To Örebro University

Alert fatigue remains a major barrier to the effective deployment of predictive models in emergency care, particularly in the context of rare but critical outcomes such as in-hospital mortality (IHM), which often occurs in less than 5.0% of patients admitted from the emergency department (ED). Severe class imbalance leads to low positive predictive value (PPV), undermining the clinical utility of even high-performance predictive models. To address this issue, we propose AI-TEW (Artificial Intelligence-powered Tiered Early Warning), a novel two-stage early warning framework designed to reduce false alarms and improve clinical interpretability. In Stage 1, a robust machine learning model was developed and validated using data from 174,292 ED visits across three hospitals in China and the United States. The model demonstrated strong discriminative ability for IHM prediction, achieving AUROCs ranging from 0.84 (95% CI, 0.81-0.86) to 0.91 (95% CI, 0.90-0.91) in internal and external validation cohorts. In Stage 2, AI-TEW implements a tiered risk stratification strategy by optimizing decision thresholds to prioritize high-risk patients, thereby increasing PPV from baseline levels of 9.8-18.8% to 32.5-40.5% across sites, while maintaining a high negative predictive value (NPV) of over 98% for low-risk individuals. To further refine alert precision, a knowledge-based filtering layer is introduced, leveraging large language models (LLM) to interpret patient-specific risk factors derived from SHAP (Shapley Additive exPlanations) method. Integrating explainable AI with clinical reasoning enhances contextual understanding and reduces spurious alerts, leading to an 11.53% increase in PPV in external validation (p = 0.0092 for MedGemma). By integrating improved predictive efficiency with interpretable, knowledge-informed filtering, AI-TEW reduces alert burden while supporting timely clinical intervention, demonstrating a promising approach to mitigating the impact of class imbalance in emergency risk prediction.

BACKGROUND: This study aimed to describe the characteristics of frequent users of ambulance services and their corresponding ambulance missions.

METHODS: A retrospective cohort study using register data from a Swedish region (approximately 308,000 inhabitants) was conducted. The dataset covered ambulance missions dispatched between 2019 and 2023; data from 2020-2022 were analyzed, with the extended timeframe (2019-2023) used to identify frequent users. Frequent users were defined as patients with four or more ambulance missions within a 365-day period, divided into moderate (patients with 4-11 missions within 365 days) and high-frequency users (≥ 12 missions within 365 days). Missions associated with these users were referred to as frequent, moderate-frequent, high-frequency, or non-frequent ambulance missions.

RESULTS: Between 2020 and 2022, 73,461 ambulance missions were utilized by 41,407 unique patients. Of these 41,407 patients, 8.4% were frequent users (7.9% moderate and 0.5% high-frequency users), but accounted for 29.0% of all ambulance missions. The remaining 91.6% were non-frequent users, accounting for 71.0% of the ambulance missions. Frequent users were older than non-frequent users (median age 75 vs 62, p < .001). Frequent ambulance missions were more often dispatched for dyspnoea (17.8% vs 11.6%, p < .001), convulsions (4.2% vs 2.1%, p < .001), and abdominal pain (10.5% vs 8.7%, p < .001) compared with non-frequent ambulance missions. Frequent ambulance missions were also more likely to be assigned to a higher triage level, and 77.7% resulted in transport to the emergency department. Compare to moderate-frequent ambulance missions, high-frequency missions were more often dispatched for psychiatric emergencies (6.2% vs 2.1%, p < .001) and intoxications (5.3% vs 2.0%, p < .001), and more frequently resulted in transport to psychiatric emergency care (3.6% vs 1.4%, p < .001), or not being conveyed (24.0% vs 16.1%, p < .001).

CONCLUSIONS: Frequent ambulance users account for a large proportion of all ambulance missions. The results indicate that frequent users are often in need of emergency care, and understanding this population's needs is essential to ensure appropriate care. However, the group is heterogeneous and can be divided into frequent and high-frequency users, which have different characteristics. We suggest that future research investigates system-level approaches to identify frequent ambulance users and implement care plans to address patients' needs and reduce ambulance utilisation.

The host immune response in sepsis involves both pro- and anti-inflammatory mechanisms, with monocytes playing a central role in the process. We have previously identified an in vitro response profile of endotoxin (LPS) tolerant primary human monocytes, consisting of eight cytokines/chemokines as well as a set of five transcription factors. In the current study, we evaluated differences in expression levels of these investigated molecular markers across different patient groups (patients with or without infection, and with or without sepsis), and their association with clinical outcomes (septic shock and in-hospital mortality), among 809 ambulance patients. The results showed that patients with sepsis displayed the lowest HLA-DRA expression levels together with the lowest TNF/IL-10 ratio, while most other cytokine/chemokines and gene expressions were elevated. Higher levels of HGF, CCL8, CCL2, TNF and IL-10, as well as upregulation of HIF1A and NFKBIA were seen in septic patients with septic shock. The data suggests that the investigated immunological markers linked to immunosuppressed monocyte responses are associated with patients with sepsis and septic shock.

BACKGROUND: Inability to walk (IATW) has predicted mortality in emergency units in low-income settings, but its value among inpatients in high-income countries is unknown. AIM: To evaluate the predictive accuracy of IATW for 30-day in-hospital mortality among inpatients in a high-income country, compare its performance with NEWS CRITICAL (the threshold for urgent review), and assess whether IATW provides complementary risk stratification when combined with NEWS.

METHODS: All adult inpatients in four Swedish hospitals were prospectively examined in a multicenter point-prevalence assessment. IATW was defined as inability to walk five steps without physical assistance or walking aids. NEWS was calculated from contemporaneous vital signs. The primary outcome was 30-day in-hospital mortality. We computed sensitivity, specificity, predictive values, and likelihood ratios. Multivariable logistic regression assessed the association between IATW and mortality, adjusted for NEWS, age, and sex.

RESULTS: Of 1842 patients, 59.9% were IATW-positive and 22.3% met NEWS CRITICAL (≥5 or any single-parameter score of 3). Mortality was 4.6%. IATW had higher sensitivity (86.9% vs 66.7%) but lower specificity (41.4% vs 79.8%) than NEWS CRITICAL; NPV was high for both (98.5% vs 98.0%). IATW was independently associated with mortality after adjustment for NEWS (OR 2.43, 95% CI 1.24-4.76; p = 0.009). Those both IATW-negative and with NEWS <5 (35.2% of the cohort) had 1.1% mortality.

CONCLUSIONS: Inability to walk (IATW) is a bedside assessment that when negative identifies patients with low 30-day in-hospital mortality and provides complementary information to NEWS. Its low complexity, high sensitivity and low negative likelihood ratios support its use in clinical risk assessments.

Introduction: The ageing of the population is expected to increase the costs and consumption of health care. Care should be patient-centred. Patient-centred care (PCC) has been shown to increase the quality of life for older patients living with frailty, as well as reduce health care costs. As one of the core principles of PCC, it is important to investigate and consider patients' wishes regarding their care.

Aim: The aim of this study was to investigate and identify factors and information that older patients find important when planning for their care.

Methods: Sixteen patients at the geriatric ward at & Ouml;rebro University Hospital in Sweden were interviewed using focus group discussions. The patients were >= 65 years of age, spoke Swedish, and were cognitively and physically able to participate in a focus group discussion. The interviews were recorded, transcribed, and analysed using content analysis.

Results: Three main categories were formed to summarize what mattered most to the patients when planning their care were that they maintained autonomy, that no harm was done and that the care had a holistic approach. These were further divided into the subcategories: the right to be well informed, taking part in planning of their care, having accessible care and a feeling of safety and security, considering emotional wellbeing, including relatives and close friends and maintaining their wellbeing.

Conclusions: What mattered most to the patients was closely related to the World Health Organization's recommendations for ethical and good quality health care. Generally, the patients agreed on what mattered most. Further studies are needed to enrich the understanding of what is important to older patients.

INTRODUCTION: As people age, frailty becomes increasingly prevalent and is associated with adverse outcomes, including higher risks of hospitalisation, institutionalisation and mortality. Early identification of frailty in ambulance and emergency department (ED) settings may support clinical decision-making and help predict outcomes for older patients. However, there is currently no consensus on how frailty should be assessed in these settings, and the clinical usefulness of available tools remains uncertain. This protocol outlines the methods for a planned scoping review that aims to identify frailty screening tools used for older adults (age ≥65 years) in the ambulance and ED settings and to evaluate existing evidence on their clinical usefulness.

METHODS AND ANALYSIS: This protocol follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for Protocols, and the final review will follow the Joanna Briggs Institute methodology and is reported in accordance with the PRISMA extension for Scoping Reviews guidelines. Studies will be included if they examine the use of frailty screening or assessment tools in individuals aged 65 years or older in ambulance and/or ED settings and report on aspects of clinical usefulness, such as feasibility, predictive validity or influence on clinical decisions. A comprehensive literature search will be conducted in PubMed, Embase, CINAHL, Scopus and Web of Science for studies published in English between January 2014 and December 2025. To refine the search strategy, an initial systematic pre-search was performed in PubMed using Medical Subject Headings terms, followed by a pilot study. A sample pilot screening of 101 references identified in the pre-search was conducted as a support for finalising the search. Ten of the papers in the pre-screening were furthermore used as a support for testing and validating the data extraction variables and quality assessment procedures. In the full scoping review, study selection and data extraction will be independently conducted by two reviewers using the Covidence software (Veritas Health Innovation, Melbourne, Australia), with any discrepancies resolved by a third reviewer. Extracted data will be summarised in tabular format and analysed through narrative synthesis. The methodological quality of included studies will be evaluated using the quality assessment tool of the National Heart, Lung, and Blood Institute for cohort studies.

ETHICS AND DISSEMINATION: Ethical approval is not required as no primary data will be collected. Findings will be disseminated through publication in a peer-reviewed journal, conference presentations and summaries shared with relevant clinical and research-related stakeholders.

BACKGROUND: Emergency departments (EDs) are often the first point of contact for patients with sepsis, and therefore play a critical role in early recognition and treatment. However, the extent to which sepsis guidelines are implemented across EDs in Europe remains unclear, and variability in adherence may impact patient outcomes. OBJECTIVE: This study aimed to assess current sepsis management practices in European EDs, evaluate adherence to international guidelines, and identify key challenges limiting effective implementation.

METHODS AND DESIGN: A structured survey was developed by sepsis experts from the European Society for Emergency Medicine (EUSEM). The questionnaire included both quantitative and open-ended items and underwent iterative refinement through pilot testing to ensure clarity and relevance.

SETTINGS AND PARTICIPANTS: The survey was distributed to EUSEM members and national emergency medicine societies across Europe. Respondents included medical directors or designated sepsis specialists, with only one response requested per ED. RESULTS: A total of 402 EDs from 28 European countries participated in the survey. While most EDs (72.5%) reported having a sepsis protocol in place, less than half implemented regular monitoring or structured training measures. The 1-h sepsis bundle was described as moderately to highly challenging to implement by the majority of the respondents; just over half (55%) reported completing all its elements within 1 h of ED presentation. Key barriers included high patient volumes, insufficient staffing, and lack of standardized sepsis definitions, leading to delays in recognition and treatment. The results also highlighted concerns regarding increased broad-spectrum antibiotic use following the implementation of the 1-h bundle, emphasizing the need for a balanced approach that considers antimicrobial stewardship.

CONCLUSION: This survey establishes a benchmark for understanding sepsis management practices in European EDs, identifying substantial variations and challenges. Areas for improvement include enhanced training to follow protocols, improved monitoring systems that measure protocol adherence, and alignment with evidence-based guidelines.

BACKGROUND: Disaster medicine involves managing situations where medical needs exceed available resources. In Sweden, disaster medicine is not yet a mandatory component of the medical education. Since the introduction of a revised six-year medical curriculum in 2021, it is unclear how disaster medicine will be integrated into the new program. This study aimed to evaluate the status of undergraduate disaster medicine education in Swedish medical schools, the teaching methodologies employed, plans for future curriculum integration, and the extent of variation across universities.

METHODS: We conducted a comprehensive review of syllabi from all Swedish medical programs to identify the inclusion of disaster medicine. Additionally, semi-structured interviews were conducted with 13 representatives from all seven Swedish medical schools, including those responsible for disaster medicine education or members of the education boards. The interviews explored teaching methods, curriculum content, and plans for the new six-year program. Data were analysed using qualitative content analysis.

RESULTS: Disaster medicine is included in the curriculum of all Swedish medical programs; however, its content, extent, and teaching approaches vary. Lectures are the most common teaching method, with some schools incorporating case discussions, tabletop exercises, and disaster simulations. Most medical faculties plan to maintain or expand their disaster medicine curriculum. However, there is no formal collaboration between universities in developing or standardizing disaster medicine education for the new curriculum.

CONCLUSION: The current level of disaster medicine education in Swedish medical schools requires enhancement in both quality and scope. Variations between universities would need to be minimized to ensure a more consistent approach. Preliminary plans for the new six-year medical program suggest that disparities between universities may persist, underscoring the need for a coordinated effort in standardizing disaster medicine education at the undergraduate level.

AIM: This study explores HICGs' experience of disaster response during a terrorist-induced major incident major incident.

DESIGN: A qualitative descriptive design with individual semi-structured interviews was used.

METHODS: This was a qualitative study based on seven individual interviews. Participants were members of hospital incident command groups during a terror attack. The interviews were transcribed verbatim and analyzed using deductive content analysis. The SRQR checklist was used to report the findings.

RESULTS: The data created from the interviews identified barriers and facilitators for hospital response as well as aligned with previously established categories: Expectations, prior experience, and uncertainty affect hospital incident command group response during a Major Incident and three categories, (I) Gaining situational awareness (containing two subcategories), (II) Transitioning to management (containing three subcategories) and (III) Experiences of hospital incident command group response (containing two subcategories). In addition, the results suggest that an exaggerated response may have led to unanticipated adverse events.

CLINICAL TRIAL NUMBER: Not applicable.

INTRODUCTION: The burden of critical illness may have been underestimated. Previous analyses have used data from intensive care units (ICUs) only, and there is a lack of evidence about where in hospitals critically ill patients receive care. This study aims to determine the burden of critical illness among adult inpatients across hospitals in different global settings.

METHODS: We performed a prospective, observational, hospital-based, point prevalence and cohort study in countries of different socioeconomic levels: Malawi, Sri Lanka and Sweden. On specific days, all adult in-patients in the eight study hospitals were examined by the study team for the presence of critical illness and followed up for hospital mortality. Patients with at least one severely deranged vital sign were classified as critically ill. The primary outcomes were the presence of critical illness and 30-day hospital mortality. In addition, we determined where the critically ill patients were being cared for and the association between critical illness and 30-day hospital mortality.

RESULTS: Among 3652 hospitalised patients, we found a point prevalence of critical illness of 12.0% (95% CI 11.0 to 13.1), with a hospital mortality of 18.7% (95% CI 15.3 to 22.6). The crude OR of death of critically ill patients compared with non-critically ill patients was 7.5 (95% CI 5.4 to 10.2). Of the critically ill patients, 96.1% (95% CI 93.9 to 97.6) were cared for in the general wards outside ICUs.

CONCLUSIONS: The study has revealed a substantial burden of critical illness in hospitals from different global settings. One in eight hospital in-patients was critically ill, 19% of the critically ill died in hospital, and 96% of the critically ill patients were cared for outside of ICUs. Implementing the most feasible and low-cost critical care in general wards throughout hospitals would impact a large number of high-risk patients and has the potential to improve outcomes across all acute care specialties.