324139 research outputs found
Sort by
Balancing act of academic clinical fellows in UK emergency medicine: a qualitative study
Get PDFBackground: Emergency medicine (EM) faces significant workforce challenges in sustaining clinical academic careers. Academic clinical fellowships (ACFs) offer protected research time, but little is known about how EM ACFs experience and navigate these posts. Methods: Semi-structured interviews were conducted with 20 current and former EM ACFs from 12 universities in England and Wales. Interviews were analysed using thematic analysis following Braun and Clarke’s six-phase approach. A mixed inductive and deductive framework was applied. Reflexivity and positionality were addressed through multi-researcher coding and consensus development. Results: Six themes were identified: (1) Elements of surprise—structural ambiguity and unexpected barriers; (2) Unclear direction—limited guidance and inconsistent supervision; (3) Loneliness—professional isolation and detachment from clinical peers; (4) Engagement—enthusiasm linked to research alignment and supervisory support; (5) Repeated generic hurdles—difficulty balancing academic and clinical demands; (6) EM-specific hurdles—reduced exposure to key rotations and limited academic mentorship within EM. Fellows reported uncertainty about extensions to training and programme variability. Conclusions: The EM ACF provides valuable entry into clinical academia; however, inconsistent structures, supervisory support and clarity in expectations hinder its full potential. Standardised induction, tailored supervision and flexible but transparent pathways are needed. These findings can inform policy, training programmes and institutional practices to better support the next generation of clinical academics in EM in the UK
The decline of child stunting in 122 countries: a systematic review of child growth studies since the 19th century
Get PDFIntroduction: Child stunting, a measure of malnutrition, is a major global health challenge affecting 148.1 million children in 2022. Global stunting rates have declined from 47.2% in 1985 to 22.3% in 2022; however, trends before the mid-1980s are unclear, including whether child stunting was previously prevalent in current high-income countries (HICs). We conducted a systematic review of child growth studies before 1990 to reconstruct historical rates of child stunting. Methods: We included reports of mean height by age and sex for children up to age 10.99 years. We excluded studies that were not representative of the targeted population and data for children under age 2. Stunting rates were computed by converting the means and SDs of height to height-for-age Z-scores (HAZ) using the WHO standard/reference, combining the HAZ distributions for all ages and measuring the share of the combined distribution below the stunting threshold. Results: We found 923 child growth studies at the community, regional and national level covering 122 countries from 1814 to 2016. We supplemented these historical studies with stunting estimates from the 1990s onward from the Joint Malnutrition Estimates database. Many current HICs had high levels of child stunting in the early 20th century, similar to low- and middle-income countries (LMICs) today. However, there was heterogeneity: stunting rates were low in Scandinavia, the European settler colonies and in the Caribbean, higher in Western Europe and exceptionally high in Japan and South Korea. Child stunting declined across the 20th century. Conclusion: The global child stunting rate was substantially higher in the early 20th century than in 1985, and the reduction of child stunting was a central feature of the health transition. The high stunting rates and subsequent reduction of stunting in HICs suggest that current HICs provide lessons for eradicating child stunting and that all LMICs can eliminate stunting
Prevalence and Correlates of Geriatric Depression: A Community‐Based Cross‐Sectional Study in Bangladesh
Get PDFObjectives: This study aimed to estimate the prevalence of geriatric depression and identify its associated socio‐demographic, lifestyle, health, and psychosocial determinants among community‐dwelling older adults in Bangladesh. Methods: A community‐based cross‐sectional study was conducted between April and December 2024 among 719 adults aged ≥ 60 years, selected by multistage random sampling across four administrative divisions. Data were collected using interviewer‐administered questionnaires. Depression was assessed using the GDS‐15 (cut‐off ≥ 5). Logistic regression models were applied to identify determinants, and adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were reported. Results: The prevalence of geriatric depression was 35.2%. Poor family bonding (AOR = 6.580, 95% CI = 3.540–12.210, p < 0.001), financial instability (AOR = 3.660, 95% CI = 2.550–5.260, p < 0.001), dissatisfaction with lifestyle (AOR = 3.590, 95% CI = 2.240–5.770, p < 0.001), and daily life stress (AOR = 3.640, 95% CI = 2.560–5.190, p < 0.001) were significant predictors. Comorbidities, lack of regular exercise, mobility, and hearing problems also increased depression risk. Conclusion: Geriatric depression is common in Bangladesh and influenced by multidimensional determinants. Integrating geriatric mental health services into primary care and strengthening family and social support systems are essential for prevention and early management
Outcomes of second‐line axicabtagene ciloleucel for large B‐cell lymphoma in the UK
Get PDFFollowing approval of axicabtagene ciloleucel (axi‐cel) as second‐line (2 L) treatment for large B‐cell lymphoma (LBCL), results from real‐world CAR T cohorts will be key to confirm safety and efficacy in standard practice. We present comprehensive clinical outcomes of LBCL patients intended to be treated with 2 L axi‐cel through the UK National CAR T service. Of 345 patients approved for 2 L axi‐cel, 302 (87.5%) were infused. The median age was 62 years (range 22–78); 21% were over 70 years. 75% of patients were approved for CAR T within 3 months from end of first‐line (1 L) therapy. 42% of patients required pre‐apheresis holding therapy, and 97% received bridging therapy. The best overall response rate was 86% (64% complete response). The 12‐month OS was 73.9% (95% CI: 68.3–78.7) for infused patients and 1.5 months (0.9–3.0) for patients not proceeding to CAR T. The 12‐month PFS was 52.4% (46.3–58.0). In multivariable analysis, advanced stage, male sex, no response to 1 L therapy, high LDH, and high CRP pre‐infusion were independently associated with PFS. Grade ≥3 CRS and ICANS rates were 5% and 18%, respectively. Outcomes in patients aged ≥70 years were similar to the younger population. In this large UK real‐world cohort of 2 L axi‐cel in LBCL, we demonstrate efficacy and toxicity outcomes comparable to the pivotal ZUMA‐7 trial, despite 42% patients requiring urgent holding therapy. Outcomes were favorable in patients aged ≥70 years, supporting the use of 2 L CAR T in older fit patients
Power transform revisited: numerically stable, and federated
No full textPower transforms are popular parametric methods for making data more Gaussian-like, and are widely used as preprocessing steps in statistical analysis and machine learning. However, we find that direct implementations of power transforms suffer from severe numerical instabilities, which can lead to incorrect results or even crashes. In this paper, we provide a comprehensive analysis of the sources of these instabilities and propose effective remedies. We further extend power transforms to the federated learning setting, addressing both numerical and distributional challenges that arise in this context. Experiments on real-world datasets demonstrate that our methods are both effective and robust, substantially improving stability compared to existing approaches
The unheard ‘scandal’: Covid‐19 vaccine‐injured people's perceptions of the national redress scheme and trust in the government in the United Kingdom
Get PDFThis article examines 218 Covid‐19 vaccine‐injured people's experiences of the process of seeking redress from the UK Government's Vaccine Damage Payment Scheme (VDPS). Moreover, it engages with the wider topic of how we can understand the government(s)’ role in dealing with an ‘emergency’ (the Covid‐19 pandemic) and the consequent (global) health crisis. Vaccine‐injured individuals and communities perceive suffering health problems after Covid‐19 vaccination as a ‘tragedy’, which requires an adequate response. However, the VDPS often does not meet their expectations, which fuels the vaccine‐injured community's narratives around an emerging ‘scandal’ over inadequate redress provision for those who suffered adverse effects following Covid‐19 vaccines. These perceptions and meaning‐making are in stark contrast to the fact that the wider population is unaffected by these issues and are generally not aware of the harms suffered or of the VDPS
Stability and perturbation analysis of non-negative super-resolution
Get PDFThe convolution of a discrete measure, x = Pki=1 aiδti , with a local window function, φ(s − t), is a common model for a measurement device whose resolution is substantially lower than that of the objects being observed. Super-resolution concerns localising the point sources {ai , ti} k i=1 with an accuracy beyond the essential support of φ(s−t), typically from m samples y(sj) = Pki=1 aiφ(sj − ti) + wj , where wj indicates an inexactness in the sample value. We consider the setting of x being non-negative and study two aspects of this problem: stability of the solutions with respect to measurement noise and perturbation of the solutions with respect to inaccuracies in the dual variable when solving the dual problem.
In Part I, we characterise non-negative solutions xˆ consistent with the samples within the bound ∥w∥2 ≤ δ. We show that the integrals of xˆ and x over (ti − E, ti + E) are close, converging to one another as E and δ approach zero. We then show how to make this general result, for windows that form a Chebyshev system, precise for the case of φ(s − t) being a Gaussian window, in which case the average error between xˆ and x is O(δ 1/6 ). The main innovation of this result is that non-negativity alone is sufficient to localise point sources beyond the sensor resolution and that, while regularisers such as total variation might be particularly effective, they are not required in the non-negative setting.
A practical approach for solving the problem is to consider its dual. In Part II, we study the stability of solutions with respect to the solutions to the dual problem. In particular, we establish a relationship between perturbations in the primal variable and perturbations the dual variable around the optimiser. We then establish a similar relationship between perturbations in the dual variable around the optimiser and the magnitude of the additive noise ∥w∥2 in the measurements
A decision-tree model to evaluate the impacts of workforce investments aimed at strengthening emergency obstetric and neonatal care (EmONC) facilities in Burundi
Get PDFIntroduction: Human resources for health (HRH) in low- and middle income countries (LMICs) are often allocated irrationally based on normative policies that ignore service delivery patterns and health outcomes, failing to account for costs and expected benefits. To help Burundi rationalise scarce resources and tackle health system inefficiencies, we developed four workforce investment scenarios aimed at strengthening emergency obstetric and neonatal care (EmONC) facilities and predicted the costs and benefits associated with these investment proposals over the 2025–2030 timeframe. Methods: We documented across the 112 Burundian EmONC facilities the available human resources and collated annual data on deliveries and obstetric complications covering 2021. Using the Workload Indicator for Staffing Needs (WISN) methodology, we estimated for each facility the workforce deficit and developed four EmONC workforce investment scenarios; no investment, a partial workforce package, and full workforce packages targeted at either all facilities or a select group of priority facilities. A decision tree model helped to predict the costs per maternal death averted. Results: A budget of US 3.27 million or allocate US 2380.58 for each maternal life saved. Conclusion: Efforts to invest across the EmONC network if simply based on assigned B/CEmONC status may poorly allocate scarce resources. If resources are limited, investing in two additional midwives across the 104 EmONC facilities would be more efficient than implementing a full workforce package. However, health benefits of any workforce investment significantly depend on an enabling environment