Emergency Medicine

This observational study aims to use electronic health records to build an International Big Data Centre in Emergency Medicine, within the Institute of Sciences in Emergency Medicine (ISEM) at the Guangdong Provincial People's Hospital. The main questions it seeks to answer are not limited to the following: * Identify the relationship between Emergency Department Length of Stay (EDLOS), Mortality, and Adverse Events (AE) * Identify the risk factors associated with high mortality and AE rate among patients who experience prolonged EDLOS * Other research questions related to emergency medicine, such as building prediction and cluster models for acute diseases

This retrospective diagnostic accuracy study evaluates the ability of two large language models (LLMs) - GPT-4o (gpt-4o-2024-11-20; OpenAI) and Claude 4.6 Sonnet (claude-sonnet-4-6; Anthropic) - to generate correct diagnoses from anonymized Turkish-language emergency department (ED) anamnesis notes, and compares their performance with the diagnosis entered by the treating emergency physician. A consensus gold standard is established by three independent board-certified emergency medicine specialists who blindly review each note and vote on the primary diagnosis using ICD-10 three-character codes; the majority vote (at least 2 of 3 specialists agreeing) constitutes the reference standard. Both LLMs are evaluated using a standardized zero-shot direct prompting strategy (temperature=0, stateless API sessions). The primary outcome is diagnostic accuracy (proportion of ICD-10 chapter-level matches) and Cohen's kappa for each LLM against the gold standard. Secondary outcomes include top-3 accuracy, treating physician accuracy, inter-model agreement, and subgroup analyses by ESI triage level and ICD-10 chapter. Inter-rater reliability among the three specialists is quantified using Fleiss' kappa. Analyses are performed in Jamovi. This study represents the first evaluation of LLM diagnostic accuracy using Turkish-language clinical notes and the first to benchmark LLM performance against an independent three-specialist majority-vote gold standard rather than against the treating physician's own diagnosis.

This prospective observational diagnostic accuracy study evaluates whether large language models (LLMs) - GPT-4o (OpenAI, gpt-4o-2024-11-20) and Claude (Anthropic, claude-sonnet-4-6) - can accurately calculate HEART scores from unstructured Turkish clinical notes and predict 30-day major adverse cardiac events (MACE) in emergency department patients presenting with non-traumatic chest pain. The study will enroll 600 consecutive adult patients. For each patient, the same anonymized data (free-text anamnesis, ECG report text, troponin value, and age) will be independently processed by both LLMs via separate API calls with deterministic settings (temperature=0, JSON format). A three-expert consensus HEART score - derived through blinded independent scoring by three emergency medicine physicians with majority-vote adjudication - serves as the reference standard for agreement analysis. Actual 30-day MACE (all-cause death, AMI Type 1/2/4b, unplanned revascularization) determined via national health database and telephone follow-up serves as the outcome for diagnostic accuracy analysis. A secondary documentation-quality sub-study will quantify how spontaneously Turkish emergency anamnesis notes capture HEART score parameters.

This study aims to understand the health and social care needs and experiences of adults who frequently use urgent and emergency care services in Dorset. Using a mixed-methods design, the study combines analysis of non-patient-identifiable business intelligence data with qualitative interviews and co-production activities. The business intelligence data contextualises patterns of high intensity service use and informs participant identification. Qualitative interviews will explore the personal, social and system-level factors that contribute to frequent attendance. Co-production activities with an advisory group, supported by The Lantern Trust in Weymouth, will use these findings to develop a preventative intervention model grounded in lived experience. The study will recruit up to 50 patients, up to 10 carers and up to 20 health and social care professionals. The findings will contribute to the development of more effective, person-centred approaches to supporting people who frequently use urgent and emergency care services and will inform national and local policy in this area.

An emergency department (ED) is a healthcare service that provides the first clinical assessment and treatment to patients with various acute conditions. These departments, however, are often overwhelmed by the large volume of patients. As a consequence, ED crowding has become a global concern and has been correlated to reduced timeliness and effectiveness of care and increased patient mortality. Concerning input, 20% to 30% of patients are brought to the ED by ambulance; the remaining are self-presenting for the vast majority. Notably, non-urgent conditions characterize a high proportion of all ED visits worldwide, and almost all of these visits involve self-presenting patients. Increasing the awareness of these patients about the mandate of EDs and the real-time situation of the neighboring emergency departments has the potential to reduce the self-presentation of patients with minor, non-urgent conditions. Such patient empowerment can be achieved through a dashboard. Concerning throughput, working in the ED requires emergency physicians and nurses to treat many patients at once while maintaining situational awareness of the surroundings. This is especially true for the head of the department, but it also holds for all physicians. It can be crucial, for example, for physicians to know if there is a bottleneck in the flow of the entire patient care process, such as a particularly high average waiting time for radiology reporting or cardiologic consultation. The availability of this information allows countermeasures to be put in place to regain efficiency. All this can be achieved through dedicated dashboards automatically fed from various information system. In addition, appropriate dashboards also enable health policymakers to monitor specific epidemiological phenomena, such as the emergence of certain infectious diseases, in a timely manner.

The peer-to-peer comparison means center-to-center comparison, which requires adjusting for possible differences among centers to be fair and convincing. The first step to reach this goal is to develop a predictive model that accurately estimates each patient's probability of being admitted, starting from clinical conditions and boundary variables. Such a model would make it possible to calculate, for each ED, the expected hospitalization rate; that is, the hospitalization rate that would have been observed if the ED had behaved like the average of the EDs that provided the data to build the model itself. Comparing the observed hospitalization rate in the single ED with the expected rate derived from the model provides a rigorous method of comparing the department with the average performance, taking into account the characteristics of the patients treated and the conditions under which the ED operated. In other words, the predictive model represents the benchmark against which each ED is evaluated.

The goal of this retrospective cohort study is to develop and validate a language model that can interpret the contents of emergency department electronic medical records and extract relevant information for research purposes in all adult patients who arrived at the participating emergency departments in a three-year period. The main question it aims to answer is: is the language model able to interpret the contents of emergency department electronic medical records and extract the requested information from them so that it can be used to make accurate analyses and predictions? The study is retrospective and data will be extracted automatically from the medical health records.

This single-center randomized controlled study aims to evaluate the effectiveness of Case-Based Learning (CBL) compared with traditional teaching in clinical training of bedside lung ultrasound (BLUE) for emergency medicine residents and medical students. The hypothesis is that CBL improves theoretical understanding, practical ultrasound skills, and clinical reasoning in emergency settings.

The investigators will implement a protocol for use of the PIVO device for blood culture collection in the emergency department. This protocol utilizes the device in various scenarios to reduce needlesticks while allowing for accurate and appropriate assessment of potential pathogens in the bloodstream. The study team will train staff on use of the device then monitor utilization and contamination rates among emergency department patients.

The goal of this clinical trial is to find out if a personalized treatment approach can improve care for people with sepsis in the emergency department (ED). Sepsis is a life-threatening condition that happens when the body has an uncontrolled response to an infection. This can lead to low blood pressure, organ failure, and death if not treated quickly. Right now, most people with sepsis receive a standard amount of fluids to raise their blood pressure. But this one-size-fits-all approach can lead to fluid overload and other complications. Because each person responds differently, this study will test whether a more personalized treatment-based on how the heart responds to fluids-can lead to safer and more effective care. The study will include 188 adults who come to the ED at the University Medical Centre Groningen (UMCG) with suspected sepsis in need of hemodynamic resuscitation. Everyone in the study will receive fluids to support their blood pressure. Participants will be randomly assigned to one of two groups: * Personalized treatment group: Fluids and vasopressors (medications that raise blood pressure) will be given based on how the heart responds to each fluid dose. This response is measured using a non-invasive monitor that tracks stroke volume index (ΔSVI)-a measure of how much blood the heart pumps. * Standard care group: Fluids will be given based on current guidelines (30 milliliters per kilogram of body weight), as decided by the treating doctor. Researchers will compare how much fluid is given during the first 3 hours of care. They will also look at: * When and how much vasopressor medicine is used * How well blood pressure and circulation respond * Signs of organ recovery or damage * How long participants stay in the hospital * Any problems or side effects during treatment The researchers hope that this personalized approach will lead to using less fluid, starting vasopressors earlier, and helping people with sepsis recover more safely and quickly.

Although the Manchester Triage System (MTS) is widely used and validated internationally, it has some limitations. Its accuracy is moderate, especially for children and the elderly. Rising patient numbers and overcrowded emergency departments increase wait times, sometimes beyond safe limits. In Slovenia, MTS has been in use for 14 years without major updates, despite a significant rise in emergency visits. The yellow triage category (60-minute wait time) includes a very diverse group of patients, some of whom might require faster care. Older patients, in particular, often show atypical symptoms and may be under-triaged. Including rapid bedside lab tests, like blood lactate levels, could improve risk assessment and triage accuracy. Elevated lactate is linked with higher mortality and can help identify critically ill patients more effectively. The proposed study is a prospective, randomized trial involving two groups of patients in the yellow triage category, all of whom will have their capillary blood lactate levels measured. Patients with normal lactate levels will be excluded. Only patients with elevated lactate will be compared. The test group will be re-triaged to the orange category and treated more urgently. The control group, despite also having high lactate levels, will remain in the yellow category, and their elevated lactate values will not be shared with the treating physician. Randomization will be based on the patient's birth date (even days = test group, odd days = control group). Only the nurse will know the result, maintaining physician blinding to avoid the Hawthorne effect-changes in behavior due to awareness of being studied. Standard lab tests will be performed later during treatment as deemed necessary by the attending doctor.

The objective of this study is to pilot a multifaceted, optimized intervention for metabolic syndrome (MetS) in emergency department patients to establish feasibility. Participants (n=20) will be randomized to intervention or control (usual care). The composite intervention will include an educational video outlining the adverse effects of MetS and the benefit of walking, a written exercise prescription with a defined goal of walking 150 minutes per week, a Fitbit accelerometer device, resources for healthy eating practices, periodic text message reminders, and an urgent referral to primary care and our health system's Healthy Me clinic for follow-up visit. Investigators hypothesize that this approach will change patient understanding and motivation to increase physical activity and healthy eating habits.

known hypertensive patients admitted to emergency department with increased blood pressure will be evaluated in terms of antihypertensive agents given at hospital, degree of blood pressure decrease, hospital stay and laboratory and imaging tests ordered. The impact of routine oral antihypertensive agents used by the patients on these parameter will be assessed.