sickness absence; work pace (1.86; 1.22–2.86) and role clarity (0.55; 0.31–1.00) were associated to unemployment; and control over working time (0.72; 0.56–0.95) decreased the risk of the early retirement.ConclusionsWork environment seems to be important for subsequent early exit from work. Physical and psychosocial demands seem to be associated to exit to a stronger extent than resources at work.

Waiting for the Payday? The Market for Startups and the Timing of Entrepreneurial Exit

Most technology startups are set up for exit through acquisition by large corporations. In choosing when to sell, startups face a trade-off. Early acquisition reduces execution errors, but later acquisition both improves the likelihood of finding a better match and benefits from increased buyer competition. Startups’ exit strategies vary considerably: Some startups aim to sell early; others remain in stealth mode by developing the invention for a late sale. We develop an analytical model to study the timing of the exit strategy. We find that startups with more capable founding teams commit to a late exit, whereas those with less capable founding teams commit to an early exit. Finally, startups with founding teams of intermediate capabilities remain flexible: They seek early offers but eventually sell late. If trying the early market is so costly that startups have to make a mutually exclusive choice between an early and late sale, startups sell inefficiently late. Instead, if they can collect early offers at no cost before deciding on the timing of sale, there are too many early acquisitions. This paper was accepted by David Simchi-Levi, business strategy.

“She is the Best Female Coach”: Female Swimming Coaches’ Experiences of Sexism

Sport participation for women and girls is at an all-time high in the United States, but women are still widely underrepresented in leadership positions and coaching (Acosta &amp; Carpenter, 2014). Women hold approximately 50% of head coaching positions of women’s teams in the National Collegiate Athletic Association, and only 18% of the head coaching positions of women’s swimming and diving teams (LaVoi &amp; Silva-Breen, 2018). Numerous barriers have been identified through previous research on the factors that inhibit upward career mobility for female coaches. Semi-structured interviews were used to examine the career experiences of 21 current or former female swimming coaches at the NCAA Division I level. The theme of sexism in coaching was pervasive and identified in five different categories: (a) misidentification, (b) differential treatment, (c) isolation, (d) tokenism, and (e) motherhood. The sexism that female coaches experience hinders upward career mobility which can lead to career dissatisfaction and early exits from the field, contributing to the underrepresentation of women in the profession.