Emergency Department Clinicians Research Articles

Can clinicians identify community-acquired pneumonia on ultralow-dose CT? A diagnostic accuracy study

BackgroundWithout increasing radiation exposure, ultralow-dose computed tomography (CT) of the chest provides improved diagnostic accuracy of radiological pneumonia diagnosis compared to a chest radiograph. Yet, radiologist resources to rapidly report the chest CTs are limited. This study aimed to assess the diagnostic accuracy of emergency clinicians’ assessments of chest ultralow-dose CTs for community-acquired pneumonia using a radiologist’s assessments as reference standard.MethodsThis was a cross-sectional diagnostic accuracy study. Ten emergency department clinicians (five junior clinicians, five consultants) assessed chest ultralow-dose CTs from acutely hospitalised patients suspected of having community-acquired pneumonia. Before assessments, the clinicians attended a focused training course on assessing ultralow-dose CTs for pneumonia. The reference standard was the assessment by an experienced emergency department radiologist. Primary outcome was the presence or absence of pulmonary opacities consistent with community-acquired pneumonia. Sensitivity, specificity, and predictive values were calculated using generalised estimating equations.ResultsAll clinicians assessed 128 ultralow-dose CTs. The prevalence of findings consistent with community-acquired pneumonia was 56%. Seventy-eight percent of the clinicians’ CT assessments matched the reference assessment. Diagnostic accuracy estimates were: sensitivity = 83% (95%CI: 77–88), specificity = 70% (95%CI: 59–81), positive predictive value = 80% (95%CI: 74–84), negative predictive value = 78% (95%CI: 73–82).ConclusionThis study found that clinicians could assess chest ultralow-dose CTs for community-acquired pneumonia with high diagnostic accuracy. A higher level of clinical experience was not associated with better diagnostic accuracy.

Care conundrum in the emergency department: The gap between clinician awareness and patient expectations surrounding advance directives.

Clinicians in the emergency department (ED) frequently encounter seriously ill patients at a time when advance directives may be pivotal in improved clinician decision-making. The objectives of this study were to identify the prevalence of advanced directives in ED patients, as well as patterns of advance care discussions between patients and providers. This study describes patients' perceptions and expectations of such serious illness discussions in an emergency care setting with the expectation of including patients as strategic members of the care team. Trained research assistants in two emergency departments surveyed patients over age 65, or their caregivers, from July 2016 to August 2018. Patients were verbally administered a standard survey tool related to advance directives and advance care planning. 497 out of 877 patients completed surveys (59.4%). 50% of patients reported having an advance care planning document. The large majority (92%) of patients with an advance directive had not been asked about it during their ED visit. When questioned about their personal preferences, 79% of patients thought emergency physicians should be aware of their wishes regarding life-sustaining treatments and end-of-life care. Paradoxically, only 38% expressed a desire to discuss advance care plans with an ED clinician. Older patients expect emergency clinicians to be aware of their care preferences, yet most are not asked about these care preferences in the ED. The large gap between patient preference and reality suggests the need for more targeted discussion by ED clinicians and translation of patient perspectives into system healthcare improvements. Future studies should explore barriers to advance care planning in the ED as well as patient preferences for these conversations to support a true healthcare learning system.

Understanding uptake of information about innovations among emergency department clinicians during the COVID-19 pandemic.

Early in the COVID-19 pandemic, little was known about managing sick patients, but emergency department (ED) clinicians had to decide which treatments and care processes to adopt. Our objective was to describe how ED clinicians learned about innovations and how they assessed them for credibility during the pandemic. We purposively sampled clinicians from hospital-based EDs to conduct focus groups with ED clinicians and staff. We used both inductive and deductive approaches to conduct thematic analysis of transcripts. We conducted focus groups with clinicians from eight EDs across the United States. We found that ED clinicians in our sample relied on friends and colleagues or departmental and institutional leadership for information on innovations. They used social media sources when they came from credible accounts but did not directly seek information from professional societies. Clinicians reported a range of challenges to obtain credible information during the pandemic, including a fractured and changing information environment, policies misaligned across clinical sites or that conflicted with clinical knowledge, high patient volume, fear of harming patients, and untimely information. Facilitators included access to experienced and trusted colleagues and leaders and practicing at multiple EDs. Participants cited anecdotal evidence, institutional practice, and word-of-mouth-rather than peer-reviewed evidence and professional society communications-as their primary sources of information about care innovations during the early phases of the pandemic. These results underscore the importance of developing trusted local mechanisms and wider networks to identify and vet information for frontline clinicians during rapidly emerging public health emergencies.

Association between homelessness and temperature-related health emergencies in the United States

Abstract Temperature-related illness (TRI) encompasses heat-related illness, such as heat exhaustion and heatstroke, and cold-related illness, such as frostbite and hypothermia. TRI is typically the result of exposure to ambient weather conditions; because of this, unhoused individuals are hypothesized to have higher risk of TRI. However, no national epidemiological studies have been completed to determine this risk. The objective of this study was to determine the association between homelessness and emergency department (ED) diagnosis of TRI in the United States. We conducted a cross-sectional study of adult ED visits in the U.S. from 2005 through 2020 using data from the National Hospital Ambulatory Medical Care Survey (NHAMCS), a nationally representative sample of non-federal ED patient visits. Housing status (housed vs. unhoused) was measured using NHAMCS patient residence category, with blank responses excluded. TRI was defined as ED clinician diagnosis of heat- or cold-related illness using ICD-9 and ICD-10 codes. Multivariable logistic regression was used to determine adjusted odds of TRI by housing status. There were 323,606 non-pediatric ED visits in the NHAMCS sample. TRI diagnosis was present in 288 (0.09%) visits. 4099 visits (0.9%) were categorized as unhoused. After adjusting for sex, mental health diagnosis, and alcohol or substance use or use disorder, the odds of TRI diagnosis in unhoused individuals was 4.08 (95% CI 2.09,7.95) compared to housed individuals. We found a higher adjusted odds of TRI diagnosis at an ED visit among unhoused individuals compared with housed individuals.

Lessons Learned Recruiting Emergency Department Clinicians for an ICT Implementation Science Study.

Collaborating with health system leaders and clinicians is essential for successful implementation of interventions, especially when using new information and communication technologies (ICTs). Recruiting clinicians in research poses challenges, such as time constraints. This poster shares lessons learned from a study conducted in a psychiatric emergency department, along with recommendations for researchers seeking to effectively recruit clinicians when introducing a new ICT.

Clinical Characteristics, Outcomes, and Interobserver Agreement of Point-of-Care Ultrasound Detected Mesenteric Adenitis in Nonsurgical Pediatric Abdominal Pain: A Retrospective Cohort Study.

Point-of-care ultrasound (PoCUS) in the emergency department (ED) may facilitate the diagnosis of nonsurgical sources of abdominal pain after surgical causes are excluded. Identifying mesenteric adenitis is a feasible PoCUS application due to its ease of use and speed. However, there are scant data regarding the diagnosis of mesenteric adenitis by PoCUS. The objective of this study was to describe the clinical characteristics, outcomes, and interobserver agreement of mesenteric adenitis identified on PoCUS in pediatric patients with nonsurgical abdominal pain. This was a retrospective review at a tertiary-care, urban pediatric ED. All cases of mesenteric adenitis diagnosed on PoCUS from January 2018 to August 2022 were reviewed. Demographics and clinical data, including relevant outcomes, were recorded. All PoCUS videos were reviewed by a senior sonologist-physician for determination of mesenteric adenitis in children 21 years and younger with nonsurgical abdominal pain. Interobserver agreement by Cohen κ was calculated between experienced and novice physician sonologists blinded to diagnosis, who reviewed 77 six-second video clips for presence or absence of mesenteric adenitis. Thirty-three subjects were identified by PoCUS to have mesenteric adenitis in the setting of nonsurgical abdominal pain presenting to our ED. Most common indications for PoCUS were for suspected appendicitis, suspected intussusception, or undifferentiated abdominal pain. Forty-six percent of patients were male; median age was 9 years (interquartile range, 4-14 years). On 4-week clinical follow-up, 1 patient returned to our ED with a surgical abdomen. Cohen κ values were 0.83 (95% confidence interval, 0.70-0.97) between experienced sonologist-physicians and 0.76 (95% confidence interval, 0.61-.90) between novice and experienced sonologist-physicians. PoCUS can identify mesenteric adenitis, typically a diagnosis of exclusion, in pediatric patients with nonsurgical abdominal pain, both by novice and experienced physician-sonologists. Use of PoCUS may help ED clinicians identify a common cause of nonsurgical abdominal pain in children.

Clinical Characteristics, Outcomes, and Interobserver Agreement of Point-of-Care Ultrasound Detected Mesenteric Adenitis in Nonsurgical Pediatric Abdominal Pain: A Retrospective Cohort Study.

Point-of-care ultrasound (PoCUS) in the emergency department (ED) may facilitate the diagnosis of nonsurgical sources of abdominal pain after surgical causes are excluded. Identifying mesenteric adenitis is a feasible PoCUS application due to its ease of use and speed. However, there are scant data regarding the diagnosis of mesenteric adenitis by PoCUS. The objective of this study was to describe the clinical characteristics, outcomes, and interobserver agreement of mesenteric adenitis identified on PoCUS in pediatric patients with nonsurgical abdominal pain. This was a retrospective review at a tertiary-care, urban pediatric ED. All cases of mesenteric adenitis diagnosed on PoCUS from January 2018 to August 2022 were reviewed. Demographics and clinical data, including relevant outcomes, were recorded. All PoCUS videos were reviewed by a senior sonologist-physician for determination of mesenteric adenitis in children 21 years and younger with nonsurgical abdominal pain. Interobserver agreement by Cohen κ was calculated between experienced and novice physician sonologists blinded to diagnosis, who reviewed 77 six-second video clips for presence or absence of mesenteric adenitis. Thirty-three subjects were identified by PoCUS to have mesenteric adenitis in the setting of nonsurgical abdominal pain presenting to our ED. Most common indications for PoCUS were for suspected appendicitis, suspected intussusception, or undifferentiated abdominal pain. Forty-six percent of patients were male; median age was 9 years (interquartile range, 4-14 years). On 4-week clinical follow-up, 1 patient returned to our ED with a surgical abdomen. Cohen κ values were 0.83 (95% confidence interval, 0.70-0.97) between experienced sonologist-physicians and 0.76 (95% confidence interval, 0.61-.90) between novice and experienced sonologist-physicians. PoCUS can identify mesenteric adenitis, typically a diagnosis of exclusion, in pediatric patients with nonsurgical abdominal pain, both by novice and experienced physician-sonologists. Use of PoCUS may help ED clinicians identify a common cause of nonsurgical abdominal pain in children.

Is palliative care a utopia for older patients with organ failure, dementia or frailty? A qualitative study through the prism of emergency department admission

BackgroundNearly three out of four older people will use the emergency department (ED) during their last year of life. However, most of them do not benefit from palliative care. Providing palliative care is a real challenge for ED clinicians who are trained in acute, life-saving medicine. Our aim is to understand the ED’s role in providing palliative care for this population.MethodsWe designed a qualitative study based on 1) interviews – conducted with older patients (≥ 75 years) with a palliative profile and their informal caregivers – and 2) focus groups – conducted with ED and primary care nurses and physicians. Palliative profiles were defined by the Supportive and Palliative Indicators tool (SPICT). Qualitative data was collected in French-speaking Belgium between July 2021 and July 2022. We used a constant inductive and comparative analysis.ResultsFive older patients with a palliative profile, four informal caregivers, 55 primary and ED caregivers participated in this study. A priori, the participants did not perceive any role for the ED in palliative care. In fact, there is widespread discomfort with caring for older patients and providing palliative care. This is explained by multiple areas of tensions. Palliative care is an approach fraught with pitfalls, i.e.: knowledge and know-how gaps, their implementation depends on patients’(co)morbidity profile and professional values, experiences and type of practice. In ED, there are constant tensions between emergency and palliative care requirements, i.e.: performance, clockwork and needs for standardised procedures versus relational care, time and diversity of palliative care projects. However, even though the ED’s role in palliative care is not recognised at first sight, we highlighted four roles assumed by ED caregivers: 1) Investigator, 2) Objectifier, 3) Palliative care provider, and 4) Decision-maker on the intensity of care. A common perception among participants was that ED caregivers can assist in the early identification of patients with a palliative profile.ConclusionsCurrently, there is widespread discomfort regarding ED caregivers caring for older patients and providing palliative care. Nonetheless, ED caregivers play four roles in palliative care for older patients. In the future, ED caregivers might also perform the role of early identifier.

Developing a Mental Health Screening Questionnaire in an Asian Children's Hospital Emergency Setting

Introduction Mental health concerns among adolescents are increasingly prevalent, yet underrecognized. Adolescents with psychological distress often present to the emergency department (ED) with somatic symptoms. Due to inadequate time for rapport building and lack of familiarity of ED clinicians with psychosocial evaluation, these concerns often get missed. We describe the development and implementation of the Youth Well Being (YWB) questionnaire, a self-administered psychosocial screening tool that aims to overcome the communication barriers to psychosocial evaluation. Methods A multidisciplinary team used a Delphi-like approach to develop the questionnaire, using the home, education, activities/peers, drugs/alcohol, suicidality, emotions/behavior, discharge resources (HEADS-ED) questionnaire as the main reference. Modifications were made based on panel members' clinical experience and adapted to suit local sociocultural context. The YWB questionnaire is administered to adolescents aged 10 to 19 years presenting to the KK Women's and Children's Hospital ED with possible psychosomatic symptoms and behavioral or mental health issues. Positive findings prompt further targeted face-to-face interviews by the clinicians to facilitate appropriate psychosocial referral. Results The 9 domains in the YWB questionnaire explore potential psychosocial difficulties that affect the adolescent's well-being and aim to uncover potential issues that could account for the adolescent's symptoms. We discuss the rationale behind the questions and response options in the YWB questionnaire. Conclusions The YWB questionnaire is the first initiative in Singapore to enable efficient psychosocial screening of at-risk adolescents in the ED. This communication tool can potentially be used in other health care settings to enable early recognition and intervention for adolescents distressed by psychosocial problems.

Risk assessment and the use of personal protective equipment in an emergency department: Differing perspectives of emergency and infection control clinicians. A video-vignette survey

The use of personal protective equipment (PPE) in emergency departments (EDs) is an important defense during infectious disease emergencies. However, what counts as appropriate PPE in EDs is contentious and inconsistently implemented in practice. An online scenario-based video survey was distributed through purposive sampling, and completed by 270 ED and infection prevention and control clinicians in Australia. A descriptive content analysis was performed on the data, and differences between groups were tested using Fisher exact test. Participants agreed that most items were required in both scenarios. Eye protection, mask use, and hand hygiene frequency were more contentious. Physicians were more likely than nurses, and ED clinicians more likely than infection prevention and control clinicians, to regard items or actions as optional rather than essential. Many ED clinicians, particularly physicians, regarded sequences as too time-consuming to be practical in a busy ED. Our findings likely reflect differences in professional roles, competing priorities, and risks, and highlight important contextual characteristics of EDs, such as diagnostic uncertainty, equipment inaccessibility, and resource constraints. To be feasible, practicable, and thereby effective, PPE guidance in the ED must be designed collaboratively with frontline ED staff, and reflects the complexities of their practice.

Review article: Patients who leave before care is completed: What does the legal duty to warn mean for emergency department clinicians?

Patients leave ED for a variety of reasons and at all stages of care. In Australian law, clinicians and health services owe a duty of care to people presenting to the ED for care, even if they have not yet entered a treatment space. There is also a positive duty to warn patients of material risks associated with their condition, proposed treatment(s), reasonable alternative treatment options and the likely effect of their healthcare decisions, including refusing treatment. This extends to a decision to leave the ED before care is completed. The form of that warning may vary based on what is known about the patient's condition and the associated risks at the time. Specific documentation of warnings given is essential.

Pneumonia diagnosis performance in the emergency department: a mixed-methods study about clinicians' experiences and exploration of individual differences and response to diagnostic performance feedback.

We sought to (1) characterize the process of diagnosing pneumonia in an emergency department (ED) and (2) examine clinician reactions to a clinician-facing diagnostic discordance feedback tool. We designed a diagnostic feedback tool, using electronic health record data from ED clinicians' patients to establish concordance or discordance between ED diagnosis, radiology reports, and hospital discharge diagnosis for pneumonia. We conducted semistructured interviews with 11 ED clinicians about pneumonia diagnosis and reactions to the feedback tool. We administered surveys measuring individual differences in mindset beliefs, comfort with feedback, and feedback tool usability. We qualitatively analyzed interview transcripts and descriptively analyzed survey data. Thematic results revealed: (1) the diagnostic process for pneumonia in the ED is characterized by diagnostic uncertainty and may be secondary to goals to treat and dispose the patient; (2) clinician diagnostic self-evaluation is a fragmented, inconsistent process of case review and follow-up that a feedback tool could fill; (3) the feedback tool was described favorably, with task and normative feedback harnessing clinician values of high-quality patient care and personal excellence; and (4) strong reactions to diagnostic feedback varied from implicit trust to profound skepticism about the validity of the concordance metric. Survey results suggested a relationship between clinicians' individual differences in learning and failure beliefs, feedback experience, and usability ratings. Clinicians value feedback on pneumonia diagnoses. Our results highlight the importance of feedback about diagnostic performance and suggest directions for considering individual differences in feedback tool design and implementation.

Let Us Keep the Emergency Department Safe: National Disability Insurance Scheme Supports Workers’ Experiences When Supporting People with Psychosocial Disability in the Emergency Department

Internationally, community healthcare is provided by personal health budgets or personalised schemes as part of a consumer-directed care emphasis. Although the introduction of a personalised scheme in Australia has improved the lives of many, people with psychosocial disability (PSD) and their families have experienced many challenges with service integration. Six focus groups with 17 National Disability Insurance Scheme (NDIS) support workers were conducted between June and November 2022. Participants were asked semistructured questions about their engagement with emergency department (ED) clinicians when presenting with a person with PSD and NDIS plan. They were also asked how they contribute to service integration to improve communication practices between the two services. A qualitative, descriptive thematic analysis approach was used. Results indicate there is a need for improved service integration pathways between the ED and NDIS services. The thematic analyses identified five main themes: the ED is harmful and not safe, but nowhere else to go; make the ED safe; further training for ED staff; improving communication across services; and alternatives to the ED. This study reports NDIS support worker’s experience when supporting a person they are caring for in the ED. NDIS support workers have an understanding of the needs of the people they care for;their wisdom can contribute greatly to ED clinical responses in crisis care. Recommendations are offered to assist with service integration between the two systems.

Evaluation of a Digital Scribe: Conversation Summarization for Emergency Department Consultation Calls.

We present a proof-of-concept digital scribe system as an Emergency Department (ED) consultation call-based clinical conversation summarization pipeline to support clinical documentation, and report its performance. We use four pre-trained large language models to establish the digital scribe system: T5-small, T5-base, PEGASUS-PubMed, and BART-Large-CNN via zero-shot and fine-tuning approaches. Our dataset includes 100 referral conversations among ED clinicians and medical records. We report the ROUGE-1, ROUGE-2, and ROUGE-L to compare model performance. In addition, we annotated transcriptions to assess the quality of generated summaries. The fine-tuned BART-Large-CNN model demonstrates greater performance in summarization tasks with the highest ROUGE scores (F1ROUGE-1=0.49, F1ROUGE-2=0.23, F1ROUGE-L=0.35) scores. In contrast, PEGASUS-PubMed lags notably (F1ROUGE-1=0.28, F1ROUGE-2=0.11, F1ROUGE-L=0.22). BART-Large-CNN's performance decreases by more than 50% with the zero-shot approach. Annotations show that BART-Large-CNN performs 71.4% recall in identifying key information and a 67.7% accuracy rate. The BART-Large-CNN model demonstrates a high level of understanding of clinical dialogue structure, indicated by its performance with and without fine-tuning. Despite some instances of high recall, there is variability in the model's performance, particularly in achieving consistent correctness, suggesting room for refinement. The model's recall ability varies across different information categories. The study provides evidence towards the potential of AI-assisted tools in assisting clinical documentation. Future work is suggested on expanding the research scope with additional language models and hybrid approaches, and comparative analysis to measure documentation burden and human factors.

Impact of a Mental Health Screening Process in a Pediatric Emergency Department.

The aim of this study was to describe how specific mental health-trained social workers can assist in the evaluations and follow-up of patients presenting with mental health concerns in the pediatric emergency department (ED). Work was performed at a quaternary children's hospital ED with 95,000 annual ED visits across 2 locations. Patients requiring mental health services identified based on presenting complaint or from universal suicide screen were included. Emergency department team first evaluates the patients for medical screening and then consults a team of social workers specialized in acute mental health screening (AMHS). The team evaluates and provides recommendation for disposition and assists in plan completion. For patients not admitted, AMHS team makes 24- and 48-hour calls to ensure safety. We collected and analyzed the data on all eligible patients from September 2015 through June 2019 for (1) demographic information, (2) trends in number of consults to AMHS, (3) disposition plans and trends by year, and (4) frequency of follow-up phone calls. A total of 5950 patient visits were reviewed, for 4454 distinct patients. Most patients were 12 to 17 years of age, female, and White, with Medicaid being the predominant insurance. The most common chief complaint was suicidal ideation/plan/attempt. Self-referrals were the majority of assessments, and 59% of patients were already receiving mental health services. Median team response time was 19 minutes. There was an upward trend in consults. Psychiatric hospitalization was the most common disposition; more than 95% of the other visits had timely follow-up phone calls. Despite an increasing number of patients presenting to the ED with mental health crisis, safe and efficient management is possible with ED staff-social worker partnership. This approach can ensure that eligible patients receive consistent and evidence-based evaluations and can allow ED clinicians to respond to medical emergencies that require their attention.

Payment Innovation in Emergency Care: A Case for Global Clinician Budgets

The fee-for-service funding model for U.S. emergency department (ED) clinician groups is increasingly fragile. Traditional FFS payment systems offer no financial incentives to improve quality, address population health, or make value-based clinical decisions. FFS also does not support maintaining ED capacity to handle peak demand periods. In FFS, clinicians rely heavily on cross subsidization where high reimbursement from commercial payors offsets low reimbursement from government payors and the uninsured. While FFS survived decades of steady cuts in government reimbursement rates, it is increasingly strained due to visit volatility and the effects of the No Surprises Act, which is driving down commercial reimbursement. Financial pressures on ED clinician groups, combined with higher hospital boarding and clinical workloads are increasing workforce attrition. Here we propose an alternative model to address some of these fundamental issues: an all-payer-funded, voluntary global budget for ED clinician services. If designed and implemented effectively, the model could effectively support robust clinician staffing over the long-term, ensure stability in clinical workload, and potentially improve the equity in payments. The model could also be combined with population health programs (e.g., pre- and post-ED telehealth, frequent ED use programs, and other innovations) which offer significant payer returns and address both quality and value. A linked program could also change hospital incentives that contribute to boarding. Strategies exist to test and refine ED clinician global budgets through existing government programs in Maryland and potentially through state-level legislation, as a precursor to broader adoption.

Innovation adoption, use and implementation in emergency departments during the COVID-19 pandemic

IntroductionDuring a public health emergency like the COVID-19 pandemic, emergency department (ED) clinicians may be able to save more lives if they rapidly identify and implement innovations that are safe...

Project to improve the management of the head injury patients presenting to the emergency department

IntroductionAt Sandwell General Hospital, there was no risk stratification tool or pathway for head injury (HI) patients presenting to the emergency department (ED). This resulted in significant delays in the...

The epidemiology of dying within 48hours of presentation to emergency departments: a retrospective cohort study of older people across Australia and New Zealand.

Emergency department (ED) clinicians are more frequently providing care, including end-of-life care, to older people. To estimate the need for ED end-of-life care for people aged ≥65years, describe characteristics of those dying within 48 hours of ED presentation and compare those dying in ED with those dying elsewhere. We conducted a retrospective cohort study analysing data from 177 hospitals in Australia and New Zealand. Data on older people presenting to ED from January to December 2018, and those who died within 48hours of ED presentation, were analysed using simple descriptive statistics and univariate logistic regression. From participating hospitals in Australia or New Zealand, 10,921 deaths in older people occurred. The 48-hour mortality rate was 6.43 per 1,000 ED presentations (95% confidence interval: 6.31-6.56). Just over a quarter (n = 3,067, 28.1%) died in ED. About one-quarter of the cohort (n = 2,887, 26.4%) was triaged into less urgent triage categories. Factors with an increased risk of dying in ED included age 65-74years, ambulance arrival, most urgent triage categories, principal diagnosis of circulatory system disorder, and not identifying as an Aboriginal or Torres Strait Islander person. Of the 7,677 older people admitted, half (n = 3,836, 50.0%) had an encounter for palliative care prior to, or during, this presentation. Our findings provide insight into the challenges of recognising the dying older patient and differentiating those appropriate for end-of-life care. We support recommendations for national advanced care planning registers and suggest a review of triage systems with an older person-focused lens.

Multisite development and validation of machine learning models to predict severe outcomes and guide decision-making for emergency department patients with influenza.

Millions of Americans are infected by influenza annually. A minority seek care in the emergency department (ED) and, of those, only a limited number experience severe disease or death. ED clinicians must distinguish those at risk for deterioration from those who can be safely discharged. We developed random forest machine learning (ML) models to estimate needs for critical care within 24h and inpatient care within 72h in ED patients with influenza. Predictor data were limited to those recorded prior to ED disposition decision: demographics, ED complaint, medical problems, vital signs, supplemental oxygen use, and laboratory results. Our study population was comprised of adults diagnosed with influenza at one of five EDs in our university health system between January 1, 2017 and May 18, 2022; visits were divided into two cohorts to facilitate model development and validation. Prediction performance was assessed by the area under the receiver operating characteristic curve (AUC) and the Brier score. Among8032 patientswith laboratory-confirmed influenza,incidence of critical care needs was 6.3% and incidence of inpatient care needswas 19.6%.Themost common reasons for ED visitwere symptoms of respiratory tract infection,fever,and shortness of breath. Model AUCswere0.89 (95% CI 0.86-0.93)for prediction of criticalcare and0.90(95% CI0.88-0.93)for inpatient care needs;Brier scores were0.026 and0.042, respectively.Importantpredictorsincludedshortness of breath,increasingrespiratory rate,and ahigh number of comorbid diseases. ML methods can be used to accurately predict clinical deterioration in ED patients with influenza and have potential to support ED disposition decision-making.