Academic Emergency Medicine最新文献

Objective: The Global Emergency Medicine Literature Review (GEMLR) highlights the highest-quality research addressing emergency care in resource-limited settings (ECRLS). This 18th edition reviews global emergency medicine (GEM) literature published during 2023.

Methods: A scoping review of GEM articles published in 2023 was performed using a systematic PubMed search and manual gray literature (GRAY) search. Reviewers and editors from 10 countries screened articles utilizing case definitions of three categories of GEM research-disaster and humanitarian response (DHR), ECRLS, and emergency medicine development (EMD). After duplicates and those not meeting authorship equity and ethical review requirements were removed, articles were scored according to rubrics for original research (OR), review articles (RE), and GRAY. Those in the top 5% from any category were summarized and critiqued in narrative review.

Results: There were 58,291 articles identified in the main search and 11,035 in the GRAY search. A total of 825 articles from the main search and 37 GRAY articles screened in and were scored. Fifty-five main search articles and one GRAY article were included after scoring, a 52.8% increase from 2022 despite <1% change in search volume. ECRLS remained the largest category (63%). As in previous years, articles frequently addressed emergencies in pediatrics (10 articles), trauma (9), prehospital care (8), maternal/neonatal care (6), education/training (6), disaster medicine (4), and airway/sedation management (4). A total of 3.5% of screened-in articles failed to meet GEMLR's new authorship equity and ethics standards.

Conclusions: The quantity and quality of GEM research continues to grow as measured by the GEMLR scoring system. A revised search string identified relevant GEM articles with broad application in global settings. New equity guidelines were successfully implemented. This review summarizes the highest quality current GEM research while providing evolving guidelines for best practices in performing this important and rapidly growing work.

Investigating error in healthcare has long been a central focus of patient safety efforts, yet this approach oversimplifies the complexities of a deeply interconnected and dynamic system. This paper argues that framing patient harm solely as "error" has failed to yield meaningful improvements, as it overlooks the broader factors contributing to adverse outcomes. Through a case study, we illustrate how an exclusive focus on error investigation missed critical insights and propose that attention should instead be directed towards identifying and managing hazards. Emphasizing the importance of treating healthcare as a sociotechnical system, we argue for the application of sociotechnical system thinking, particularly in addressing diagnostic errors in high-pressure environments like emergency medicine. The traditional methods, such as root cause analysis, error investigation, and the Swiss cheese model, are no longer sufficient. These outdated frameworks fail to address the systemic challenges clinicians face and often misguide safety interventions. The paper calls for a paradigm shift towards re-designing healthcare systems using human factors research to better align with the complexities of clinical work and the hazards that lead to harm.

Background: Diagnostic stewardship is the effort to optimize diagnostic testing to reduce errors while avoiding overtesting and overtreatment. Abdominal pain and appendicitis in children are essential use cases. Delayed diagnosis of appendicitis can be dangerous and even life-threatening, but overtesting is harmful.

Methods: We conducted a retrospective cohort study of children aged 5-17 years presenting with abdominal pain to 26 EDs within the Michigan Emergency Department Improvement Collaborative (MEDIC) from May 1, 2016, to February 29, 2024. We defined two outcome measures summarized by ED. First, we describe the cross-sectional imaging:appendicitis visits ratio, defined as the count of ED visits resulting in any cross-sectional imaging (CT or MRI) divided by the count of ED visits with a diagnosis of appendicitis. Second, we describe the delayed diagnosis rate, defined by an ED visit for abdominal pain resulting in a discharge and subsequent return visit with a diagnosis of appendicitis within 7 days.

Results: The sample included 120,112 pediatric visits for abdominal pain at 26 EDs; 4967 (4.1%) were diagnosed with appendicitis. The cross-sectional imaging:appendicitis visits ratio varied by site, from as low as 0.2 (95% confidence interval [CI] 0.1-0.2) at a pediatric site to as high as 7.9 (95% CI 4.8-16.4) at an urban ED. The proportion of pediatric ED visits for abdominal pain that resulted in an identified delayed diagnosis of appendicitis was 0.1% (141/120,112). All but four sites had fewer than 10 cases of delayed diagnosis across the study period.

Conclusions: In this retrospective cohort study of 120,000+ ED visits for pediatric abdominal pain, we found that the ratio of visits with cross-sectional imaging to diagnosed cases of appendicitis varied widely across EDs. Delayed diagnosis of appendicitis was uncommon. Adherence to best practices and improved imaging quality may hold promise to improve diagnostic stewardship for children with abdominal pain across EDs.

Objectives: We applied three electronic triggers to study frequency and contributory factors of missed opportunities for improving diagnosis (MOIDs) in pediatric emergency departments (EDs): return visits within 10 days resulting in admission (Trigger 1), care escalation within 24 h of ED presentation (Trigger 2), and death within 24 h of ED visit (Trigger 3).

Methods: We created an electronic query and reporting template for the triggers and applied them to electronic health record systems of five pediatric EDs for visits from 2019. Clinician reviewers manually screened identified charts and initially categorized them as "unlikely for MOIDs" or "unable to rule out MOIDs" without a detailed chart review. For the latter category, reviewers performed a detailed chart review using the Revised Safer Dx Instrument to determine the presence of a MOID.

Results: A total of 2937 ED records met trigger criteria (Trigger 1 1996 [68%], Trigger 2 829 [28%], Trigger 3 112 [4%]), of which 2786 (95%) were categorized as unlikely for MOIDs. The Revised Safer Dx Instrument was applied to 151 (5%) records and 76 (50%) had MOIDs. The overall frequency of MOIDs was 2.6% for the entire cohort, 3.0% for Trigger 1, 1.9% for Trigger 2, and 0% for Trigger 3. Brain lesions, infections, or hemorrhage; pneumonias and lung abscess; and appendicitis were the top three missed diagnoses. The majority (54%) of MOIDs cases resulted in patient harm. Contributory factors were related to patient-provider (52.6%), followed by patient factors (21.1%), system factors (13.2%), and provider factors (10.5%).

Conclusions: Using electronic triggers with selective record review is an effective process to screen for harmful diagnostic errors in EDs: detailed review of 5% of charts revealed MOIDs in half, of which half were harmful to the patient. With further refining, triggers can be used as effective patient safety tools to monitor diagnostic quality.

Background: Nonconsent to pulmonary vascular (or advanced) imaging for suspected pulmonary embolism (PE) in pregnancy can delay diagnosis and treatment, increasing risk of adverse outcomes. We sought to understand factors associated with consent and understand outcomes after nonconsent.

Methods: This retrospective cohort study was undertaken across 21 community hospitals from October 1, 2021, through March 31, 2023. We included gravid patients undergoing diagnostics for suspected PE who were recommended advanced imaging. The primary outcome was verbal consent to advanced imaging. Diagnostic settings were nonobstetric (99% emergency departments [EDs]) and obstetrics (labor and delivery and outpatient clinics). Using quasi-Poisson regression, we calculated adjusted relative risks (aRRs) of consenting with 95% confidence intervals (CIs). We also reported symptom resolution and delayed imaging at follow-up and 90-day PE outcomes.

Results: Imaging was recommended for 405 outpatients: median age was 30.5 years; 50% were in the third trimester. Evaluation was more common in nonobstetric (83%) than obstetric settings (17%). Overall, 314 (78%) agreed to imaging and 91 (22%) declined imaging. Consenting was more prevalent in obstetric settings compared with nonobstetric settings: 99% versus 73% (p < 0.001). When adjusted for demographic and clinical variables, including pretest probability, only obstetric setting was independently associated with consenting: aRR 1.26 (95% CI 1.09-1.44). Seventy-nine (87%) patients declining imaging had 30-day follow-up. Eight of 12 who reported persistent or worsening symptoms on follow-up were again recommended advanced imaging and consented. Imaging was negative. None who initially declined imaging were diagnosed with PE or died within 90 days.

Conclusions: One in five gravid patients suspected of PE declined advanced imaging, more commonly in nonobstetric (principally ED) settings than obstetric settings. Patients symptomatic on follow-up responded favorably to subsequent imaging recommendations without 90-day outcomes. Improving the communication and documentation of informed consent and securing close follow-up for nonconsenters may mitigate risks of missed and delayed PE diagnosis.

Objectives: For emergency department (ED) patients, lung cancer may be detected early through incidental lung nodules (ILNs) discovered on chest CTs. However, there are significant errors in the communication and follow-up of incidental findings on ED imaging, particularly due to unstructured radiology reports. Natural language processing (NLP) can aid in identifying ILNs requiring follow-up, potentially reducing errors from missed follow-up. We sought to develop an open-access, three-step NLP pipeline specifically for this purpose.

Methods: This retrospective used a cohort of 26,545 chest CTs performed in three EDs from 2014 to 2021. Randomly selected chest CT reports were annotated by MD raters using Prodigy software to develop a stepwise NLP "pipeline" that first excluded prior or known malignancy, determined the presence of a lung nodule, and then categorized any recommended follow-up. NLP was developed using a RoBERTa large language model on the SpaCy platform and deployed as open-access software using Docker. After NLP development it was applied to 1000 CT reports that were manually reviewed to determine accuracy using accepted NLP metrics of precision (positive predictive value), recall (sensitivity), and F1 score (which balances precision and recall).

Results: Precision, recall, and F1 score were 0.85, 0.71, and 0.77, respectively, for malignancy; 0.87, 0.83, and 0.85 for nodule; and 0.82, 0.90, and 0.85 for follow-up. Overall accuracy for follow-up in the absence of malignancy with a nodule present was 93.3%. The overall recommended follow-up rate was 12.4%, with 10.1% of patients having evidence of known or prior malignancy.

Conclusions: We developed an accurate, open-access pipeline to identify ILNs with recommended follow-up on ED chest CTs. While the prevalence of recommended follow-up is lower than some prior studies, it more accurately reflects the prevalence of truly incidental findings without prior or known malignancy. Incorporating this tool could reduce errors by improving the identification, communication, and tracking of ILNs.