Fidelity Data Research Articles

The structured health intervention for truckers (SHIFT) cluster randomised controlled trial: a mixed methods process evaluation

BackgroundThis paper presents the mixed methods process evaluation of the randomised controlled trial (RCT) of the Structured Health Intervention For Truckers (SHIFT), a multi-component intervention targeting physical activity and positive lifestyle behaviours in a cohort of 382 truck drivers in the UK. The SHIFT RCT found a significant difference in daily steps between intervention and control groups at 6-months in favour of the intervention participants.MethodsSHIFT was evaluated within a cluster-RCT and involved 25 transport sites (12 intervention and 13 control sites). Intervention components included an education session, Fitbit, text messages, and cab workout equipment. Participants completed questionnaires at baseline and 6-months follow-up. Semi-structured focus groups/interviews were conducted with drivers (n = 19) and managers (n = 18) from each site, after completion of the final follow-up health assessment (16-18 months post-randomisation). Questionnaires and interviews collected information on fidelity, dose, context, implementation, barriers, sustainability, and contamination.ResultsQuestionnaire and interview data from intervention participants indicated favourable attitudes towards SHIFT, specifically towards the Fitbit with a high proportion of drivers reporting regularly using it (89.1%). 79.2% of intervention participants attended the education session, which was deemed useful for facilitating improvements in knowledge and behaviour change, dietary changes were predominantly recalled. Despite not being part of the intervention, participants reported that feedback from the health assessments motivated them to change aspects of their lifestyle (intervention = 91.1%, control = 67.5%). The cab workout equipment was used less and spoken unfavourably of in the interviews. The main barriers to a healthy lifestyle at work were reported as long hours and irregular shift patterns. The most suggested improvement for the intervention was more frequent contact with drivers. Managers were positive about the objectives of SHIFT, however almost all mentioned the challenges related to implementation, specifically in smaller sites.ConclusionsOverall, SHIFT was predominantly implemented as intended, with minimal discrepancies seen between the delivery and protocol. Having said this, transport sites each have distinct characteristics, which may require adaptations to individual settings to encourage participation. Managers and drivers reported enthusiasm and necessity for SHIFT to be included in future Certificate of Professional Competence training.Trial registrationISRCTN10483894 (date registered: 01/03/2017).

Molecular Structure Evaluation and Image-Guided Atomistic Representation of Hard Carbon Electrodes

Construction of large-scale atomistic representations of hard carbon electrodes aids exploration of structure-property relationships. These representations of practical value need to agree with experimental data, specifically the distribution of structural features. The molecular structure of a commercial hard carbon was evaluated by HRTEM image analysis in combination with LDIMS, FT-IR, XPS, XRD, SAXS, and gas sorption. In particular, an improved algorithm was applied to automatically calculate the interlayer spacing by finding LCS (longest common subsequence), which can extract more high fidelity data of fringe pairs from the HRTEM image analysis. Hard carbon is a partially ordered system, with order varying over length scales. Thus, a large-scale atomistic representastion (C48025H1857O811N198S127) in a 100 × 100 × 100 Å cubic cell was generated using an image-guided construction approach, better capturing the structural diversity, micropore distribution, and spatial arrangement necessary to represent carbon electrode behavior. A wide variety of chemical and physical parameters were consistent with experimental data. Such structural model that depicts experimentally-determined characteristics will provide valuable strategies for the development of high-performance carbon electrodes.

SCDeep: Single-Channel Depth Encoding for 3D-Range Geometry Compression Utilizing Deep-Learning Techniques

Recent advances in optics and computing technologies have encouraged many applications to adopt the use of three-dimensional (3D) data for the measurement and visualization of the world around us. Modern 3D-range scanning systems have become much faster than real-time and are able to capture data with incredible precision. However, increasingly fast acquisition speeds and high fidelity data come with increased storage and transmission costs. In order to enable applications that wish to utilize these technologies, efforts must be made to compress the raw data into more manageable formats. One common approach to compressing 3D-range geometry is to encode its depth information within the three color channels of a traditional 24-bit RGB image. To further reduce file sizes, this paper evaluates two novel approaches to the recovery of floating-point 3D range data from only a single-channel 8-bit image using machine learning techniques. Specifically, the recovery of depth data from a single channel is enabled through the use of both semantic image segmentation and end-to-end depth synthesis. These two distinct approaches show that machine learning techniques can be utilized to enable significant file size reduction while maintaining reconstruction accuracy suitable for many applications. For example, a complex set of depth data encoded using the proposed method, stored in the JPG 20 format, and recovered using semantic segmentation techniques was able to achieve an average RMS reconstruction accuracy of 99.18% while achieving an average compression ratio of 106:1 when compared to the raw floating-point data. When end-to-end synthesis techniques were applied to the same encoded dataset, an average reconstruction accuracy of 99.59% was experimentally demonstrated for the same average compression ratio.

Modified Multifidelity Surrogate Model Based on Radial Basis Function with Adaptive Scale Factor

Multifidelity surrogates (MFSs) replace computationally intensive models by synergistically combining information from different fidelity data with a significant improvement in modeling efficiency. In this paper, a modified MFS (MMFS) model based on a radial basis function (RBF) is proposed, in which two fidelities of information can be analyzed by adaptively obtaining the scale factor. In the MMFS, an RBF was employed to establish the low-fidelity model. The correlation matrix of the high-fidelity samples and corresponding low-fidelity responses were integrated into an expansion matrix to determine the scaling function parameters. The shape parameters of the basis function were optimized by minimizing the leave-one-out cross-validation error of the high-fidelity sample points. The performance of the MMFS was compared with those of other MFS models (MFS-RBF and cooperative RBF) and single-fidelity RBF using four benchmark test functions, by which the impacts of different high-fidelity sample sizes on the prediction accuracy were also analyzed. The sensitivity of the MMFS model to the randomness of the design of experiments (DoE) was investigated by repeating sampling plans with 20 different DoEs. Stress analysis of the steel plate is presented to highlight the prediction ability of the proposed MMFS model. This research proposes a new multifidelity modeling method that can fully use two fidelity sample sets, rapidly calculate model parameters, and exhibit good prediction accuracy and robustness.

해양공학수조 환경을 고려한 수중함 자유항주모형의 수중 복합항법 기반 경로점 추종 및 고장 허용 제어기 개발

The Submarine Free Running Model Test (FRMT) refers to a test using a submarine scale model with self-propulsion and remote control capability. The main purpose of the submarine FRMT is to predict the maneuverability of a full-scale vehicle, in addition to the possibility of providing high fidelity data on the maneuvering characteristics in various situations. In case the submarine FRMT was performed below a certain depth, the model would not be capable of receiving GPS or RF signals for remote control. Accordingly, an appropriate operational algorithm is required to autonomously carry out a given task. This paper presents the development results of the underwater integrated navigation and control algorithm for performing the submarine FRMT in an ocean engineering basin. In the ocean engineering basin, it is possible to exclude the influence of disturbance caused by waves and currents. However, it is essential to comprehensively consider collision avoidance and mission control within a limited area of operation. The navigation system is composed of an Inertial Navigation System (INS), a Doppler Velocity Logger (DVL), and a depth sensor. Moreover, for additional calibration, an underwater ultrasonic range sensor and the basin wall position information are utilized. Considering modeling uncertainties and a controller failure situation, a robust controller based on a sliding mode controller and a nonlinear disturbance observer was used. The simulated proposed algorithm was evaluated by analyzing and comparing it with conventional navigation and control methods.

Multi-fidelity modeling to predict the rheological properties of a suspension of fibers using neural networks and Gaussian processes

Unveiling the rheological properties of fiber suspensions is of paramount interest to many industrial applications. There are multiple factors, such as fiber aspect ratio and volume fraction, that play a significant role in altering the rheological behavior of suspensions. Three-dimensional (3D) numerical simulations of coupled differential equations of the suspension of fibers are computationally expensive and time-consuming. Machine learning algorithms can be trained on the available data and make predictions for the cases where no numerical data are available. However, some widely used machine learning surrogates, such as neural networks, require a relatively large training dataset to produce accurate predictions. Multi-fidelity models, which combine high-fidelity data from numerical simulations and less expensive lower fidelity data from resources such as simplified constitutive equations, can pave the way for more accurate predictions. Here, we focus on neural networks and the Gaussian processes with two levels of fidelity, i.e., high and low fidelity networks, to predict the steady-state rheological properties, and compare them to the single-fidelity network. High-fidelity data are obtained from direct numerical simulations based on an immersed boundary method to couple the fluid and solid motion. The low-fidelity data are produced by using constitutive equations. Multiple neural networks and the Gaussian process structures are used for the hyperparameter tuning purpose. Results indicate that with the best choice of hyperparameters, both the multi-fidelity Gaussian processes and neural networks are capable of making predictions with a high level of accuracy with neural networks demonstrating marginally better performance.

Cognitive-Behavioral Social Skills Training Adapted for Youth at Clinical High Risk for Psychosis.

Interventions for functional impairments in adolescents and young adults at clinical high risk (CHR) for psychosis are needed. Cognitive-Behavioral Social Skills Training (CBSST) has been found to improve functioning in patients with schizophrenia. The CBSST manual was adapted for CHR and implemented across 3 sites. The key changes that were made were to present a focus of normalization and destigmatization of attenuated psychotic symptoms and since CBSST has a major focus on role plays, problem solving and challenging thoughts, examples of these were changed to be more appropriate for this young CHR population. We describe the manual modifications and present fidelity data to examine the success of training and supervision methods in a multi-site randomized controlled trial of CBSST in CHR youth. Fidelity was high and comparable across sites. Case vignettes are presented to demonstrate how CBSST techniques were adapted for UHR individuals to target functional impairments.

Optimizing Outcomes for Children With Phonological Impairment: A Systematic Search and Review of Outcome and Experience Measures Reported in Intervention Research.

Reporting of outcome and experience measures is critical to our understanding of the effect of intervention for speech sound disorders (SSD) in children. There is currently no agreed-upon set of measures for reporting intervention outcomes and experiences. In this article, we introduce the Speech Outcome Reporting Taxonomy (SORT), a tool designed to assist with the classification of outcome and experience measures. In a systematic search and review using the SORT, we explore the type and frequency of these measures reported in intervention research addressing phonological impairment in children. Given the integral relationship between intervention fidelity and intervention outcomes, reporting of fidelity is also examined. Five literature databases were searched to identify articles written or translated into English published between 1975 and 2020. Using the SORT, outcome and experience measures were extracted and categorized. The number of intervention studies reporting fidelity was determined. A total of 220 articles met inclusion criteria. The most frequently reported outcome domain was broad generalization measures (n = 142, 64.5%), followed by specific measures of generalization of an intervention target (n = 133, 60.5%). Eleven (5.0%) articles reported measures of the impact of the phonological impairment on children's activity, participation, quality of life, or others. Twenty articles (9.1%) reported on parent, child, or clinician experience or child engagement. Fidelity data were reported for 13.4% of studies of interventions. The measurement of intervention outcomes is challenging yet important. No single type of measure was reported across all articles. Through using tailored measures closely related to intervention targets in combination with a universal set of measures of intelligibility, the impact of phonological impairment on children's lives, and the experience of receiving and providing intervention, researchers and clinicians could work together to progress insights and innovations in science and practice for children with SSD. https://doi.org/10.23641/asha.19497803.

A randomized controlled trial of a proportionate universal parenting program delivery model (E-SEE Steps) to enhance child social-emotional wellbeing.

BackgroundEvidence for parenting programs to improve wellbeing in children under three is inconclusive. We investigated the fidelity, impact, and cost-effectiveness of two parenting programs delivered within a longitudinal proportionate delivery model (‘E-SEE Steps’).MethodsEligible parents with a child ≤ 8 weeks were recruited into a parallel two-arm, assessor blinded, randomized controlled, community-based, trial with embedded economic and process evaluations. Post-baseline randomization applied a 5:1 (intervention-to-control) ratio, stratified by primary (child social-emotional wellbeing (ASQ:SE-2)) and key secondary (maternal depression (PHQ-9)) outcome scores, sex, and site. All intervention parents received the Incredible Years® Baby Book (IY-B), and were offered the targeted Infant (IY-I)/Toddler (IY-T) program if eligible, based on ASQ:SE-2/PHQ-9 scores. Control families received usual services. Fidelity data were analysed descriptively. Primary analysis applied intention to treat. Effectiveness analysis fitted a marginal model to outcome scores. Cost-effectiveness analysis involved Incremental Cost-Effectiveness Ratios (ICERs).ResultsThe target sample (N = 606) was not achieved; 341 mothers were randomized (285:56), 322 (94%) were retained to study end. Of those eligible for the IY-I (n = 101), and IY-T (n = 101) programs, 51 and 21 respectively, attended. Eight (of 14) groups met the 80% self-reported fidelity criteria. No significant differences between arms were found for adjusted mean difference scores; ASQ:SE-2 (3.02, 95% CI: -0.03, 6.08, p = 0.052), PHQ-9 (-0.61; 95% CI: -1.34, 0.12, p = 0.1). E-SEE Steps had higher costs, but improved mothers’ Health-related Quality of Life (0.031 Quality Adjusted Life Year (QALY) gain), ICER of £20,062 per QALY compared to control. Serious adverse events (n = 86) were unrelated to the intervention.ConclusionsE-SEE Steps was not effective, but was borderline cost-effective. The model was delivered with varying fidelity, with lower-than-expected IY-T uptake. Changes to delivery systems and the individual programs may be needed prior to future evaluation.Trial registrationInternational Standard Randomized Controlled Trial Number: ISRCTN11079129.

A 2-D reduced dynamic model for a shell-and-tube based metal hydride reactor for geometry and operation condition optimal design

An MH reactor is a complex multi-physics system and its optimal design is critical to reduce the operation and investment cost. Geometry and operation condition optimal design reported in the literature is often based on parametric study of high fidelity simulations, which is computationally expensive. In this paper, a framework to develop a fast reduced model for a two-dimensional shell-and-tube based MH reactor is proposed. This framework is generally applicable to parabolic PDEs. A lumped capacity model is used as a reduced model basis. Tuning factors that aim to correct the transportation phenomena in a lumped capacity model are introduced. The tuning factors are correlated to key dimensionless group parameters of the system and are constructed by using data from a high fidelity model (HFM). Comparing to pure data-driven models, the proposed modeling framework has good extrapolation ability and hence requires less high fidelity data. The tuning factors establish parametric dependence of the system for efficient optimal design.

A first principles investigation of ternary and quaternary II–VI zincblende semiconductor alloys

One of the most common ways of tuning the stability, electronic structure, and optical behavior of semiconductors is via composition engineering. By mixing multiple isovalent elements at any cation or anion site, new compositions may be generated with markedly different properties than end-point compositions, and not always lying within a predictable trend. In this work, we explore the trends in lattice constant, electronic band gap, formation and mixing energy, and optical absorption behavior in a series of II–VI zincblende semiconductors with Cd/Zn at the cation site and S/Se/Te at the anion site, using multiple levels of density functional theory approximations. We find that while the GGA-PBE functional reproduces all trends correctly, full geometry optimization with the HSE06 functional predicts band gaps with much higher experimental accuracy. We find that all mixed S–Se and mixed Cd–Zn compounds show linear trends in band gap, rising from Se to S and Cd to Zn, respectively, whereas all Se–Te mixed compounds exhibit band gap bowing. All mixing energy curves, calculated based on decomposition to end point compositions, show inverted bowing behavior but with small positive mixing energy values <50 meV per formula unit, indicating robust stability of all solid solutions. Formation energies, calculated based on decomposition to elemental species, always show linear trends and remain sufficiently negative for all binaries, ternaries and quaternaries, whereas lattice constants show expected linear trends. We further report trends in optical absorption spectra and relationships between PBE and HSE computed properties, which reveal equations that can be used to accurately predict higher fidelity data. This work lays out systematic trends in the stability and optoelectronic characteristics of Cd–Zn–S–Se–Te alloys and enables the selection of optimal compositions for desired applications.

Agents for sequential learning using multiple-fidelity data

Sequential learning for materials discovery is a paradigm where a computational agent solicits new data to simultaneously update a model in service of exploration (finding the largest number of materials that meet some criteria) or exploitation (finding materials with an ideal figure of merit). In real-world discovery campaigns, new data acquisition may be costly and an optimal strategy may involve using and acquiring data with different levels of fidelity, such as first-principles calculation to supplement an experiment. In this work, we introduce agents which can operate on multiple data fidelities, and benchmark their performance on an emulated discovery campaign to find materials with desired band gap values. The fidelities of data come from the results of DFT calculations as low fidelity and experimental results as high fidelity. We demonstrate performance gains of agents which incorporate multi-fidelity data in two contexts: either using a large body of low fidelity data as a prior knowledge base or acquiring low fidelity data in-tandem with experimental data. This advance provides a tool that enables materials scientists to test various acquisition and model hyperparameters to maximize the discovery rate of their own multi-fidelity sequential learning campaigns for materials discovery. This may also serve as a reference point for those who are interested in practical strategies that can be used when multiple data sources are available for active or sequential learning campaigns.

Error Quantification for the Assessment of Data-Driven Turbulence Models

Data-driven turbulence modelling is becoming common practice in the field of fluid mechanics. Complex machine learning methods are applied to large high fidelity data sets in an attempt to discover relationships between mean flow features and turbulence model parameters. However, a clear discrepancy is emerging between complex models that appear to fit the high fidelity data well a priori and simpler models which subsequently hold up in a posteriori testing through CFD simulations. With this in mind, a novel error quantification technique is proposed consisting of an upper and lower bound, against which data-driven turbulence models can be systematically assessed. At the lower bound is models that are linear in either the full set or a subset of the input features, where feature selection is used to determine the best model. Any machine learning technique must be able to improve on this performance for the extra complexity in training to be of practical use. The upper bound is found by the stable insertion of the high fidelity data for the Reynolds stresses into CFD simulation. Three machine learning methods, Gene Expression Programming, Deep Neural Networks and Gaussian Mixtures Models are presented and assessed on this error quantification technique. We further show that for the simple canonical cases often used to develop data-driven methods, lower bound linear models can provide very satisfactory accuracy and stability with limited scope for substantial improvement through more complex machine learning methods.

Machine Learning for the Development of Data-Driven Turbulence Closures in Coolant Systems

Abstract This work shows the application of Gene Expression Programming to augment RANS turbulence closures for flows though complex geometries. Specifically, an optimized internal cooling channel of a turbine blade, designed for additive manufacturing. One of the challenges in internal cooling design is the heat transfer accuracy of the RANS formulation in comparison to higher fidelity methods, which are still not used in design on account of their computational cost. However, high fidelity data can be extremely valuable for improving current lower fidelity models and this work shows the application of data-driven approaches to develop turbulence closures for an internally ribbed duct. Different approaches are compared and the results of the improved model are illustrated; first on the same geometry, and then for an unseen predictive case. The work shows the potential of using data-driven models for accurate heat transfer predictions even in non-conventional configurations and indicates the ability of closures learnt from complex flow cases to adapt successfully to unseen test cases.

Implementation of competence committees during the transition to CBME in Canada: A national fidelity-focused evaluation

Purpose This study evaluated the fidelity of competence committee (CC) implementation in Canadian postgraduate specialist training programs during the transition to competency-based medical education (CBME). Methods A national survey of CC chairs was distributed to all CBME training programs in November 2019. Survey questions were derived from guiding documents published by the Royal College of Physicians and Surgeons of Canada reflecting intended processes and design. Results Response rate was 39% (113/293) with representation from all eligible disciplines. Committee size ranged from 3 to 20 members, 42% of programs included external members, and 20% included a resident representative. Most programs (72%) reported that a primary review and synthesis of resident assessment data occurs prior to the meeting, with some data reviewed collectively during meetings. When determining entrustable professional activity (EPA) achievement, most programs followed the national specialty guidelines closely with some exceptions (53%). Documented concerns about professionalism, EPA narrative comments, and EPA entrustment scores were most highly weighted when determining resident progress decisions. Conclusions Heterogeneity in CC implementation likely reflects local adaptations, but may also explain some of the variable challenges faced by programs during the transition to CBME. Our results offer educational leaders important fidelity data that can help inform the larger evaluation and transformation of CBME.

Feasibility of deploying peer coaches to mentor frontline home health aides and promote mobility among individuals recovering from a stroke: pilot test of a randomized controlled trial

BackgroundEach year, approximately 100,000 individuals receive home health services after a stroke. Evidence has shown the benefits of home-based stroke rehabilitation, but little is known about resource-efficient ways to enhance its effectiveness, nor has anyone explored the value of leveraging low-cost home health aides (HHAs) to reinforce repetitive task training, a key component of home-based rehabilitation. We developed and piloted a Stroke Homehealth Aide Recovery Program (SHARP) that deployed specially trained HHAs as “peer coaches” to mentor frontline aides and help individuals recovering from stroke increase their mobility through greater adherence to repetitive exercise regimens. We assessed the feasibility of SHARP and its readiness for a full-scale randomized controlled trial (RCT). Specifically, we examined (1) the practicability of recruitment and randomization procedures, (2) program acceptability, (3) intervention fidelity, and (4) the performance of outcome measures.MethodsThis was a feasibility study including a pilot RCT. Target enrollment was 60 individuals receiving post-stroke home health services, who were randomized to SHARP + usual home care or usual care only. The protocol specified a 30-day intervention with four planned in-home coach visits, including one joint coach/physical therapist visit. The primary participant outcome was 60-day change in mobility, using the performance-based Timed Up and Go and 4-Meter Walk Gait Speed tests. Interviews with participants, coaches, physical therapists, and frontline aides provided acceptability data. Enrollment figures, visit tracking reports, and audio recordings provided intervention fidelity data. Mixed methods included thematic analysis of qualitative data and quantitative analysis of structured data to examine the intervention feasibility and performance of outcome measures.ResultsAchieving the 60-participant enrollment target required modifying participant eligibility criteria to accommodate a decline in the receipt of HHA services among individuals receiving home care after a stroke. This modification entailed intervention redesign. Acceptability was high among coaches and participants but lower among therapists and frontline aides. Intervention fidelity was mixed: 87% of intervention participants received all four planned coach visits; however, no joint coach/therapist visits occurred. Sixty-day follow-up retention was 78%. However, baseline and follow-up performance-based primary outcome mobility assessments could be completed for only 55% of participants.ConclusionsThe trial was not feasible in its current form. Before progressing to a definitive trial, significant program redesign would be required to address issues affecting enrollment, coach/HHA/therapist coordination, and implementation of performance-based outcome measures.Trial registrationClinicalTrials.gov, NCT04840407. Retrospectively registered on 9 April 2021

Online reconstruction of 3D temperature field fused with POD-based reduced order approach and sparse sensor data

In power engineering, analysis of the thermal behavior of steam turbine and development of control strategies depend on online evaluations of the temperature fields of in-service components. Due to the complexity of structure and the variability of operating conditions, an online reconstruction technique fused with proper orthogonal decomposition (POD) and sparse sensor data is described herein to monitor the three-dimensional (3D) transient temperature fields of a steam turbine casing. The POD-based reduced order approach was driven by the high-fidelity data of the steam turbine under several start-up conditions, and the POD was used to extract a reduced set of vectors in the orthogonal basis to capture the spatial distribution modes of 3D temperature field. As a result, the temperature fields of steam turbine casing can be reconstructed using the reduced set of vectors in the orthogonal POD basis and the corresponding coefficients, and then it is feasible to online monitor the global temperature distribution by minimizing deviations between the predicted and measured temperatures. The feasibility and accuracy of the reconstruction approach were demonstrated using two numerical examples in comparison with high fidelity data. Finally, real sensor-measured sequential temperatures were used to dynamically determine the real temperature field of the casing. Notably, the reconstructed temperatures were highly consistent with the sensor-measured temperatures, and the time delay of almost every reconstruction process was less than 0.02 s. This proposed reconstruction approach can provide reliable online monitoring for the transient temperature fields of a 3D complex component based on sparse sensors, and it will provide effective support for the design, operation and maintenance.

A Synthesis of the Conceptualization and Measurement of Implementation Fidelity in Mathematics Intervention Research

In intervention studies, high rates of implementation fidelity are important markers of a study’s success; however, the definition of implementation fidelity is both complex and dynamic. In this synthesis, we examined the dimensions of implementation fidelity measured and reported in 99 studies in which researchers utilized a mathematics intervention for elementary students (i.e., Grades 1–5). We examined implementation fidelity following recommendations made by Dane and Schneider (1998), O’Donnell (2008), and DeFouw et al. (2009) to capture a comprehensive representation of the implementation fidelity data collected and reported within mathematics intervention studies. We organized our conceptualization of implementation fidelity into four overarching categories and nine dimensions within those categories: intervention design (i.e., theories of change and logistics), fidelity of implementor (i.e., adherence, quality of delivery, dosage, and implementor knowledge or experience), student engagement, and treatment analysis (i.e., treatment differentiation and analysis of implementation fidelity). Overall, findings indicate many author teams reported adherence data and dosage data, but significantly fewer studies reported quality of delivery data, student engagement data, or treatment differentiation data. In addition, author teams were more likely to report at least one form of logistics, such as implementor support, than theories of change. Implications for research and practice are discussed.

Multi-fidelity Data Aggregation using Convolutional Neural Networks

Multi-fidelity data exist in almost every engineering and science discipline, which can be from simulation, experiments, and a hybrid form. High fidelity data are usually associated with higher accuracy and expense (e.g., high resolution experimental testing or finer scale simulation), while low-fidelity data are on the opposite side in terms of the accuracy and cost. Multi-fidelity data aggregation (MDA) in this study refers to the process of combining two or multiple sources of different fidelity data to have a high accuracy estimation and low computational cost. MDA has a wide range of application in engineering and science, such as multiscale simulation, multi-resolution imaging, and hybrid simulation-testing. This paper presents a novel framework named Multi-fidelity Data Aggregation using Convolutional Neural Networks (MDA-CNN) for multi-fidelity modeling. The MDA-CNN architecture has three components: multi-fidelity data compiling, multi-fidelity perceptive field and convolution, and deep neural network for mapping. This framework captures and utilizes implicit relationships between any high-fidelity datum and all available low-fidelity data using a defined local perceptive field and convolution. Most existing strategies rely on the collocation method and interpolation, which focuses on the single point relationship. The proposed method has several unique benefits. First, the proposed framework treats the multi-fidelity data as image data and processes them using CNN, which is very scalable to high dimensional data with more than two fidelities. Second, the flexibility of nonlinear mapping in neural network facilitates the multi-fidelity aggregation and does not need to assume specific relationships among multiple fidelities. Third, the proposed framework does not assume that multi-fidelity data are at the same order or from the same physical mechanisms (e.g., assumptions are needed for some error estimation-based multi-fidelity model). Thus, the proposed method can handle data aggregation from multiple sources across different scales, such as different order derivatives and other correlated phenomenon data in a single framework. The proposed MDA-CNN is validated using extensive numerical examples and experimental data with multi-source and multi-fidelity data. Discussions are given to illustrate the benefits and limitations of the proposed framework. Conclusions and future work are presented based on the observations in the current study.

Risk factors for stroke in penetrating carotid trauma-An analysis from the PROOVIT Registry.

Penetrating carotid injuries are associated with an up to 20% risk of stroke. This study evaluated patients in the American Association for Surgery of Trauma PROspective Observational Vascular Injury Trial, with the aim of determining factors associated with stroke and stroke or death. Penetrating extracranial carotid injuries in the American Association for Surgery of Trauma PROspective Observational Vascular Injury Trial registry from 2012 to 2020 were queried. Isolated external carotid injuries were excluded. Patients with documented postinjury in-hospital stroke were compared with those without. Significant predictors (p < 0.1) for stroke and stroke or death on univariate analysis were included in multivariate analyses. One hundred two patients from 17 institutions were included. Mean age was 35 ± 18 years, and 80% were male. Average Glasgow Coma Scale (GCS) score on presentation was 9 ± 5, with an Injury Severity Score [ISS] of 22 ± 13. Operative management occurred in 51% of patients who were significantly more hypotensive (systolic blood pressure: 109 vs. 131 mm Hg; p = 0.015) with a lower initial pH (7.17 vs. 7.31; p = 0.001) and presented with hard signs of vascular injury (74% vs. 26%; p < 0.001). Overall stroke rate was 17% (23% operative vs. 10% nonoperative, p = 0.076). Rate of stroke or death was 27% (64% operative and 36% nonoperative). On multivariate analysis, lower GCS (p = 0.05) and completion angiography (p = 0.04) were associated with stroke. Likewise lower GCS (p = 0.015) and ISS (p = 0.04) were associated with stroke or death. Penetrating carotid trauma undergoing operative management had a stroke rate of 23%. Low GCS on arrival and need for completion angiography are independently associated with postinjury in-hospital stroke, whereas low GCS on arrival and ISS were associated with stroke or death. The ideal treatment strategy remains elusive, thus a dedicated multicenter study may help to achieve higher fidelity data on this rare but devastating injury. Prognostic and Epidemiological, Level III.