We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
To save content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about saving content to .
To save content items to your Kindle, first ensure no-reply@cambridge.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about saving to your Kindle.
Note you can select to save to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
The development of Generative Artificial Intelligence (GenAI) has led to intense wonder, surprise, excitement, and concern within the language teaching profession. These tools offer the potential to assist language teachers in helping their learners achieve their language learning goals, and at the same time, risk disrupting language teaching and learning processes, the teaching profession, and possibly the instrumental needs to learn foreign languages. This Element provides an accessible introduction and guide to the use of GenAI for language teaching. It aims to facilitate language teachers' development of the professional knowledge and skills they need to use GenAI responsibly, ethically and effectively. The Element It is a valuable resource for pre-service and in-service language teachers of all experience levels. Each section includes helpful tips and questions for reflection to get teachers started with GenAI while ensuring they engage critically and responsibly with these tools. Evidence-informed approaches are promoted throughout the Element.
It remains unclear which individuals with subthreshold depression benefit most from psychological intervention, and what long-term effects this has on symptom deterioration, response and remission.
Aims
To synthesise psychological intervention benefits in adults with subthreshold depression up to 2 years, and explore participant-level effect-modifiers.
Method
Randomised trials comparing psychological intervention with inactive control were identified via systematic search. Authors were contacted to obtain individual participant data (IPD), analysed using Bayesian one-stage meta-analysis. Treatment–covariate interactions were added to examine moderators. Hierarchical-additive models were used to explore treatment benefits conditional on baseline Patient Health Questionnaire 9 (PHQ-9) values.
Results
IPD of 10 671 individuals (50 studies) could be included. We found significant effects on depressive symptom severity up to 12 months (standardised mean-difference [s.m.d.] = −0.48 to −0.27). Effects could not be ascertained up to 24 months (s.m.d. = −0.18). Similar findings emerged for 50% symptom reduction (relative risk = 1.27–2.79), reliable improvement (relative risk = 1.38–3.17), deterioration (relative risk = 0.67–0.54) and close-to-symptom-free status (relative risk = 1.41–2.80). Among participant-level moderators, only initial depression and anxiety severity were highly credible (P > 0.99). Predicted treatment benefits decreased with lower symptom severity but remained minimally important even for very mild symptoms (s.m.d. = −0.33 for PHQ-9 = 5).
Conclusions
Psychological intervention reduces the symptom burden in individuals with subthreshold depression up to 1 year, and protects against symptom deterioration. Benefits up to 2 years are less certain. We find strong support for intervention in subthreshold depression, particularly with PHQ-9 scores ≥ 10. For very mild symptoms, scalable treatments could be an attractive option.
Objectives/Goals: Ventilator-associated pneumonia (VAP) is an infection caused by bacteria, viruses, or fungi during mechanical ventilation. We analyzed a cohort of COVID-19 patients admitted to the intensive care unit with respiratory failure with different VAP outcomes. We hypothesize that the multiomics data can help predict VAP development within this cohort. Methods/Study Population: We recruited participants from a cohort on a NYU IRB protocol (i22–00616), who had COVID19 respiratory failure, admitted to ICU, and required invasive mechanical ventilation (n = 245). We collected and analyzed research specimens (bronchoalveolar lavage [BAL, n = 213], tracheal aspirates [n = 246], background [n = 18]) and clinical cultures (sputum and BAL) for 245 participants. A panel of experts adjudicated VAP within the cohort, resulting in 92 VAP diagnoses. We annotated metatranscriptome (Illumina NovaSeq) using a Kraken/Bracken database, and KEGG for functional annotation of transcriptome data (Illumina HiSeq). We used edgeR (v.4.0.16) to analyze differential expression of metatranscriptome and transcriptome data. Results/Anticipated Results: We diagnosed VAP in n = 92 (38%) participants. We found significant differences in days of overall hospital stay (p Discussion/Significance of Impact: VAP is a serious complication of mechanical ventilation, and oral commensals alter the lung microbiome and host immunity. We identified a transcriptome-metatranscriptome signature that identifies those at VAP risk. VAP was associate with both pro- and anti-inflammatory gene expression resulting in increased risk for lower airway infection.
Adults rate the speech of children assigned male at birth (AMAB) and assigned female at birth (AFAB) as young as 2.5 years of age differently on a scale of definitely a boy to definitely a girl (Munson et al., 2022), despite the lack of consistent sex dimorphism in children’s speech production mechanisms. This study used longitudinal data to examine the acoustic differences between AMAB and AFAB children and the association between the acoustic measures and perceived gender ratings of children’s speech. We found differences between AMAB and AFAB children in two acoustic parameters that mark gender in adult speech: the spectral centroid of /s/ and the overall scaling of resonant frequencies in vowels. These results demonstrate that children as young as 3 years old speak in ways that reflect their sex assigned at birth. We interpret this as evidence that children manipulate their speech apparatus volitionally to mark gender through speech.
In responding to a Chemical, Biological, Radiological, and Nuclear explosive (CBRNe) disaster, clinical leaders have important decision-making responsibilities which include implementing hospital disaster protocols or incident command systems, managing staffing, and allocating resources. Despite emergency care clinical leaders’ integral role, there is minimal literature regarding the strategies they may use during CBRNe disasters. The aim of this study was to explore emergency care clinical leaders’ strategies related to managing patients following a CBRNe disaster.
Methods
Focus groups across 5 tertiary hospitals and 1 rural hospital in Queensland, Australia. Thirty-six hospital clinical leaders from the 6 study sites crucial to hospital disaster response participated in 6 focus groups undertaken between February and May 2021 that explored strategies and decision making to optimize patient care following a CBRNe disaster.
Results
Analysis revealed the use of rehearsals, adopting new models of care, enacting current surge management processes, and applying organization lessons were facilitating strategies. Barriers to management were identified, including resource constraints and sites operating over capacity.
Conclusions
Enhanced education and training of clinical leaders, flexible models of care, and existing established processes and tested frameworks could strengthen a hospital’s response when managing patients following a CBRNe disaster.
The presence of an intraluminal thrombus in acutely symptomatic carotid stenosis is thought to represent a high-risk lesion for short-term stroke reccurrence though evidence on natural history and treatment is lacking, leading to equipoise and much variation in practice. The objective of this study was to map these variations in practice (medical management and timing of revascularization), determine the considerations that influence clinician decision-making in this condition and gather opinions that inform the development and design of future trials in the area.
Methods:
This was a mixed-methods study using both quantitative survey methods and qualitative interview-based methods. International perspectives were gathered by distributing a case-based survey via the “Practice Current” section of Neurology: Clinical Practice and interviewing international experts using established qualitative research methods.
Results:
The presence of an intraluminal thrombus significantly increased the likelihood of using a regimen containing anticoagulation agents (p < 0.001) in acutely symptomatic carotid stenosis in the case-based survey. Themes that emerged from qualitative interview analysis were therapeutic uncertainty regarding anticoagulation, decision to reimage, revascularization choices and future trial design and anticipated challenges.
Conclusion:
Results of this study demonstrate a preference for anticoagulation and delayed revascularization after reimaging to examine for clot resolution, though much equipoise remains. While there is interest from international experts in future trials, further study is needed to understand the natural history of this condition in order to inform trial design.
High-cardinality categorical features are pervasive in actuarial data (e.g., occupation in commercial property insurance). Standard categorical encoding methods like one-hot encoding are inadequate in these settings.
In this work, we present a novel Generalised Linear Mixed Model Neural Network (“GLMMNet”) approach to the modelling of high-cardinality categorical features. The GLMMNet integrates a generalised linear mixed model in a deep learning framework, offering the predictive power of neural networks and the transparency of random effects estimates, the latter of which cannot be obtained from the entity embedding models. Further, its flexibility to deal with any distribution in the exponential dispersion (ED) family makes it widely applicable to many actuarial contexts and beyond. In order to facilitate the application of GLMMNet to large datasets, we use variational inference to estimate its parameters—both traditional mean field and versions utilising textual information underlying the high-cardinality categorical features.
We illustrate and compare the GLMMNet against existing approaches in a range of simulation experiments as well as in a real-life insurance case study. A notable feature for both our simulation experiment and the real-life case study is a comparatively low signal-to-noise ratio, which is a feature common in actuarial applications. We find that the GLMMNet often outperforms or at least performs comparably with an entity-embedded neural network in these settings, while providing the additional benefit of transparency, which is particularly valuable in practical applications.
Importantly, while our model was motivated by actuarial applications, it can have wider applicability. The GLMMNet would suit any applications that involve high-cardinality categorical variables and where the response cannot be sufficiently modelled by a Gaussian distribution, especially where the inherent noisiness of the data is relatively high.
Traditional techniques for calculating outstanding claim liabilities such as the chain-ladder are notoriously at risk of being distorted by outliers in past claims data. Unfortunately, the literature in robust methods of reserving is scant, with notable exceptions such as Verdonck & Debruyne (2011, Insurance: Mathematics and Economics, 48, 85–98) and Verdonck & Van Wouwe (2011, Insurance: Mathematics and Economics,49, 188–193). In this paper, we put forward two alternative robust bivariate chain-ladder techniques to extend the approach of Verdonck & Van Wouwe (2011, Insurance: Mathematics and Economics,49, 188–193). The first technique is based on Adjusted Outlyingness (Hubert & Van der Veeken, 2008. Journal of Chemometrics,22, 235–246) and explicitly incorporates skewness into the analysis while providing a unique measure of outlyingness for each observation. The second technique is based on bagdistance (Hubert et al., 2016. Statistics: Methodology, 1–23) which is derived from the bagplot; however; it is able to provide a unique measure of outlyingness and a means to adjust outlying observations based on this measure.
Furthermore, we extend our robust bivariate chain-ladder approach to an N-dimensional framework. The implementation of the methods, especially beyond bivariate, is not trivial. This is illustrated on a trivariate data set from Australian general insurers and results under the different outlier detection and treatment mechanisms are compared.
Drawing upon recent studies that empirically estimate both the benefits and costs of trade, this paper addresses a simple and important question: By how much do the benefits of increased global trade outweigh the costs? To the best of our knowledge, this is the first attempt to answer this question at global and World Bank income-grouping levels using empirically estimated relationships from the trade cost literature. Using a structural gravity model, we simulate changes in three primary trade constraints: a 10% reduction in tariff levels, a 10% reduction in effective distance, and a 10% increase in free trade agreement depth. The projection leads to a roughly 5% increase in global trade by value. Our model suggests that increased trade has an incredibly high benefit–cost ratio (BCR) for the developing world with an order-of-magnitude estimate for low- and lower–middle-income countries of 100 and for upper–middle-income countries of 50. However, the BCR for high-income countries is substantially lower, with a value closer to 5. Overall, the results suggest that free trade leads to substantial net benefits globally, generating US$ 700 billion in benefits (0.83% of global GDP) and US$ 100 billion in costs (0.12% of global GDP) in the first year, a differential that grows over time. Sensitivity analyses suggest that our BCRs are on the lower end of a plausible range. The results point to the incredible value of free trade, particularly for developing countries, and reiterate the importance of considering distributional impacts when implementing trade reforms.
With persistent incidence, incomplete vaccination rates, confounding respiratory illnesses, and few therapeutic interventions available, COVID-19 continues to be a burden on the pediatric population. During a surge, it is difficult for hospitals to direct limited healthcare resources effectively. While the overwhelming majority of pediatric infections are mild, there have been life-threatening exceptions that illuminated the need to proactively identify pediatric patients at risk of severe COVID-19 and other respiratory infectious diseases. However, a nationwide capability for developing validated computational tools to identify pediatric patients at risk using real-world data does not exist.
Methods:
HHS ASPR BARDA sought, through the power of competition in a challenge, to create computational models to address two clinically important questions using the National COVID Cohort Collaborative: (1) Of pediatric patients who test positive for COVID-19 in an outpatient setting, who are at risk for hospitalization? (2) Of pediatric patients who test positive for COVID-19 and are hospitalized, who are at risk for needing mechanical ventilation or cardiovascular interventions?
Results:
This challenge was the first, multi-agency, coordinated computational challenge carried out by the federal government as a response to a public health emergency. Fifty-five computational models were evaluated across both tasks and two winners and three honorable mentions were selected.
Conclusion:
This challenge serves as a framework for how the government, research communities, and large data repositories can be brought together to source solutions when resources are strapped during a pandemic.
Whereas the beneficial effect of antiplatelet therapy for recurrent stroke prevention has been well established, uncertainties remain regarding the optimal antithrombotic regimen for recently symptomatic carotid stenosis. We sought to explore the approaches of stroke physicians to antithrombotic management of patients with symptomatic carotid stenosis.
Methods:
We employed a qualitative descriptive methodology to explore the decision-making approaches and opinions of physicians regarding antithrombotic regimens for symptomatic carotid stenosis. We conducted semi-structured interviews with a purposive sample of 22 stroke physicians (11 neurologists, 3 geriatricians, 5 interventional-neuroradiologists, and 3 neurosurgeons) from 16 centers on four continents to discuss symptomatic carotid stenosis management. We then conducted thematic analysis on the transcripts.
Results:
Important themes revealed from our analysis included limitations of existing clinical trial evidence, competing surgeon versus neurologist/internist preferences, and the choice of antiplatelet therapy while awaiting revascularization. There was a greater concern for adverse events while using multiple antiplatelet agents (e.g., dual-antiplatelet therapy (DAPT)) in patients undergoing carotid endarterectomy compared to carotid artery stenting. Regional variations included more frequent use of single antiplatelet agents among European participants. Areas of uncertainty included antithrombotic management if already on an antiplatelet agent, implications of nonstenotic features of carotid disease, the role of newer antiplatelet agents or anticoagulants, platelet aggregation testing, and timing of DAPT.
Conclusion:
Our qualitative findings can help physicians critically examine the rationale underlying their own antithrombotic approaches to symptomatic carotid stenosis. Future clinical trials may wish to accommodate identified variations in practice patterns and areas of uncertainty to better inform clinical practice.
Influenza virus infections can lead to a number of secondary complications, including sepsis. We applied linear regression models to mortality and hospital admission data coded for septicaemia from 1998 to 2019 in Hong Kong, and estimated that septicaemia was associated with an annual average excess mortality rate of 0.23 (95% CI 0.04–0.40) per 100 000 persons per year and an excess septicaemia hospitalisation rate of 1.73 (95% CI 0.94–2.50) per 100 000 persons per year. The highest excess morbidity and mortality was found in older adults and young children, and during influenza A(H3N2) epidemics.
Lockdown during the pandemic has had significant impacts on public mental health. Previous studies suggest an increase in self-harm and suicide in children and adolescents. There has been little research on the roles of stringent lockdown.
Aims
To investigate the mediating and predictive roles of lockdown policy stringency measures in self-harm and emergency psychiatric presentations.
Method
This was a retrospective cohort study. We analysed data of 2073 psychiatric emergency presentations of children and adolescents from 23 hospital catchment areas in ten countries, in March to April 2019 and 2020.
Results
Lockdown measure stringency mediated the reduction in psychiatric emergency presentations (incidence rate ratio of the natural indirect effect [IRRNIE] = 0.41, 95% CI [0.35, 0.48]) and self-harm presentations (IRRNIE = 0.49, 95% CI [0.39, 0.60]) in 2020 compared with 2019. Self-harm presentations among male and looked after children were likely to increase in parallel with lockdown stringency. Self-harm presentations precipitated by social isolation increased with stringency, whereas school pressure and rows with a friend became less likely precipitants. Children from more deprived neighbourhoods were less likely to present to emergency departments when lockdown became more stringent,
Conclusions
Lockdown may produce differential effects among children and adolescents who self-harm. Development in community or remote mental health services is crucial to offset potential barriers to access to emergency psychiatric care, especially for the most deprived youths. Governments should aim to reduce unnecessary fear of help-seeking and keep lockdown as short as possible. Underlying mediation mechanisms of stringent measures and potential psychosocial inequalities warrant further research.
To determine the usefulness of adjusting antibiotic use (AU) by prevalence of bacterial isolates as an alternative method for risk adjustment beyond hospital characteristics.
AU in days of therapy per 1,000 patient days and microbiologic data from 2015 and 2016 were collected from 26 hospitals. The prevalences of Pseudomonas aeruginosa, extended-spectrum β-lactamase (ESBL)–producing bacteria, methicillin-resistant Staphylococcus aureus (MRSA), and vancomycin-resistant enterococci (VRE) were calculated and compared to the average prevalence of all hospitals in the network. This proportion was used to calculate the adjusted AU (a-AU) for various categories of antimicrobials. For example, a-AU of antipseudomonal β-lactams (APBL) was the AU of APBL divided by (prevalence of P. aeruginosa at that hospital divided by the average prevalence of P. aeruginosa). Hospitals were categorized by bed size and ranked by AU and a-AU, and the rankings were compared.
Results:
Most hospitals in 2015 and 2016, respectively, moved ≥2 positions in the ranking using a-AU of APBL (15 of 24, 63%; 22 of 26, 85%), carbapenems (14 of 23, 61%; 22 of 25; 88%), anti-MRSA agents (13 of 23, 57%; 18 of 26, 69%), and anti-VRE agents (18 of 24, 75%; 15 of 26, 58%). Use of a-AU resulted in a shift in quartile of hospital ranking for 50% of APBL agents, 57% of carbapenems, 35% of anti-MRSA agents, and 75% of anti-VRE agents in 2015 and 50% of APBL agents, 28% of carbapenems, 50% of anti-MRSA agents, and 58% of anti-VRE agents in 2016.
Conclusions:
The a-AU considerably changes how hospitals compare among each other within a network. Adjusting AU by microbiological burden allows for a more balanced comparison among hospitals with variable baseline rates of resistant bacteria.
Property-based random testing á la QuickCheck requires building efficient generators for well-distributed random data satisfying complex logical predicates, but writing these generators can be difficult and error prone. This chapter introduces a probabilistic domain-specific language in which generators are conveniently expressed by decorating predicates with lightweight annotations to control both the distribution of generated values and the amount of constraint solving that happens before each variable is instantiated. This language, called Luck, makes generators easier to write, read and maintain. We give Luck a probabilistic formal semantics and prove several fundamental properties, including the soundness and completeness of random generation with respect to a standard predicate semantics. We evaluate Luck on common examples from the property-based testing literature and on two significant case studies, showing that it can be used in complex domains with comparable bug-finding effectiveness and a significant reduction in testing code size compared to handwritten generators.
The risk of environmental contamination by severe acute respiratory coronavirus virus 2 (SARS-CoV-2) in the intensive care unit (ICU) is unclear. We evaluated the extent of environmental contamination in the ICU and correlated this with patient and disease factors, including the impact of different ventilatory modalities.
Methods:
In this observational study, surface environmental samples collected from ICU patient rooms and common areas were tested for SARS-CoV-2 by polymerase chain reaction (PCR). Select samples from the common area were tested by cell culture. Clinical data were collected and correlated to the presence of environmental contamination. Results were compared to historical data from a previous study in general wards.
Results:
In total, 200 samples from 20 patient rooms and 75 samples from common areas and the staff pantry were tested. The results showed that 14 rooms had at least 1 site contaminated, with an overall contamination rate of 14% (28 of 200 samples). Environmental contamination was not associated with day of illness, ventilatory mode, aerosol-generating procedures, or viral load. The frequency of environmental contamination was lower in the ICU than in general ward rooms. Eight samples from the common area were positive, though all were negative on cell culture.
Conclusion:
Environmental contamination in the ICU was lower than in the general wards. The use of mechanical ventilation or high-flow nasal oxygen was not associated with greater surface contamination, supporting their use and safety from an infection control perspective. Transmission risk via environmental surfaces in the ICUs is likely to be low. Nonetheless, infection control practices should be strictly reinforced, and transmission risk via droplet or airborne spread remains.
The coronavirus disease 2019 (COVID-19) pandemic has led to significant strain on front-line healthcare workers.
Aims
In this multicentre study, we compared the psychological outcomes during the COVID-19 pandemic in various countries in the Asia-Pacific region and identified factors associated with adverse psychological outcomes.
Method
From 29 April to 4 June 2020, the study recruited healthcare workers from major healthcare institutions in five countries in the Asia-Pacific region. A self-administrated survey that collected information on prior medical conditions, presence of symptoms, and scores on the Depression Anxiety Stress Scales and the Impact of Events Scale-Revised were used. The prevalence of depression, anxiety, stress and post-traumatic stress disorder (PTSD) relating to COVID-19 was compared, and multivariable logistic regression identified independent factors associated with adverse psychological outcomes within each country.
Results
A total of 1146 participants from India, Indonesia, Singapore, Malaysia and Vietnam were studied. Despite having the lowest volume of cases, Vietnam displayed the highest prevalence of PTSD. In contrast, Singapore reported the highest case volume, but had a lower prevalence of depression and anxiety. In the multivariable analysis, we found that non-medically trained personnel, the presence of physical symptoms and presence of prior medical conditions were independent predictors across the participating countries.
Conclusions
This study highlights that the varied prevalence of psychological adversity among healthcare workers is independent of the burden of COVID-19 cases within each country. Early psychological interventions may be beneficial for the vulnerable groups of healthcare workers with presence of physical symptoms, prior medical conditions and those who are not medically trained.
Rapid detection and isolation of coronavirus disease 2019 (COVID-19) patients is the only means of reducing hospital transmission. We describe the impact of implementation of on-site severe acute respiratory coronavirus virus 2 (SARS-CoV-2) reverse-transcription polymerase chain reaction (RT-PCR) testing on reducing turnaround time, isolation duration, pathology test ordering, and antibiotic use in patients who do not have COVID-19.
Introducing common shocks is a popular dependence modelling approach, with some recent applications in loss reserving. The main advantage of this approach is the ability to capture structural dependence coming from known relationships. In addition, it helps with the parsimonious construction of correlation matrices of large dimensions. However, complications arise in the presence of “unbalanced data”, that is, when (expected) magnitude of observations over a single triangle, or between triangles, can vary substantially. Specifically, if a single common shock is applied to all of these cells, it can contribute insignificantly to the larger values and/or swamp the smaller ones, unless careful adjustments are made. This problem is further complicated in applications involving negative claim amounts. In this paper, we address this problem in the loss reserving context using a common shock Tweedie approach for unbalanced data. We show that the solution not only provides a much better balance of the common shock proportions relative to the unbalanced data, but it is also parsimonious. Finally, the common shock Tweedie model also provides distributional tractability.
Previous research has suggested an association between depression and subsequent acute stroke incidence, but few studies have examined any effect modification by sociodemographic factors. In addition, no studies have investigated this association among primary care recipients with hypertension.
Methods
We examined the anonymized records of all public general outpatient visits by patients aged 45+ during January 2007–December 2010 in Hong Kong to extract primary care patients with hypertension for analysis. We took the last consultation date as the baseline and followed them up for 4 years (until 2011–2014) to observe any subsequent acute hospitalization due to stroke. Mixed-effects Cox models (random intercept across 74 included clinics) were implemented to examine the association between depression (ICPC diagnosis or anti-depressant prescription) at baseline and the hazard of acute stroke (ICD-9: 430–437.9). Effect modification by age, sex, and recipient status of social security assistance was examined in extended models with respective interaction terms specified.
Results
In total, 396 858 eligible patients were included, with 9099 (2.3%) having depression, and 10 851 (2.7%) eventually hospitalized for stroke. From the adjusted analysis, baseline depression was associated with a 17% increased hazard of acute stroke hospitalization [95% confidence interval (CI) 1.03–1.32]. This association was suggested to be even stronger among men than among women (hazard ratio = 1.29, 95% CI 1.00–1.67).
Conclusion
Depression is more strongly associated with acute stroke incidence among male than female primary care patients with hypertension. More integrated services are warranted to address their needs.