70040 research outputs found
Sort by
Development of an internationally agreed national minimum dataset for low back pain: a modified Delphi study
No full textBACKGROUND: Low back pain is the leading cause of disability worldwide and the burden is expected to rise due to increases in ageing and population growth. The 2018 Lancet Low Back Pain Series proposed urgent actions to reverse the rising trend of disability due to low back pain including the need for a set of data indicators to routinely monitor progress in achieving these actions worldwide.
PURPOSE: To reach international consensus on a national minimum dataset of low back pain indicators that could be used across all countries to monitor progress in improving care and reducing disability from low back pain.
STUDY DESIGN: Modified Delphi study.
SAMPLE: Nineteen participants attended a preliminary workshop at the 2023 International Back and Neck Pain Forum, The Netherlands. Subsequently, 305 and 339 participants (researchers, clinicians, policy makers, educators, and consumers) completed Delphi surveys Rounds 1 and 2, respectively.
OUTCOME MEASURES: A 9-point Likert scale rated importance and feasibility of low back pain indicators (1 = not important or feasible; 9 = extremely important or feasible, 0 unsure). In Round 2, indicators that achieved consensus on importance and feasibility were ranked "most important" (top-ranked) to "least important."
METHODS: Workshop participants were divided into four groups and asked to independently consider the importance and feasibility of 105 indicators identified from previous reviews divided into six themes. Participants could remove indicators considered unimportant or not feasible. This required unanimous agreement from independent workshop groups. Participants could also suggest improvements to the wording of indicators, the best unit of measure and additional missing indicators.
RESULTS: Thirty-eight indicators were recommended by workshop participants for inclusion in the Delphi study. Survey responses over two rounds reached consensus for 21 indicators (11 burden, 10 care) ranked from most to least important after reaching consensus on importance and feasibility in Round 1. Number of work days lost and number of opioid prescriptions for low back pain were the highest ranked indicators for burden and care, respectively. Importance rankings were similar across subgroups comparing high-income and low- and middle-income countries, and consumers and non-consumers.
CONCLUSION: We reached international consensus that 21 indicators could be used to monitor progress in improving care and outcomes for people with low back pain globally. Future work is needed to confirm the acceptability and feasibility of these indicators across countries, and, if implemented, to determine their value over time
A scoping review of clinical decision support systems for child and adolescent mental health
No full textThis scoping review maps the literature on clinical decision support systems (CDSSs) that aid clinicians in treatment decision-making for youth mental health. The objectives were to identify CDSSs used in youth mental health services, as well as grade and evaluate these systems with regard to performance, usability, and clinical utility. Electronic databases were systematically searched for publications related to the development, implementation, or evaluation of clinical decision support systems. Identified CDSSs were grouped according to user setting (primary or mental health care) and three functional areas including screening and assessment, risk prediction, and treatment recommendation. We assessed and graded the systems based on phases of evaluation, as well as the level and direction of evidence. This review identified 10 CDSSs at various stages of development and implementation used across primary and mental health care settings. Of these, one CDSS is in development phase, four have been implemented, and five have evidence of post-implementation investigations. Outside of primary care settings, few CDSSs exist to support clinicians making clinical treatment decisions for youth mental health. Availability of CDSSs is limited and country-specific. In terms of functionality, successful implementation is evident in the functional area of screening and assessment, with more work required in the functional areas of risk prediction and treatment recommendation to assist clinicians in providing youth mental health care. There is potential for machine learning in developing these functional areas
Estimating Blood Donation Eligibility, Attitudes, and Knowledge among the Australian Population
No full textBackground
A safe and sufficient blood supply is a cornerstone of healthcare systems, yet its sustainability depends on the eligibility and willingness of the population to donate blood. This thesis explores the prevalence of eligible blood donors in Australia and the impact of exclusion factors on the donor-eligible pool; data which can be used by blood collection agencies to model the impact of changes to the criteria. This thesis also evaluates the impact of deferral policies on the eligibility of gay, bisexual, and other men who have sex with men (GBM), and the acceptability of proposed gender-neutral assessments (GNA). Lastly, it examines public knowledge of blood donation eligibility and criteria, and common exclusion factors for blood donation.
Methods
Data were derived from a nationally representative survey of Australians aged over 18 years old. Respondents provided information about their eligibility, donation history, comfort with GNA questions, and knowledge of deferral criteria. Results were weighted to match the Australian general population.
Results
An estimated 57.3% of Australians aged 18–74 was estimated to be eligible to donate blood under current criteria, yet only 14.2% of this group had donated in the past two years, revealing a large untapped donor pool of potential donors. Among GBM, current deferral policies exclude a large proportion from donation, with only 40.2% eligible under existing criteria. A transition to new criteria for plasma donation and hypothetical GNA scenarios could increase eligibility for GBM to 73.6%, potentially enhancing donation rates.
Most respondents expressed comfort with proposed GNA questions about new sexual partners (73.1%) and anal sex (64.0%), although discomfort was more pronounced among certain subgroups, such as individuals identifying as Muslim, Eastern Orthodox, or born in the Middle East, Southern Europe, or Asia. Additionally, the study revealed significant gaps in public knowledge of blood donation criteria, with 21.4% of respondents unaware of their existence. Misunderstanding of deferral periods—either overestimating or underestimating—was prevalent. Underestimating deferral periods could lead to individuals attempting to donate but receive a deferral. While deferrals are essential to ensure the safety of the donor and the blood supply, they can negatively impact on donor return rates. Conversely, overestimating deferral periods may unnecessarily discourage potential donors who mistakenly believe they are ineligible, from donating blood.
Conclusion
This research highlights the need for blood donation criteria to undergo regular review in line with scientific evidence and epidemiology to avoid unnecessary deferrals. There is a need for targeted education campaigns to improve public understanding of blood donation criteria and to address cultural sensitivities surrounding proposed changes to deferral policies and GNA. Addressing these gaps and fostering inclusivity can enhance donor retention and recruitment, contributing to a more sustainable and effective blood donation system in Australia
High Order Implicit Time Integration Methods in Structural Dynamics
No full textStructural dynamic analysis is essential for predicting the behavior of structures subjected to dynamic loads. It enables engineers to design safer and more reliable systems. Time integration methods play a vital role in the transient analysis, where accurate representation of low-frequency responses and controlled dissipation of high-frequency components are critical. These methods discretize the continuous time domain into smaller intervals, allowing the numerical solution of equations of motion to predict the behavior of structures under dynamic loading. This thesis introduces a series of novel high-order implicit time integration schemes that are able to achieve high accuracy, controllable numerical dissipation, and computational efficiency.
The first scheme incorporates user-controlled numerical dissipation through a spectral radius parameter in the high-frequency limit. By employing a weighted Padé expansion mixing the diagonal and sub-diagonal Padé expansions of the matrix exponential, the scheme is able to transit between A-stable (minimal dissipation) and L-stable (maximum dissipation) configurations. The algorithm is designed to share similar format as the Newmark method. Through numerical simulations, the scheme shows the effectiveness in minimizing spurious oscillations while preserving low-frequency accuracy.
The second schemes use rational approximations of the matrix exponential in the solution of the equations of motion to have the same effective stiffness matrix in all sub steps. For each sub-step, the implicit equation is in the same form as the Newmark method. Two scenarios are developed: M-schemes allow for adjustable numerical dissipation via a user-specified parameter, while (M+1)-schemes offer higher accuracy with non-adjustable dissipation.
The third scheme is built from the previous two methods, utilizing partial fractions to decompose the rational approximations. This method eliminates the need to compute the inversion of the mass matrix in the implicit equations. An efficient algorithm to compute the acceleration at the same order of accuracy as that in displacement and velocity is developed, which involves only vector operations.
Overall, a comprehensive framework for high-order implicit time integration methods in structural dynamics is established, which has the ability to offer both high accuracy and computational efficiency. Practical guidelines for parameter and time step selection are provided in structural dynamics modeling
Signal Control Optimisation for Continuous Flow Intersections with Connected and Autonomous Vehicles Considering Vulnerable Road Users
No full textThis research investigates the optimisation of traffic signal control for Continuous Flow Intersections (CFIs), also known as Displaced Left-Turn (DLT) intersections, which are designed to improve traffic flow and safety by reducing conflict points between left-turn and through movements. While previous studies have explored signal optimisation for CFIs, few have incorporated the impact of Connected and Autonomous Vehicles (CAVs) or considered pedestrian safety in their designs. This research addresses these gaps by developing and extending a robust signal optimisation framework in three key directions. First, a foundational two-step optimisation model for CFI signal control was developed and tested under more than 300 demand scenarios, demonstrating significant improvements in intersection efficiency. Compared to a state-of-practice benchmark, the proposed model reduced average vehicle delays by 17% and queue lengths by 32%, particularly excelling in scenarios with high left-turn demand. Building on this foundation, the second part of the study introduces CAVs into the optimisation framework. A real-time model was developed to account for mixed autonomy traffic, dynamically adjusting signal timings based on real-time data from CAVs to estimate lane-specific saturation flow rates. Simulation results across varying demand patterns showed the model's effectiveness, reducing average delays by 14% and queue lengths by 16% compared to a fixed signal control benchmark, highlighting the potential of CAVs to enhance traffic management at CFIs. The final part extends the model to incorporate pedestrian safety, a key factor often overlooked in CFI signal design. A chance-constrained optimisation framework was developed to manage the probability of pedestrian red-light violations, using a triangular distribution to model the relationship between pedestrian waiting times and violation risks. The model balances pedestrian safety with vehicle efficiency, and simulation experiments revealed that CFIs with refuge islands required shorter red times for pedestrians, offering valuable insights for safer intersection design. This research advances CFI signal optimisation by integrating real-time CAV data and pedestrian safety, offering a comprehensive and flexible framework for enhancing both efficiency and safety at complex intersections
PrEP use and willingness cascades among GBMSM in 15 Asian countries/territories: an analysis of the PrEP APPEAL survey
No full textIntroduction: Despite the high HIV incidence among gay, bisexual and other men who have sex with men (GBMSM) and the demonstrated effectiveness of HIV pre-exposure prophylaxis (PrEP), PrEP is not accessible at scale across Asia. To help inform future scaling efforts, our study aimed to examine PrEP use and willingness to use among GBMSM to identify opportunities and target groups for upscaling PrEP. Methods: The PrEP APPEAL survey was a cross-sectional survey, promoted through online advertising and community organizations, from May to November 2022. Eligible participants were adult GBMSM who self-identified as HIV negative residing in Asia. We constructed two cascades: PrEP use (comprising awareness, lifetime use and current use of PrEP) and PrEP willingness among participants who were aware of PrEP but had never used it (comprising HIV exposure risk, willingness in PrEP and willingness to pay for PrEP). Multivariable logistic regression models identified factors associated with lifetime PrEP use and PrEP willingness. Results: Of 15,339 participants, 1440 were excluded due to missing data, leaving 13,899 for analysis. Most lived in large or capital cities (68.3%) and in lower-middle-income countries (45.1%). The median age was 30 (25−36) years old. For the PrEP use cascade, 82.2% (n = 11,427/13,899) of participants were aware of PrEP, 35.0% (n = 4000/11,427) had used it before and 70.1% (n = 2803/4000) of them were currently on PrEP. For the PrEP willingness cascade, 54.8% of (n = 4068/7427) PrEP-naïve participants engaged in one or more behaviours with a higher risk of HIV acquisition, 73.7% (n = 2996/4068) of them expressed willingness to use PrEP and 83.0% (n = 2487/2996) of them were willing to pay for PrEP. Multivariable logistic regression models identified system-level (PrEP availability, accessibility and affordability) predictors of PrEP use. Individual-level behaviours associated with higher HIV acquisition risks were associated with PrEP use and willingness. Conclusions: While PrEP uptake was suboptimal, there was high awareness and willingness in PrEP among GBMSM. This is encouraging for future scale-up efforts. Future PrEP programmes should address system-level barriers to support PrEP uptake
Gender and Perpetration in Histories of the Third Reich: Enduring Misrepresentations. An Analysis of the Histories of Female Perpetrators and their Representation Across Time and Modality.
No full textThis thesis seeks to examine the intersection of gender and atrocity by analysing the ways in which female perpetrators during the Third Reich and Holocaust have or indeed have not been represented in collective public memories and narratives. This work hypothesises that female perpetrators, either through their literal absence, comparative marginalisation, or depiction as monstrosity, are persistently framed by particular (erroneous) gendered norms of femininity. Within such a framework, female violence and complicity are either not permissible and thus absent from representation or preferentially depicted as aberration, typically in the guise of the monstrous “Female Nazi.” These representations and silences curate a collective, public, and cultural narrative that does not allow space for the lived histories of “ordinary women” who may express political agency or participate in atrocity. Using an inter-disciplinary approach, this thesis examines a cross section of women’s histories and their representation in war crimes trials, film, media, and museum spaces over time to expose a dominant and preferential structure that overwhelmingly designates men as perpetrators and women as victims. This work presents a breadth and depth of research that signals the need for a re-evaluation and corrective of the master historical canon regarding the motivations and behaviours of German women during the Third Reich, and its reciprocal public representations
Micronutrient Fortification of Wheat Flour and its Efficacy upon Nutrition Status and Growth in Papua New Guinean School Children
No full textMicronutrient fortification is a public health strategy employed globally to address population-level undernutrition; however, clinical trials are recommended to understand intervention efficacy.
A double-blind, controlled clinical trial (CCT) investigating the efficacy of wheat flour fortification was conducted in two urban schools (coded as School A and B, respectively) in the city of Lae, Papua New Guinea (PNG) from February through November 2018. Eligible students aged 6-12 years (N = 423) were allocated on a school cohort-basis to receive biscuits containing iron, zinc, thiamine, riboflavin, niacin, folic acid, B12 and vitamin A; or an unfortified biscuit, intended for a period of up to 100 school days. The treatment group (n=190) received nutrient levels commensurate with 75g daily intake of wheat flour fortified with reference to WHO recommendations for cereal staples and provided in the form of biscuits. This dissertation reports the investigation of anthropometric indices, iron and anaemia status, zinc, and vitamin B12 only.
Due to an unforeseen, and unavoidable school closure in the treatment group (School B) and associated feeding difficulties, the study ran for 97 days of feeding for the treatment cohort, ending on Friday 9th November 2018, and the control group (School A) ran for 83 days, concluding on Friday 17th August 2018.
At baseline, for the entire study population, anaemia prevalence was 60.0%, Iron deficiency anaemia (IDA) 12.5%, vitamin B12 deficiency 10.3%, and zinc deficiency 0.99%. Stunting prevalence was 9.6%, and underweight 10.6%.
At follow-up, a reduction in anaemia was observed in both schools (P < .01) and subsequently, a highly significant decrease in IDA within the control group (P = .008), but not within the treatment group (P = .227). An increase in stunting (HAZ <-2) was observed in the control group (P < .001), and the treatment group, but this was not statistically significant (P = .180).
Of note, was the highly significant increase in the prevalence of zinc deficiency within both schools at follow-up; 22.0% in the control group, and 37.5% in the treatment group (P < .001). No significant differences were observed in other indicators.
Although this study is one of very few investigating the prevalence of nutritional and anthropometric deficits in children aged 6-12 years in PNG, there were many factors such as school closures, absenteeism and distribution issues leading to a loss to follow-up of 24.6% that interfered with the assessment of the efficacy of fortification. The high rates of anaemia and zinc deficiency suggest the continuing need for investigation into the underlying aetiology, which may extend beyond dietary and lifestyle factors, and address non-communicable disease if mandatory fortification of staple foods is to be effective
Learning-Based Approaches for Optimal Control and Resilient Operation of Battery Energy Storage Systems in Cyber-Physical Environments
No full textThe increasing deployment of Battery Energy Storage Systems (BESS) is transforming modern power grids by providing essential support for the integration of renewable energy sources, enhancing grid stability, and enabling more efficient energy management. As BESS become more widespread, they are required to perform complex functions, such as balancing supply and demand, regulating voltage, and managing frequency fluctuations. However, these advanced roles introduce significant challenges, ensuring that BESS operate optimally under varying conditions and in the face of growing cybersecurity threats. Additionally, the accurate estimation of the State of Health (SoH) of BESS is critical for maintaining their long-term performance and preventing costly failures. This thesis aims to address these multifaceted challenges by developing and integrating innovative strategies that optimize BESS control, fortify their cybersecurity, and improve SoH estimation. By leveraging learning-based technologies such as deep reinforcement learning (DRL), meta-learning, and machine learning, this research aims to push the boundaries of BESS capabilities, ensuring they remain reliable, efficient, and secure as they become increasingly integral to the future energy landscape.
The first major contribution of this thesis is the development of a DRL framework tailored for the adaptive and robust control of BESS. This framework addresses the dynamic complexities and uncertainties inherent in BESS operations, outperforming traditional gradient-based methods in optimizing system performance under various operational scenarios. A hierarchical DRL approach is introduced, ensuring that BESS can effectively manage their dynamic behavior while maintaining safe and efficient operations across diverse conditions.
The second contribution is the introduction of a DRL-inspired Growth Meta-heuristic Optimizer, designed in the form of a meta-learning algorithm for multi-objective optimization within Cyber-Physical-Social Systems (CPSS). This novel optimizer is capable of learning from prior optimization tasks, improving its performance over time. It effectively balances conflicting objectives, adheres to operational constraints, and significantly enhances overall system performance. Extensive simulations demonstrate the practical utility of this meta-learning optimizer in real-world energy management scenarios, highlighting its superiority over existing methods.
Addressing the increasing cybersecurity threats posed to BESS, the third contribution of this thesis focuses on the development of a comprehensive framework for detecting and mitigating DRL-based cyber-attacks. As energy systems become more digitalized and interconnected, safeguarding BESS from these sophisticated threats is critical to maintaining their integrity and reliable operation. This framework provides a robust defense mechanism, ensuring that BESS can continue to operate securely within CPSS.
Finally, this thesis advances the field of SoH estimation for BESS by employing vision transformers and Gaussian process models. This approach offers a highly accurate and reliable method for monitoring battery health over time, which is crucial for extending the lifespan and improving the efficiency of BESS.
Through the integration of these methodologies, this thesis provides a comprehensive solution to the pressing challenges faced by BESS in modern power systems. The research contributes significantly to the development of resilient, efficient, and secure energy storage systems, supporting the sustainable evolution of the energy landscape
Swimming in Data: Using AI and distance sampling to improve the analysis of remote underwater videos in estimating fish abundance
No full textWith the boom in technological advancements, remote underwater monitoring videos have become a popular method for sampling fish assemblages by marine ecologists, providing cost-effective access to large amounts of non-invasive monitoring data in previously inaccessible locations. However, the widespread uptake of this method is hindered by time intensive manual data extraction techniques that are typically analysed using biased and underpowered metrics, complicating the inference required for marine fisheries management. This thesis makes three contributions, to improve the utility of Remote Underwater Videos (RUVs).
Firstly, animal movement behaviour is exploited to enhance automated AI object detection of taxa in underwater monitoring videos. My method can also be used with terrestrial camera traps, so this use case is additionally investigated. My findings indicate that datasets with limited training data benefit from incorporating movement information into the detection pipeline.
Secondly, utilising AI predictions supplemented with Monte Carlo Markov Chain simulations, I assess the properties of the popular metric MaxN against the rarely used MeanCount to analyse underwater monitoring video data, providing key insights for marine ecologists. I find that estimates of MaxN are biased towards the movement and schooling behaviour of a species, as well as the length of a video, whilst MeanCount is unaffected by these factors and has greater power to detect effects of interest. However, both metrics are biased by imperfect detection (e.g. low water visibility), and although correlated with abundance, they do not estimate absolute abundance.
Finally, by extracting distance information from stereo remote underwater videos, I apply distance sampling techniques to develop methods for estimating fish density that account for imperfect detection. Distance sampling, widely used in terrestrial wildlife studies, records distance from the observer for each detection, and hence estimates how the probability of detection varies with distance from the observer. This allows the estimation of absolute abundance, correcting for imperfect detection. Despite stereo RUVs containing distance information and being widely used by marine ecologists to obtain fish length and biomass, they have never been analysed in this fashion before. My simulations show that abundance estimates derived from distance sampling are found to be unbiased under imperfect detection, thus it is possible to obtain accurate estimates of fish density in a broad range of practical settings, using equipment and software that marine ecologists are already frequently implementing, but in new ways