The American journal of emergency medicine
-
The manual recording of electronic health records (EHRs) by clinicians in the emergency department (ED) is time-consuming and challenging. In light of recent advancements in large language models (LLMs) such as GPT and BERT, this study aimed to design and validate LLMs for automatic clinical diagnoses. The models were designed to identify 12 medical symptoms and 2 patient histories from simulated clinician-patient conversations within 6 primary symptom scenarios in emergency triage rooms. ⋯ This paper highlights the potential of LLMs for automatic EHR recording in Korean EDs. The KLUE-RoBERTa-based model demonstrated superior classification performance. Furthermore, XAI using SHAP provided reliable explanations for model outputs. The reliability of these explanations was confirmed by a Turing test.
-
Major trauma is a leading cause of unexpected death globally, with increasing age-adjusted death rates for unintentional injuries. Field triage schemes (FTSs) assist emergency medical technicians in identifying appropriate medical care facilities for patients. While full FTSs may improve sensitivity, step-by-step field triage is time-consuming. A simplified FTS (sFTS) that uses only physiological and anatomical criteria may offer a more rapid decision-making process. However, evidence for this approach is limited, and its performance in identifying all age groups requiring trauma center resources in Asia remains unclear. ⋯ sFTS using only physiological and anatomical criteria is suboptimal for Asian adult patients with trauma of all age groups. Adjusting the physiological criteria and adding a shock index as a triage tool can improve the sensitivity of severely injured patients, particularly in young age groups. A swift field triage process can maintain acceptable sensitivity and specificity in severely injured patients.
-
Timely hospital presentation and treatment are critical for recovery from coronavirus disease (COVID-19). However, the relationship between symptom onset-to-door time and key clinical outcomes, such as inpatient mortality, has been poorly understood due to the difficulty of retrospectively measuring symptom onset in observational data. This study examines the association between patient-reported symptom onset-to-door time (ODT) and mortality among patients hospitalized and treated for COVID-19 disease. ⋯ More days between symptom onset and hospital arrival were associated with lower mortality among hospitalized patients treated for COVID-19 disease, particularly if they did not have severe illness at ED presentation. However, onset-to-door time was not associated with mortality among hospitalized patients with severe illness at ED presentation. Collectively, these results suggest that non-severely ill COVID-19 patients who require hospitalization are less likely to decompensate with each passing day without severe illness. These findings may continue to guide clinical care delivery for hospitalized COVID-19 patients.
-
To evaluate if the change in end-tidal carbon dioxide (ETCO2) over time has improved discriminatory value for determining resuscitation futility compared to a single ETCO2 value in prolonged, refractory non-shockable out-of-hospital cardiac arrest (OHCA). ⋯ Nearly one-sixth of EMS-treated adult OHCA patients had refractory non-shockable arrests after at least 30 min of ongoing resuscitation. In this group, the ETCO2 trend following advanced airway placement may be more accurate in guiding termination of resuscitation than an absolute ETCO2 cutoff of 10 or 20 mmHg.
-
Boarding admitted patients in the emergency department is an important cause of throughput delays and safety risks in adults, though has been less studied in children. We assessed changes in boarding in a pediatric ED (PED) from 2018 to 2022 and modeled associations between boarding and select quality metrics. ⋯ Since July 2021, PED boarding time increased for admitted children across acute and critical admissions. The relationship between acute care boarding and longer hospital LOS suggests a resource-inefficient, self-perpetuating cycle that demands multi-disciplinary solutions.