Explore the vast range of titles available.

Large language models (LLMs) have demonstrated impressive performances in various medical domains, prompting an exploration of their potential utility within the high-demand setting of emergency department (ED) triage. This study evaluated the triage proficiency of different LLMs and ChatGPT, an LLM-based chatbot, compared to professionally trained ED staff and untrained personnel. We further explored whether LLM responses could guide untrained staff in effective triage. This study aimed to assess the efficacy of LLMs and the associated product ChatGPT in ED triage compared to personnel of varying training status and to investigate if the models' responses can enhance the triage proficiency of untrained personnel. A total of 124 anonymized case vignettes were triaged by untrained doctors; different versions of currently available LLMs; ChatGPT; and professionally trained raters, who subsequently agreed on a consensus set according to the Manchester Triage System (MTS). The prototypical vignettes were adapted from cases at a tertiary ED in Germany. The main outcome was the level of agreement between raters' MTS level assignments, measured via quadratic-weighted Cohen κ. The extent of over- and undertriage was also determined. Notably, instances of ChatGPT were prompted using zero-shot approaches without extensive background information on the MTS. The tested LLMs included raw GPT-4, Llama 3 70B, Gemini 1.5, and Mixtral 8x7b. GPT-4-based ChatGPT and untrained doctors showed substantial agreement with the consensus triage of professional raters (κ=mean 0.67, SD 0.037 and κ=mean 0.68, SD 0.056, respectively), significantly exceeding the performance of GPT-3.5-based ChatGPT (κ=mean 0.54, SD 0.024; P<.001). When untrained doctors used this LLM for second-opinion triage, there was a slight but statistically insignificant performance increase (κ=mean 0.70, SD 0.047; P=.97). Other tested LLMs performed similar to or worse than GPT-4-based ChatGPT or showed odd triaging behavior with the used parameters. LLMs and ChatGPT models tended toward overtriage, whereas untrained doctors undertriaged. While LLMs and the LLM-based product ChatGPT do not yet match professionally trained raters, their best models' triage proficiency equals that of untrained ED doctors. In its current form, LLMs or ChatGPT thus did not demonstrate gold-standard performance in ED triage and, in the setting of this study, failed to significantly improve untrained doctors' triage when used as decision support. Notable performance enhancements in newer LLM versions over older ones hint at future improvements with further technological development and specific training.

Context Specialized outpatient palliative care (SOPC) aims at relieving symptoms and providing psychosocial aid in the outpatient setting. SOPC also supports patients in setting up advance directives so that their will is respected in emergency situations. Objectives The aim of this study is to analyze the impact of SOPC on the completion of advance care directives with the focus on medical emergency identification (ID) cards (= living will in a credit card format on yellow paper) and their impact on hospital admissions. Methods All adult patients who were admitted to the SOPC service at a tertiary care center in Germany between 07/2022 and 06/2023 were included in this retrospective cohort study. Demographic data, level of care, information on advance care directives, hospitalizations, adherence to patient wishes, and tumor specific treatments were collected. The data were analyzed using descriptive and inferential measures. Results During the study period, 359 patients were included (52.0% female, mean age 74 ± 13 years). A medical emergency ID card was set up by 32.6% ( n  = 117/359). It was significantly more likely to be created during SOPC than prior to SOPC involvement (before: 20%, after: 80%; p  < 0.05). Patients who expressed not wanting to be admitted to hospital in their medical emergency ID card saw significantly less hospital admissions ( p  < 0.05). Conclusion SOPC supports patients in setting up medical emergency ID cards. These help in respecting patients’ wishes and prevent unwanted admissions to hospitals, thereby reducing strain on emergency services and emergency departments.

Background The management of critically ill patients, arriving at the emergency department (ED), requires structured care in critical care facilities, particularly in the resuscitation room. This study examines the significance of initial vital signs and blood gas analysis (BGA)-derived values as clinically useful early indicators of mortality risk in critically ill patients, both during in the resuscitation room care and within the following 30 days, with a focus on evaluating the individual predictive performance of accessible clinical parameters. Methods We pooled data from two consecutive retrospective observational studies in a German university ED to analyze an unselected patient population of non-traumatic critically ill patients. Vital signs, such as heart rate, systolic blood pressure, and BGA values (including pH, bicarbonate, carbon dioxide, glucose, lactate, electrolyte levels) on admission to the ED, were used to estimate the impact on both resuscitation room and 30-day mortality. Results In 1,536 critically ill patients, pH, lactate and bicarbonate were found to be potential predictors of resuscitation room mortality. In contrast, vital signs showed limited reliability in predicting outcomes. Of all tested variables, pH demonstrated the highest area under the curve (AUC) value among the analyzed markers for resuscitation room mortality (AUC 0.81 [95% CI 0.75–0.87]). However, the AUC of pH for 30-day mortality decreased to 0.64 ([0.6 – 0.68], indicating a complex interplay of factors influencing long-term outcome. A subgroup analysis based on pH showed a substantial increase in resuscitation room and 30-day mortality for patients with a pH below 7.2 as well as a second increase below 7.0. Conclusion Our study highlights important parameters for the assessment of critically ill patients at ED admission that are helpful for formulating immediate medical decisions. Acidosis on the initial BGA appears to be a relevant prognostic marker for mortality in critically ill, non-traumatic patients and may aid in early risk assessment, regardless of the underlying condition. Early detection of acidosis could facilitate rapid decision-making and timely identification of patients requiring intensive care.

BackgroundThe adjacent conditions infection, sepsis, and septic shock are among the most common causes of treatment in the emergency department (ED). Most available data come from intensive care units (ICU) and include nosocomial infections acquired during hospitalization. Epidemiological data from German EDs are not yet available, although the ED is one of the first points of contact for patients. The aim of this study was to investigate the epidemiology, causes, diagnosis, mortality, and treatment of patients with infections in the ED.Materials and methodsIn this retrospective, single-center observational study, routinely collected data from the patient data management system and from the hospital information system were analyzed. All adult patients who presented to the ED in connection with an infection during the study period from 01/01 to 28/02/2019 were included. Exclusion criteria were age ≤ 17 years and incomplete records. Three groups (I. Infection, II. Sepsis, and III. Septic shock) were defined according to SEPSIS-3 definitions.ResultsDuring the study period, a total of 6,607 patients were treated in the ED. Of these patients, 19.3% ( n = 1,278) had an infection (mean age 56 ± 23 years, 50% female). The sites of infection were distributed as follows: Respiratory tract 35%, genitourinary tract 18%, maxillofacial/ears/nose/throat 14%, intraabdominal 13%, soft tissues 10%, central nervous system 1%, other cause 3%, or unknown cause 6%. Infection only, sepsis and septic shock were present in 86, 10, and 3%, respectively. There were significant differences in vital signs as well as in the various emergency sepsis scores across the predefined groups [I vs. II vs. III: SOFA (pts.): 1 ± 1 vs. 4 ± 2 vs. 7 ± 3 ( p < 0.0001), systolic blood pressure (mmHg): 137 ± 25 vs. 128 ± 32 vs. 107 ± 34 ( p < 0.05), heart rate (bpm): 92 ± 18 vs. 99 ± 23 vs. 113 ± 30 ( p < 0.05), respiratory rate (min-1): 18 ± 4 vs. 20 ± 7 vs. 24 ± 10 ( p < 0.05)]. In the three groups, blood cultures were obtained in 34, 81, and 86%, of cases, respectively and antibiotics were administered in the ED in 50, 89, and 86%, of cases respectively. The 30-day mortality rate in the three groups was 1.6, 12.0, and 38.1%, respectively.ConclusionThis study is the first to show the incidence, management, and outcome of patients classified as infection, sepsis, and septic shock in a German ED. The findings of our real-world data are important for quality management and enable the optimization of treatment pathways for patients with infectious diseases.

Background. Sepsis is defined as a life-threatening organ dysfunction due to a dysregulated host response to infection. Sepsis is the dominant cause of acute kidney injury (AKI), accounting for nearly 50% of episodes of acute renal failure. Signaling cascades and pathways within the kidney are largely unknown and analysis of these molecular mechanisms may enhance knowledge on pathophysiology and possible therapeutic options. Material and Methods. 26 male Wistar rats were assigned to either a sham group (control, N=6) or sepsis group (N=20; cecal ligature and puncture model, 24 and 48 hours after CLP). Surviving rats (n=12) were decapitated at 24 hours (early phase; n=6) or 48 hours (late phase; n=6) after CLP and kidneys removed for proteomic analysis. 2D-DIGE and DeCyder 2D software (t-test, P<0.01) were used for analysis of significantly regulated protein spots. MALDI-TOF in combination with peptide mass fingerprinting (PMF) as well as Western Blot analysis was used for protein identification. Bioinformatic network analyses (STRING, GeneMania, and PCViz) were used to describe protein-protein interactions. Results. 12 spots were identified with significantly altered proteins (P<0.01) in the three analyzed groups. Two spots could not be identified. Four different proteins were found significantly changed among the groups: major urinary protein (MUP5), cytochrome c oxidase subunit B (COX5b), myosin-6 (MYH6), and myosin-7 (MYH7). A significant correlation with the proteins was found for mitochondrial energy production and electron transport. Conclusions. COX5B could be a promising biomarker candidate since a significant association was found during experimental sepsis in the present study. For future research, COX5B should be evaluated as a biomarker in both human urine and serum to identify sepsis.

Administration of oxygen is one of the most common interventions in medicine. Previous research showed that differential regulated proteins could be linked to hyperoxia-associated signaling cascades in different tissues. However, it still remains unclear which signaling pathways are activated by hyperoxia. The present study analyses hyperoxia-induced protein alterations in lung, brain, and kidney tissue using a proteomic and bioinformatic approach. Pooled data of 36 Wistar rats exposed to hyperoxia were used. To identify possible hyperoxia biomarkers, and to evaluate the relationship between protein alterations in hyperoxia affected organs and blood, proteomics data from brain, lung, and kidney were analyzed. Functional network analyses (IPA®, PathwaysStudio®, and GENEmania®) in combination with hierarchical cluster analysis (Perseus®) was used to identify relevant pathways and key proteins. Data of 54 2D-gels with more than 2500 significantly regulated spots per gel were collected. Thirty-eight differentially expressed proteins were identified and consecutively analyzed by bioinformatic methods. Most differences between hyperoxia and normoxia (21 proteins up-regulated, 17 proteins down-regulated) were found immediately after hyperoxia (15 protein spots), followed by day 3 (13 spots), and day 7 (10 spots). A highly significant association with inflammation and the inflammatory response was found. Cell proliferation, oxidative stress, apoptosis and cell death as well as cellular functions were revealed to be affected. Three hours of hyperoxia resulted in significant alterations of protein expression in different organs (brain, lung, kidney) up to seven days after exposure. Further studies are required to interpret the relevance of protein alterations in signaling cascades during/after hyperoxia.

Data on the incidence of in-flight medical emergencies on-board civil aircraft are uncommon and rarely published. Such data could provide information regarding required medical equipment on-board aircraft and requisite training for cabin crew. The aim of the present study was to gather data on the incidences, nature, and medical equipment for in-flight medical emergencies by way of a survey of physician members of a German aerospace medical society. Using unipark.de (QuestBack GmbH, Cologne, Germany), an online survey was developed and used to gather specific information. Members of the German Society for Aviation and Space Medicine (Deutsche Gesellschaft für Luft- und Raumfahrtmedizin e.V.; DGLRM) were invited to participate in the survey during a 4-week period (21 March 2015 to 20 April 2015). Chi-square test was used for statistical analysis ( <0.05 was considered significant). Altogether, 121 members of the society responded to the survey (n=335 sent out). Of the 121 respondents, n=54 (44.6%) of the participants (89.9% male and 10.1% female; mean age, 54.1 years; n=121) were involved in at least one in-flight medical emergency. Demographic parameters in this survey were in concordance with the society members' demographics. The mean duration of flights was 5.7 hours and the respondents performed 7.1 airline flights per year (median). Cardiovascular (40.0%) and neurological disorders (17.8%) were the most frequent diagnoses. The medical equipment (78.7%) provided was sufficient. An emergency diversion was undertaken in 10.6% of the cases. Although using a different method of data acquisition, this survey confirms previous data on the nature of emergencies and gives plausible numbers. Our data strongly argue for the establishment of a standardized database for recording the incidence and nature of in-flight medical emergencies. Such a database could inform on required medical equipment and cabin crew training.

By the end of the year 2016, approximately 3 billion people worldwide travelled by commercial air transport. Between 1 out of 14,000 and 1 out of 50,000 passengers will experience acute medical problems/emergencies during a flight (i.e., in-flight medical emergency). Cardiac arrest accounts for 0.3% of all in-flight medical emergencies. So far, no specific guideline exists for the management and treatment of in-flight cardiac arrest (IFCA). A task force with clinical and investigational expertise in aviation, aviation medicine, and emergency medicine was created to develop a consensus based on scientific evidence and compiled a guideline for the management and treatment of in-flight cardiac arrests. Using the GRADE, RAND, and DELPHI methods, a systematic literature search was performed in PubMed. Specific recommendations have been developed for the treatment of IFCA. A total of 29 specific recommendations for the treatment and management of in-flight cardiac arrests were generated. The main recommendations included emergency equipments as well as communication of the emergency. Training of the crew is of utmost importance, and should ideally have a focus on CPR in aircraft. The decision for a diversion should be considered very carefully.

According to the WHO classification, anaplastic large cell lymphoma (ALCL) is a distinct T-cell lymphoma entity with a number of morphological variants. The characteristic feature of lymphohistiocytic variant of ALCL according to the WHO classification is the abundance of histiocytes that exceed and mask the tumour cell population. In the current, study we reanalysed a historical series of 17 lymphomas, diagnosed as lymphohistiocytic lymphoma according to the criteria of the Kiel classification, with the presence of large purple macrophages (LPM) as the decisive finding for diagnosing this lymphoma subtype. We assessed the cellular composition of the tumour and correlated the results with the definition of lymphohistiocytic variant of ALCL given in the WHO classification. Although all cases in our cohort matched the criteria of ALCL according to the WHO, in 30% of the cases, the total amount of macrophages did not exceed the number of CD30-positive tumour cells. Our results indicate that the presence of LPM might be helpful to identify this subgroup of ALCL. Because the distinction of morphological subtypes of ALCL is of clinical relevance, improved criteria for subtyping ALCL are urgently needed that might include the presence LPM as one criteria.