Queen Mary, University of London

Recently self-supervised learning (SSL) has achieved impressive performance in computer vision (CV) and natural language processing (NLP) tasks, and some early attempts are made in the area of finance. In this paper, we apply SSL to extract features from financial time series data, and use those features to measure the similarities between assets in the market. As similarity measurement is the key to portfolio diversification, we consider two portfolio optimisation problems: index tracking (IT) and minimum variance portfolio (MVP), with the additional diversification terms linked to different similarity measurements, which are sourced from different SSL algorithms. Both IT and MVP are both convex optimisation problems with deterministic solutions, therefore the performance difference is traced back to SSL algorithms, rather than other factors. Extensive experiments are conducted with eight SSL algorithms, and the analysis of the results of the experiments demonstrates the advantages of SSL over non-SSL alternatives.

Absolute camera pose regression typically estimates the position and orientation of a camera solely based on the captured image, trained with a convolutional backbone with multilayer perceptron heads for a single reference scene only. Recently, leading pose regression results have been achieved on multiple datasets by extending this approach to learn multiple scenes by leveraging data from different modalities, especially by fusing RGB and point cloud data. In this work, we propose to use cross-attention Transformers to learn multi-scene absolute camera pose regression, where cross-attention modules are used to aggregate activation maps with self-attention from different data modalities and convert latent features and scene index into candidate pose predictions. This mechanism allows our model to focus more on the general localization features. We evaluate our approach on the popular indoor benchmark dataset 7-scenes and compare it against both state-of-the-art (SOTA) single-scene and multiple absolute pose regression models. Our main result is the rotation accuracy is improved using our method more than slightly improved position accuracy compared to existing multi-scene methods.

A well-known theorem of Philip Hall states that if a group 𝐺 has a nilpotent normal subgroup 𝑁 such that G / N ′ G/N^{\prime} is nilpotent, then 𝐺 itself is nilpotent. We say that a group class 𝔛 is a Hall class if it contains every group 𝐺 admitting a nilpotent normal subgroup 𝑁 such that G / N ′ G/N^{\prime} belongs to 𝔛. Examples have been given in [F. de Giovanni, M. Trombetti and B. A. F. Wehfritz, Hall classes of groups, to appear] to show that finite-by-𝔛 groups do not form a Hall class for many natural choices of the Hall class 𝔛. Although these examples are often linear, our aim here is to prove that the situation is much better within certain natural subclasses of the universe of linear groups.

Advances in image reconstruction using either single or multimodality imaging data provide increasingly accurate three-dimensional (3D) patient’s arterial models for shear stress evaluation using computational fluid dynamics (CFD). We aim to evaluate the impacts on endothelial shear stress (ESS) derived from a simple image reconstruction using 3D-quantitative coronary angiography (3D-QCA) versus a multimodality reconstruction method using optical coherence tomography (OCT) in patients’ vessels treated with bioresorbable scaffolds. Seven vessels at baseline and five-year follow-up of seven patients from a previous CFD investigation were retrospectively selected for a head-to-head comparison of angiography-derived versus OCT-derived ESS. 3D-QCA significantly underestimated the minimum stent area [MSA] (-2.38mm²) and the stent length (-1.46 mm) compared to OCT-fusion method reconstructions. After carefully co-registering the region of interest for all cases with a sophisticated statistical method, the difference in MSA measurements as well as the inability of angiography to visualise the strut footprint in the lumen surface have translated to higher angiography-derived ESS than OCT-derived ESS (1.76 Pa or 1.52 times for the overlapping segment). The difference in ESS widened with a more restricted region of interest (1.97 Pa or 1.63 times within the scaffold segment). Angiography and OCT offer two distinctive methods of ESS calculation. Angiography-derived ESS tends to overestimate the ESS compared to OCT-derived ESS. Further investigations into ESS analysis resolution play a vital role in adopting OCT-derived ESS.

Patterning of two or more liquids, either homogeneous in each phase or mixed with particles (including biological matter, such as cells and proteins), by controlling their flow dynamics, is relevant to several applications. Examples include dynamic spatial confinement of liquids in microfluidic systems (such as lab-on-a-chip and organ-on-a-chip devices) or structuring of polymers to modulate various properties (such as strength, conductivity, transparency and surface finishing). State-of-the-art strategies use various technologies, including positioners, shakers and acoustic actuators, which often combine limited versatility of mixing with significant inefficiency, energy consumption, and noise, as well as tendency to increase the temperature of the liquids. Here, we describe a new kind of robotic mixers of liquids, based on electro-responsive smart materials (dielectric elastomer actuators). We show for the first time how an efficient soft robotic device can be used to produce, via combinations of rotations and translations, various spatial patterns in liquids and maintain them stable for a few minutes. Moreover, we show that, as compared to a conventional orbital shaker, the new type of robotic device can mix liquids with a higher efficacy (~ 94% relative to ~ 80%, after 8 min of mixing) and with a significantly lower increase of the liquids’ temperature (+ 1 °C relative to + 5 °C, after 6 h of mixing). This is especially beneficial when mixing should occur according to controllable spatial features and should involve temperature-sensitive matter (such as biological cells, proteins, pre-polymers and other thermolabile molecules).

Background and objectives Alemtuzumab demonstrated superior efficacy versus subcutaneous interferon (IFN) beta-1a in participants with relapsing-remitting multiple sclerosis in the 2-year CARE-MS I and II trials. Efficacy was maintained in the 4-year CARE-MS extension, during which alemtuzumab-treated participants (‘alemtuzumab-only’) could receive additional courses upon disease activity, and IFN-treated participants switched to alemtuzumab (‘IFN-alemtuzumab’). Participants who completed the CARE-MS extension could enroll in the open-label TOPAZ study which assessed safety and efficacy for 5–7 years (11–13 years after alemtuzumab/IFN initiation). Methods Participants received additional alemtuzumab courses as needed. Assessments included adverse events (AEs; primary outcome), annualized relapse rate (ARR), 6-month confirmed disability worsening [CDW; ⩾1.0-point Expanded Disability Status Scale (EDSS) score increase or ⩾1.5 if baseline EDSS = 0], and 6-month confirmed disease improvement [CDI; >1.0-point EDSS decrease (baseline score ⩾2.0)]. Results 43.5% of alemtuzumab-only participants from CARE-MS II and 54.2% from CARE-MS I received no additional alemtuzumab courses; 30.0% and 20.9%, respectively, received one additional course (the median). Incidences of AEs, including thyroid AEs and infections, declined over time. The safety profile of alemtuzumab was similar for participants who received zero, one, or two additional courses. For CARE-MS II participants, who had inadequate response to previous treatment, ARR remained low during Years 3–13 for the alemtuzumab-only [0.17; 95% confidence interval (CI) 0.15–0.20] and IFN-alemtuzumab (0.14; 0.11–0.17) groups. At Year 11, the proportions of participants who were either free from CDW or who had CDI were higher in the alemtuzumab-only group (58% and 49%, respectively) than in the IFN-alemtuzumab group (51% and 37%). For CARE-MS I participants, who were previously treatment-naïve, clinical outcomes remained improved, and no between-group differences were apparent. Conclusion Safety risks associated with alemtuzumab treatment declined over time. Clinical benefits were maintained up to 11–13 years, and most participants did not require more than one additional course. Clinicaltrials.gov identifiers NCT00530348; NCT00548405; NCT00930553; NCT02255656

Background Periprosthetic fractures are rare but devastating complications of knee replacement, often requiring complex surgery with substantial morbidity and mortality. It is not known how the fracture rates after total knee replacement (TKR) and unicompartmental knee replacement (UKR) compare. We performed the first matched study comparing TKR and UKR periprosthetic fracture rates. Methods This study involved 54,215 UKRs and 54,215 TKRs, identified in the National Joint Registry and Hospital Episodes Statistics database, which were propensity score-matched on patient and surgical factors. The International Classification of Diseases, Tenth Revision, (ICD-10) code M96.6 was used to identify periprosthetic fractures at ≤3 and >3 months postoperatively, as well as estimate rates at up to 10 years. Subgroup analyses were performed in different age groups (<55, 55 to 64, 65 to 74, and ≥75 years), body mass index (BMI) categories (normal, 18.5 to <25 kg/m ² ; overweight, 25 to <30 kg/m ² ; obese, 30 to <40 kg/m ² ; and morbidly obese, ≥40 kg/m ² ), and sexes. Results The 3-month fracture rate was 0.09% (n = 50) in the UKR group and 0.05% (n = 25) in the TKR group, with this difference being significant (odds ratio [OR], 2.0; p = 0.004). The rate of fractures occurring at >3 months was 0.32% (n = 171) in the UKR group and 0.61% (n = 329) in the TKR group (OR, 0.51; p < 0.001). At 10 years, the cumulative incidence of fractures was 0.6% after UKR versus 1% after TKR (OR, 0.68; p < 0.001). Fracture rates increased with increasing age, decreasing BMI, and female sex for both UKRs and TKRs. Conclusions The fracture risk was small after both UKR and TKR, with small absolute differences between implant types. During the first 3 postoperative months, the fracture rate after UKR was 0.1% and was about twice as high as that after TKR. However, over the first 10 years, the cumulative fracture rate after TKR was 1% and was almost twice as high as that after UKR. Fracture rates after both UKR and TKR were higher in women, patients ≥75 years of age, and patients with normal weight. Level of Evidence Therapeutic Level III . See Instructions for Authors for a complete description of levels of evidence.

Introduction Patient-reported outcomes (PRO) are currently collected from trial participants using paper questionnaires by the Clinical Trials and Statistics Unit at The Institute of Cancer Research (ICR-CTSU). Streamlining PRO collection using electronic questionnaires (ePRO) may improve data collection and patient experience. Here, we outline our protocol for a Study within a trial of electronic versus paper-based Patient-Reported oUtcomes CollEction (SPRUCE), which investigates the acceptability of ePRO in oncology clinical trials. Methods and analysis SPRUCE was developed alongside patient and public contributors. SPRUCE runs in multiple host trials with a partially randomised patient preference design, allowing participants to be randomised or choose their preference of electronic or paper questionnaires. Questionnaires are scheduled in accordance with host trial follow-up. The primary objective will assess differences in return rates (compliance) between ePRO and paper PROs at the first timepoint post-host trial intervention in the randomised group. Paper PRO compliance is expected to be 90%. 244 randomised participants are required to exclude ≤80% compliance rates with ePRO (10% non-inferiority margin, with 80% power and one-sided alpha=0.05). SPRUCE aims to assess acceptability of ePRO in oncology clinical trials, establish whether ePRO is acceptable to ICR-CTSU trial participants and can capture complete PRO data, consistent with paper PROs. Ethics and dissemination The SPRUCE protocol (ICR-CTSU/2021/10074) was approved by the Coventry and Warwick Central Research Ethics Committee (21/WM/0223) on 21 October 2021. Results will be disseminated via presentations, publications and lay summaries. No participant identifiable data will be included. Trial registration SWAT169.

Application of an integrated spatial approach combining several sources of remote sensing data to include both channel and floodplain morphological and sedimentological impacts has identified the geomorphological effects of three extreme flood events during a single flood season (2019-20) along a 16km reach of the River Teme, UK. This combined approach allowed the assessment of in-channel pattern development, incision, and aggradation; lateral bank migration; and overbank sedimentation and scour by out-of-channel flows. Rates of change during the event period were compared with those in the previous 10 years. The approach also allowed the role of vegetation and cultivation, both bankside and out on floodplains, to be assessed with variations in the extent of riparian wood and channel slope driving contrasts in the extent of the response. The spatial impacts from such extreme events are highly localized, varied in kind, and can be considered for both rivers and floodplains together. Erosional effects were distinctively distributed and not simply contributions to ongoing meander development; channel aggradation was localised, and overbank sedimentation explicably patchy. In reaches without woody vegetation, differences in channel and floodplain slope, local floodplain relief as created by prior events, and the impact of man-made structures were factors that drove variations in flood response. This study strongly underlines the role of continuous riparian vegetation in maintaining bank stability and constraining lateral channel migration, but also to the potential influence of floodplain vegetation and planting for 'natural engineering' in the context of floodplains as well as channels in comparable environments. Maintenance of riparian vegetation in the context of landowner and Natural England conflict and management of future flood risk is important, but also highlighted is the need to consider the role of hedgerows and wider planting for constraining soil and riverbank erosion during flood events.

We introduce the notion of frequency-constrained substring complexity. For any finite string, it counts the distinct substrings of the string per length and frequency class. For a string x of length n and a partition of [n] in \(\tau \) intervals, \(\mathcal {I}=I_1,\ldots ,I_\tau \), the frequency-constrained substring complexity of x is the function \(f_{x,\mathcal {I}}(i,j)\) that maps i, j to the number of distinct substrings of length i of x occurring at least \(\alpha _j\) and at most \(\beta _j\) times in x, where \(I_j=[\alpha _j,\beta _j]\). We extend this notion as follows. For a string x, a dictionary \(\mathcal {D}\) of d strings (documents), and a partition of [d] in \(\tau \) intervals \(I_1,\ldots ,I_\tau \), we define a 2D array \(S=S[1\mathinner {.\,.}|x|,1\mathinner {.\,.}\tau ]\) as follows: S[i, j] is the number of distinct substrings of length i of x occurring in at least \(\alpha _j\) and at most \(\beta _j\) documents, where \(I_j=[\alpha _j,\beta _j]\). Array S can thus be seen as the distribution of the substring complexity of x into \(\tau \) document frequency classes. We show that after a linear-time preprocessing of \(\mathcal {D}\), for any x and any partition of [d] in \(\tau \) intervals given online, array S can be computed in near-optimal \(\mathcal {O}(|x| \tau \log \log d)\) time.

During COVID‐19 lockdowns in England, ‘key workers’ including factory workers, carers and cleaners had to continue to travel to workplaces. Those in key worker jobs were often from more marginalised communities, including migrant workers in precarious employment. Recognising space as materially and socially produced, this qualitative study explores migrant workers’ experiences of navigating COVID‐19 risks at work and its impacts on their home spaces. Migrant workers in precarious employment often described workplace COVID‐19 protection measures as inadequate. They experienced work space COVID‐19 risks as extending far beyond physical work boundaries. They developed their own protection measures to try to avoid infection and to keep the virus away from family members. Their protection measures included disinfecting uniforms, restricting leisure activities and physically separating themselves from their families. Inadequate workplace COVID‐19 protection measures limited workers' ability to reduce risks. In future outbreaks, support for workers in precarious jobs should include free testing, paid sick leave and accommodation to allow for self‐isolation to help reduce risks to workers’ families. Work environments should not be viewed as discrete risk spaces when planning response measures; responses and risk reduction approaches must also take into account impacts on workers’ lives beyond the workplace.

Earth’s orbit and rotation produces systematic variations in geomagnetic activity, most notably via the changing orientation of the dayside magnetospheric magnetic field with respect to the heliospheric magnetic field (HMF). Aside from these geometric effects, it is generally assumed that the solar wind in near-Earth is uniformly sampled. But systematic changes in the intrinsic solar wind conditions in near-Earth space could arise due to the annual variations in Earth heliocentric distance and heliographic latitude. In this study, we use 24 years of Advanced Composition Explorer data to investigate the annual variations in the scalar properties of the solar wind, namely the solar wind proton density, the radial solar wind speed and the HMF intensity. All parameters do show some degree of systematic annual variation, with amplitudes of around 10 to 20%. For HMF intensity, the variation is in phase with the Earth’s heliocentric distance variation, and scaling observations for distance largely explains the observed variation. For proton density and solar wind speed, however, the phase of the annual variation is inconsistent with Earth’s heliocentric distance. Instead, we attribute the variations in speed and density to Earth’s heliographic latitude variation and systematic sampling of higher speed solar wind at higher latitudes. Indeed, these annual variations are most strongly ordered at solar minimum. Conversely, combining scalar solar wind parameters to produce estimates of dynamic pressure and potential power input to the magnetosphere results in solar maximum exhibiting a greater annual variation, with an amplitude of around 40%. This suggests Earth’s position in the heliosphere makes a significant contribution to annual variations in space weather, in addition to the already well-studied geometric effects.

Fluidity, the ability of liquids to flow, is the key property distinguishing liquids from solids. This fluidity is set by the mobile transit atoms moving from one quasi-equilibrium point to the next. The nature of this transit motion is unknown. Here, we show that flow-enabling transits form a dynamically distinct sub-ensemble where atoms move on average faster than the overall system, with a manifestly non-Maxwellian velocity distribution. This is in contrast to solids and gases where no distinction of different ensembles can be made and where the distribution is always Maxwellian. The non-Maxwellian distribution is described by an exponent α\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha$$\end{document} corresponding to high dimensionality of space. This is generally similar to extra synthetic dimensions in topological quantum matter, albeit higher dimensionality in liquids is not integer but is fractional. The dimensionality is close to 4 at melting and exceeds 4 at high temperature. α\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha$$\end{document} has a maximum as a function of temperature and pressure in liquid and supercritical states, returning to its Maxwell value in the solid and gas states.

Total hip arthroplasty (THA) is a successful orthopaedic surgical procedure, and its longevity depends on bearing components and implant fixation. Optimizing polyethylene and ceramics has led to improved wear parameters and contributed to improved long-term outcomes. The present systematic review investigated whether time span from implantation, patient characteristics and performance status exert an influence on liner wear and head migration in THA. This study was conducted in conformity to the 2020 PRISMA guidelines. All the clinical investigations which reported quantitative data on the amount of liner wear and head migration in THA were considered. Only studies which reported quantitative data at least on one of the following patient characteristics were suitable: mean age, mean BMI (kg/m²), sex, side, time span between the index THA and the last follow-up (months) were eligible. A multiple linear model regression analysis was employed to verify the association between patient characteristics and the amount of liner wear and/or head migration. The Pearson Product-Moment Correlation Coefficient was used to assess the association between variables. Data from 12,629 patients were considered. The mean length of the follow-up was 90.5 ± 50.9 months. The mean age of patients at surgery was 58.4 ± 9.4 years, and the mean BMI was 27.2 ± 2.5 kg/m². 57% (7199 of 12,629 patients) were women, and in 44% (5557 of 12,629 patients) THAs were performed on the left. The mean pre-operative Harris hip score was 46.5 ± 6.0 points. There was evidence of a moderate positive association between the amount of liner wear and the time elapsed between the index surgery to the follow-up (P = 0.02). There was evidence of a moderate positive association between the amount of head migration and the time elapsed between the index surgery to the follow-up (P = 0.01). No further statistically significant association was found. The time elapsed between the index surgery to the follow-up was the most important factor which influence the head migration and liner wear in THA. Patients’ characteristics and preoperative physical activity did not influence the amount of head migration and liner wear.

With a proliferation of scholarly work focusing on populist, far-left, and far-right parties, questions have arisen about the correct ways to ideologically classify such parties. To ensure transparency and uniformity in research, the discipline could benefit from a systematic procedure. In this letter, we discuss how we have employed the method of ‘Expert-informed Qualitative Comparative Classification’ (EiQCC) to construct the newest version of The PopuList (3.0) – a database of populist, far-left, and far-right parties in Europe since 1989. This method takes into account the in-depth knowledge of national party experts while allowing for systematic comparative analysis across cases and over time. We also examine how scholars have made use of the previous versions of the dataset, explain how the new version of The PopuList differs from previous ones, and compare it to other data. We conclude with a discussion of the strengths and limitations of The PopuList dataset.

Background On World AIDS Day 2021, the UK Government committed £20 million to expand opt-out HIV testing in EDs in extremely high HIV prevalence (>5/1000) areas as part of their commitment to achieve zero new HIV infections, AIDS and HIV-related deaths by 2030. 34 EDs in London, Brighton, Greater Manchester and Blackpool were included. The initiative started in April 2022 and expanded to include hepatitis C (HCV) and hepatitis B (HBV) testing in collaboration with the HCV Elimination programme. Methods All adults undergoing blood tests in EDs had BBV testing (4th generation HIV test, HBV surface antigen and HCV antibody with reflex RNA if HCV antibody positive) unless they opted out. An opt-out approach was taken, based on successful pilots, to maximise uptake and minimise impact on ED workload. Testing information was displayed using accessible and translated posters in EDs. HIV/Sexual Health and Hepatology managed all reactive/positive results. Results By March 2023, 33 EDs had implemented HIV testing, 25 HCV and 19 HBV. From April 2022 through March 2023 there were 1,384,378 adult ED attendances with blood tests and 853,015 HIV, 346,041 HBV and 452,284 HCV tests were performed. Median test uptake increased from 51.1% (HIV), 16.3% (HBV), 23.7% (HCV) in April 2022 to 62% (HIV), 57% (HBV) and 62% (HCV) in March 2023.ED opt-out BBV testing identified 2002 people who were newly diagnosed (343 HIV, 1190 HBV, 484 HCV) and 473 who were previously diagnosed but not in care (HIV 209, HBV 156, HCV 108). Initial linkage to care was 339/552 (61%) for HIV (268/343 (78%) for new HIV diagnoses), 329/1346 (24%) for HBV and 292/592 (49%) for HCV. Conclusions Opt-out BBV testing in EDs has proven extremely effective for making new BBV diagnoses and re-engaging those previously diagnosed but not in care. We found very high rates of HBV. Initial linkage to care is encouraging and is expected to increase over time.

Aims Human islet transplantation as a therapy for type 1 diabetes is compromised by the loss of functional beta cells in the immediate post‐transplantation period. Mesenchymal stromal cells (MSCs) and MSC‐derived secretory peptides improve the outcomes of islet transplantation in rodent models of diabetes. Here we utilized a mouse model for human islet transplantation and assessed the effects of a cocktail of MSC‐secreted peptides (screened by MSC‐secretome for human islet GPCRs) on the functional survival of human islets. Methods Human islets from 9 donors (Age: 36‐57; BMI: 20‐35) were treated with a cocktail of human recombinant annexin A1 (ANXA1), stromal cell‐derived factor‐1 (SDF‐1/CXCL12) and complement component C3 (C3a). Glucose stimulated insulin secretion (GSIS) was assessed in static incubation and cytokine‐induced apoptosis was assessed by measuring caspase 3/7 activity. mRNA expression levels were determined by qPCR. Human islet function in vivo was assessed using a novel model for human islet transplantation into a T1D mouse model. Human islet function in vivo was assessed using islet transplantation under the kidney capsule of immunodeficient mice prior to STZ destruction of endogenous mouse beta cells to model T1DM. Results Pre‐treatment with a cocktail of MSC‐secreted peptides increased GSIS in vitro and protected against cytokine‐induced apoptosis in human islets isolated from nine donors. Animals transplanted with either treated or untreated human islets remained normoglycemic for up to 28 days after STZ‐administration to ablate the endogenous mouse beta cells, whereas non‐transplanted animals showed significantly increased blood glucose immediately after STZ administration. Removal of the human islet graft by nephrectomy resulted in rapid increases in blood glucose to similar levels as the non‐transplanted controls. Pre‐treating human islets with the MSC‐derived cocktail significantly improved glucose tolerance in graft recipients, consistent with enhanced functional survival of the treated islets in vivo . Conclusion Pre‐treating human islets before transplantation with a defined cocktail of MSC‐derived molecules could be employed to improve the quality of human islets for transplantation therapy for type 1 diabetes.

Hand surgeons have the potential to improve patient care, both with their own research and by using evidenced-based practice. In this first part of a two-part article, we describe key steps for the analysis of clinical data using quantitative methodology. We aim to describe the principles of medical statistics and their relevance and use in hand surgery, with contemporaneous examples. Hand surgeons seek expertise and guidance in the clinical domain to improve their practice and patient care. Part of this process involves the critical analysis and appraisal of the research of others.

Introduction Alcohol abuse is the leading global cause of premature mortality among 15–49 year olds and is associated with an increased risk of liver disease, cardiovascular disease, diabetes, anxiety and depression. Regular exercise reduces the risk of several associated comorbidities and ameliorates depression, anxiety and stress; known facilitators of relapse among those abusing alcohol. However, the current evidence base for exercise training among those with alcohol dependency is sparse. To address limitations of the small number of previous studies (no objective measurement alcohol consumption and a lack of comparison to non-exercise standard treatment), we investigated whether exercise alongside standard care provides additional benefits regarding alcohol consumption and mental health, versus standard care treatment alone, among people abusing alcohol. Methods The study was a two-arm, parallel-group randomised controlled pilot trial comparing the impact of a 12 week group exercise programme (comprising circuit, bodyweight and interval aerobic training), alongside alcohol treatment standard care (comprising group/individual counselling and pharmaceutical interventions), versus standard care alone on changes in alcohol consumption, alcohol cravings and mental health outcomes. Sixty eight treatment-seeking alcohol abuse patients, aged 18–65 years and with a baseline AUDIT score ≥ 8, were recruited (30 randomized to the exercise/standard care group; 38 to the standard care only group) with 57 participants completing up to the Week 13 follow-up. Results Participants in both groups self-reported a comparable reduction in alcohol consumption (7 day drinking diary/AUDIT score) and craving between baseline and the Week 13 follow-up; however objective alcohol consumption (PEth level) was only reduced among those in the exercise/standard care group (table 1). Despite this, there was no significant difference in the change in PEth level scores between the two groups (p>0.05; table 1). Similarly, despite a trend of more pronounced improvements in mental health scores (PHQ-9 depression; GAD anxiety) between baseline and the Week 13 follow-up among participants in the exercise/standard care group, there was no significant differences in the change in these endpoints between the two groups over time (p>0.05; table 1).View this table: • View inline • View popup • Download powerpoint Abstract P9 Table 1 Alcohol consumption and mental health endpoint data for participants in both the combined exercise programme and standard care group (Exercise + Std Care) and standard care only group (Std Care Only) who attended the baseline and Week 13 follow up visits [Mean (standard error)] Discussion A 12 week group exercise programme, alongside standard care, provided no statistically significant changes in alcohol consumption, cravings and depression/anxiety scores versus standard care alone among people attending treatment for alcohol abuse, despite the fact that only participants in the exercise/standard care group reduced their objective alcohol consumption.