143174 research outputs found
Sort by
The impact and return-on-investment of evidence-based practice in conservation and environmental management: a machine learning-assisted scoping review protocol
No full textEvidence-based Practice (EBP) is a vital principle, with its origins in the 1970s, that has transformed the disciplines of medicine and healthcare. The use of best available evidence to inform decisions and best practice has since spread across other disciplines, including in the environmental sciences through evidence-based conservation and environmental management. However, ironically there only appears to be a single scoping review on the impacts and return-on-investment of EBP in healthcare and it is unclear whether any such evidence exists in the broad field of conservation and environmental management. In this scoping review, we aim to explore the extent to which evaluations of the impacts and return-on-investment of EBP and evidence use have been conducted in conservation and environmental management on both human and environmental outcomes. We will search at least ten different electronic bibliographic platforms, databases, and search engines for published and grey literature, from 1992 to 2025 – there will be no geographical or language restrictions on the documents included. A machine learning-assisted review process will be followed using open source tools (ASReview and SysRev) and following the comprehensive SYstematic review Methodology Blending Active Learning and Snowballing (SYMBALS). The findings from the scoping review will be useful to inform organisations and practitioners considering implementing EBP on its benefits and costs and will also highlight potential research gaps on the impact of EBP and evidence use
Overcoming catastrophic forgetting in radar and LiDAR object detection in rain via layer freezing and data augmentation
No full textAdvanced Driver-Assistance Systems (ADAS) use sensors like radar, LiDAR, and cameras for reliable vehicle perception in different weather conditions. While LiDAR and cameras offer high-resolution perception in clear weather, radar excels in adverse conditions such as low light, fog, or rain. Adapting systems trained on clear-weather data to cope with adverse weather often causes catastrophic forgetting, significantly reducing their initial performance after re-training. Unsupervised domain adaptation (UDA) techniques aim to address this but are complex. In this paper, we examine catastrophic forgetting effects on radar and LiDAR, proposing methods to reduce it: model freezing, pre-training with mixed data, and adding simulated data. Our experiments on the well-established RADIATE dataset show these methods improve clear-weather retention and rain detection, with radar showing a 6.59% reduction in forgetting and a 17.19% rain detection gain, and LiDAR a 13.62% reduction in forgetting and 24% improvement with simulations
Comparing artificial intelligence- vs clinician-authored summaries of simulated primary care electronic health records
No full textObjective
To compare clinical summaries generated from simulated patient primary care electronic health records (EHRs) by GPT-4, to summaries generated by clinicians on multiple domains of quality including utility, concision, accuracy, and bias.
Materials and Methods
Seven primary care physicians generated 70 simulated patient EHR notes, each representing 10 patient contacts with the practice over at least 2 years. Each record was summarized by a different clinician and by GPT-4. artificial intelligence (AI)- and clinician-authored summaries were rated blind by clinicians according to 8 domains of quality and an overall rating.
Results
The median time taken for a clinician to read through and assimilate the information in the EHRs before summarizing, was 7 minutes. Clinicians rated clinician-authored summaries higher than AI-authored summaries overall (7.39 vs 7.00 out of 10; P = .02), but with greater variability in clinician-authored summary ratings. AI and clinician-authored summaries had similar accuracy and AI-authored summaries were less likely to omit important information and more likely to use patient-friendly language.
Discussion
Although AI-authored summaries were rated slightly lower overall compared with clinician-authored summaries, they demonstrated similar accuracy and greater consistency. This demonstrates potential applications for generating summaries in primary care, particularly given the substantial time taken for clinicians to undertake this work.
Conclusion
The results suggest the feasibility, utility and acceptability of using AI-authored summaries to integrate into EHRs to support clinicians in primary care. AI summarization tools have the potential to improve healthcare productivity, including by enabling clinicians to spend more time on direct patient care
On Gaussian multiplicative chaos and random geometry
No full textThis thesis explores three interconnected topics: Gaussian multiplicative chaos (GMC), random planar maps, and random geometric objects in dimensions three and higher. The theory of GMC constructs random measures by formally exponentiating a log-correlated Gaussian field scaled by a parameter . These measures undergo a phase transition at , where is the ambient dimension. The regime is supercritical. First, we investigate supercritical GMC measures. Specifically, we analyse the local structure of -scale invariant fields around their extremal points and gain a deeper understanding of the freezing phenomenon in supercritical GMC. We also establish the universality of supercritical GMC: for general convolution approximations of log-correlated Gaussian fields, the supercritical chaos measures converge to a nontrivial limit, which depends on the regularisation only through a multiplicative constant. Additionally, we rigorously prove that subcritical GMC satisfies the multifractal formalism. Next, we study random planar maps and their connection to Liouville quantum gravity (LQG), a canonical model of fractal surfaces with random geometry. We focus on the Smith embedding of planar maps, defined via discrete harmonic functions on the set of primal and dual vertices. We show that if a sequence of embedded planar maps satisfies an invariance principle assumption, then the a priori embedding is asymptotically close to an affine transformation of the Smith embedding. As a consequence, we establish the convergence of the Smith embedding of mated-CRT maps to LQG. Finally, we extend to higher dimensions certain random geometric objects originally introduced in the two-dimensional setting. Specifically, we construct the higher-dimensional analogue of the Liouville Brownian motion and analyse its spectral dimension. Furthermore, in arbitrary even dimensions, we construct the higher-dimensional analogue of the quantum cone.Open Acces
Investigating electric fields and energy conversion in earth's turbulent magnetosheath
No full textTurbulence plays a key role in plasma dynamics and heating throughout the Universe. Complex and unpredictable nonlinear motion leads to a transfer of energy from large to small scales, where it is dissipated into heat. In the absence of collisions, turbulence proceeds to microphysical scales where it is not known how reversible kinetic processes lead to net dissipation. Earth's magnetosheath provides an opportunity to directly probe this open question using in situ measurements. This thesis presents an investigation into the fundamental physics of turbulence using a database of magnetosheath observations by NASA's Magnetospheric Multiscale (MMS) mission. The comparative importance of different electric field contributions is studied for 60 magnetosheath intervals, revealing that how the nonlinear dynamics operate depends on both the ambient plasma conditions and the turbulence driver. The relationships uncovered provide observational constraints on turbulence theory, bounding the applicability of magnetohydrodynamic and kinetic Alfvén wave physics, with applications to other plasma regimes and simulation studies. The role of these electric field dynamics in facilitating turbulent dissipation is explored by quantifying the non-ideal energy conversion related to specific kinetic processes. It is found that energy conversion is enhanced in regions susceptible to anisotropy-driven kinetic microinstabilities, suggesting these mechanisms play a fundamental role in dissipating turbulent energy. The interpretation is that turbulence drives conditions to become locally unstable, triggering instabilities which redistribute the free energy in favour of growing wave modes. Despite these processes being reversible, the net transfer of energy is from the fields into the particles, consistent with net dissipation. Turbulence is also found to generate significantly non-Maxwellian velocity distributions, indicating that it provides additional sources of free energy whereby instabilities can thermalise the plasma.Open Acces
Addendum to consensus opinion from the International Deep Endometriosis Analysis (IDEA) group: sonographic evaluation of superficial endometriosis
No full textTraditionally, laparoscopy was considered to be the gold standard for the examination of endometriotic lesions, because it allows their direct visualization. Several international and national guidelines have now shifted their focus to non-invasive imaging-based diagnosis of deep endometriosis in preference to surgery, while laparoscopy is used to diagnose superficial endometriosis in patients with painful symptoms and negative ultrasound and/or magnetic resonance imaging findings for ovarian or deep endometriosis. In recent years, however, the role of gynecological ultrasound in the diagnosis of superficial endometriosis has been studied more extensively. The purpose of this addendum to the International Deep Endometriosis Analysis consensus opinion is to describe a standardized ultrasound protocol for the diagnosis of superficial endometriosis and to highlight the sonographic characteristics of these lesions. © 2025 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology
Biomechanics of soft tissue-device interactions-an investigation into skin and arterial tissue
No full textThis thesis adopts a multidisciplinary approach, integrating mathematical modelling, experimental testing, and numerical simulations, to investigate the complex interactions between soft tissues and medical devices. The work focuses on two main areas: characterising the contact behaviour of skin and mouse arterial tissue during indentation, and analysing artery–stent interactions using finite element simulations.
In Chapter 3, a novel method is developed to determine the Ogden hyperelastic parameters of soft tissues using indentation experiments on artificial skin. By incrementally applying Hayes’ model under the assumption that soft tissues behave approximately linearly at strain increments ≤1%, an equivalent shear modulus is calculated for each increment. Principal stresses are derived using Hooke’s law, and combined with principal strains to obtain compound stress–strain curves. The Ogden parameters are then extracted via curve fitting. Total contact force is computed by summing incremental Hayes forces.
Chapter 4 applies this methodology to mouse arterial tissue and introduces an approach to determine the equivalent thickness for Hayes’ model. The optimal thickness is defined as the one that minimises the variance of the calculated equivalent shear moduli. Based on 26 indentation tests from five mouse artery samples, the estimated thickness ranges from 0.05 to 0.5 mm, with an average shear modulus of 1.22 kPa.
Chapter 5 employs finite element simulations to study artery–stent interactions. The arterial model includes intima, media, adventitia, and a lipid pool, each assigned anisotropic hyperelastic properties. Contact is defined using hard contact with friction coefficients from 0.01 to 0.1. Simulations investigate different stent surface geometries. Results show stress and strain concentration in three critical areas: the contact interface, the cap’s right shoulder, and around the lipid pool. Larger friction and high-amplitude stent surfaces increase stresses, while small-amplitude convex surfaces reduce average shear stress, suggesting their potential to reduce arterial injury during stenting.Open Acces
Comparative analysis of sporadic and colitis associated cancers
No full textColitis associated cancers (CACs) are a sequelae of inflammatory bowel disease (IBD) which is difficult to detect during surveillance, leading to high rates of interval cancers. Despite the well defined risk factor of IBD in the development of cancer, and the frequent application of metataxonomic and mass spectrometry techniques for investigation to IBD, little is known about the structure and function of the microbiota in CAC, features of tissue level metabolism, or whether patient risk can be identified.
Open access multi-omics data in patients with long disease duration (>10 years) was analysed to identify potential risk factors. Potential alterations in stool bile acid concentrations not explained by diet were identified, as well as transcriptional differences in uninvolved ileum between patients with over or under 10 years IBD duration. However these may be related recruitment sites.
A novel tool to predict the function of the microbiota was developed and benchmarked against state of the art techniques, finding that prediction accuracy is a function of metagenome size, and remains stable with growing database size.
A prospective cohort of patients undergoing colonic resection for IBD, CAC, and sporadic CRC were recruited and a comprehensive analysis of the tissue-associated microbiota undertaken by 16S rRNA gene sequencing. A depletion of Fusobacterium in CAC tumour sites was confirmed, and potential taxa of interest identified.
Mass spectrometry imaging of CAC tumours was performed for the first time, with features suggestive of less intra-tumour heterogeneity than sporadic CRC. Supervised classification of tumours was performed using both metataxonomic and metabolomic datasets, but both performed poorly.Open Acces
Determining genotype and antimicrobial resistance of Salmonella Typhi in environmental samples by amplicon sequencing
No full textBackground
Estimates of the burden of typhoid fever due to Salmonella enterica serovar Typhi (S. Typhi) rely on data from clinical surveillance, which is rarely done in low income settings and is also limited by the poor sensitivity of the assays used and the reliance on health seeking by patients. Environmental surveillance for S. Typhi shed by symptomatic and asymptomatic individuals in wastewater offers a sensitive surveillance tool that could help to inform burden estimates. Sequencing S. Typhi direct from wastewater concentrates has the potential to identify circulating genotypes and associated antimicrobial resistance (AMR) genes, supporting public health interventions such as vaccination and antimicrobial usage.
Methodology and principal findings
We designed a multiplex targeted amplicon sequencing protocol for genotyping and determining AMR in S. Typhi from wastewater samples, targeting SNPs that identify genotypes of interest and both chromosomal and plasmid-borne AMR. PCR products were sequenced using the Oxford Nanopore Technologies (ONT) MinION, and genotypes and AMR identified using the GenoTyphi program.
We tested this approach on samples from south India from both hospital outflow and wastewater collected from the community. All samples tested were suspected to be positive for S. Typhi following quantitative PCR for ttr, tviB, and staG gene targets. Out of 110 samples tested we were able to determine a genotype and/or AMR for 8. All samples that gave a genotype call suggested a genotype consistent with those found in clinical cases in India during the same time period and produced consensus sequences that clustered with S. Typhi when included in a phylogenetic tree.
Conclusions
In this study, we provide proof of concept data for amplicon sequencing of S. Typhi in wastewater which with further optimisation could be used to complement clinical surveillance data or provide data on S. Typhi presence in the absence of clinical surveillance. This information can inform public health interventions, and the concept could be applied to other pathogens of interest for genotyping from environmental surveillance samples
The effect of sexual and gender minority violence on depression, hazardous drinking, condom use, and HIV acquisition: an individual participant data meta-analysis of the CohMSM, HPTN 075, and Anza Mapema cohort studies in Africa
No full textSome sexual and gender minorities (SGM), including men who have sex with men and transgender women, are disproportionately vulnerable to HIV. Many SGM in Africa report experiencing verbal or physical violence due to their sexual and/or gender identities or behaviours. The pathways linking such SGM violence to HIV acquisition are complex. We described experiences of verbal and physical SGM violence and explored pathways to HIV acquisition among SGM assigned male sex at birth using a two-stage individual participant data meta-analysis of three African cohort studies: CohMSM (Burkina Faso, Côte d’Ivoire, Mali, Togo), HPTN 075 (Kenya, Malawi, South Africa), and Anza Mapema (Kenya). SGM violence was assessed at baseline and follow-up visits. We fit log-linear sequential conditional mean models using generalised estimating equations to estimate risk ratios linking SGM violence, moderate-to-severe depressive symptoms, hazardous drinking, condom use, and HIV acquisition, adjusted for baseline confounders and previous exposure and outcome. We pooled study estimates using random effects meta-analysis. SGM violence, mostly verbal, was reported by 36% (570/1590) participants at baseline (past 6–12 months), and 20% (321/1590) during the first year of follow-up (past 3–6 months). Baseline SGM violence was not associated with HIV acquisition (pooled adjusted risk ratio [aRR] = 1.0, 95% CI 0.5–1.9). During follow-up, SGM violence also showed no clear relationship with HIV, but was linked to depressive symptoms at the same visit (pooled aRR = 1.7, 1.3–2.1), in turn associated with hazardous drinking (pooled aRR = 1.4, 1.1–1.7). Impacts on condom use were inconclusive. SGM in Africa face high rates of violence, which are associated with depressive symptoms and hazardous drinking–potential routes to HIV vulnerability. While our study did not conclusively demonstrate higher HIV incidence among SGM reporting violence, interventions to reduce violence and support mental health remain crucial