Direct Hospital Admission Research Articles

Real-world approaches to outpatient treatment of status migrainosus: A survey study.

Identify how the American Headache Society (AHS) membership manages status migrainosus (SM) among outpatients. SM is defined as a debilitating migraine attack lasting more than 72 h. There is no standard of care for SM, including whether a 72-h duration is required before the attack can be treated as SM. The Refractory Headache Special Interest Group from AHS developed a four-question survey distributed to AHS members enquiring (1) whether they treat severe refractory migraine attacks the same as SM regardless of duration, (2) what their first step in SM management is, (3) what the top three medications they use for SM are, and (4) whether they are United Council for Neurologic Subspecialties (UCNS) certified. The survey was conducted in January 2022. Descriptive statistical analyses were performed. Responses were received from 196 of 1859 (10.5%) AHS members; 64.3% were UCNS certified in headache management. Respondents treated 69.4% (136/196) of patients with a severe refractory migraine attack as SM before the 72-h period had elapsed. Most (76.0%, 149/196) chose "treat remotely using outpatient medications at home" as the first step, 11.2% (22/196) preferred procedures, 6.1% (12/196) favored an infusion center, 6.1% (12/196) sent patients to the emergency department (ED) or urgent care, and 0.5% (1/196) preferred direct hospital admission. The top five preferred medications were as follows: (1) corticosteroids (71.4%, 140/196), (2) nonsteroidal anti-inflammatory drugs (NSAIDs) (50.1%, 99/196), (3) neuroleptics (46.9%, 92/196), (4) triptans (30.6%, 60/196), and (5) dihydroergotamine (DHE) (21.4%, 42/196). Healthcare professionals with expertise in headache medicine typically treated severe migraine attacks early and did not wait 72 h to fulfill the diagnostic criteria for SM. Outpatient management with one or more medications for home use was preferred by most respondents; few opted for ED referrals. Finally, corticosteroids, NSAIDs, neuroleptics, triptans, and DHE were the top five preferred treatments for home SM management.

Long-term prescribed opioid use after hospitalization or emergency department presentation among opioid naïve adults (2014-2020)-A population-based descriptive cohort study.

The aim of this work is to describe opioid initiation and long-term use after emergency department (ED) visits or hospitalizations in New South Wales, Australia, by patient, admission and clinical characteristics. This is a population-based cohort study, including all hospitalizations and ED visits between 2014 and 2020, linked to medicine dispensings, deaths and cancer registrations (Medicines Intelligence Data Platform), among adults with no opioid dispensings in the previous year. Outcome measures were opioid initiations (dispensed within 7days of discharge) and long-term use (90 days of continuous exposure, 90-270 days after initiation). The cohort included 16 153 096 admissions by 4.2 million opioid-naïve adults; 39.0% were ED presentations without hospital admission, 16.8% hospital admissions via ED and 44.2% direct hospital admissions. Opioids were initiated post-discharge for 6.2% of ED, 8.3% of hospital via ED and 10.0% of direct hospital admissions; of these 1.0%, 2.5% and 0.5% progressed to long-term opioid use, respectively. Initiation was lowest in obstetric admissions without surgery (1.0%), and highest among trauma admissions (25.4%), obstetric admissions with surgical intervention (19.8%) and non-trauma surgical admissions (12.0%). Long-term use was highest among medical admissions via ED (3.5%), trauma admissions (2.3%) and ED alone (1.0%). From 2014 to 2020, overall opioid initiations decreased 16% from 8.7% to 7.2%, and long-term opioid use decreased 33% from 1.3% to 0.8%. Both opioid initiation and long-term use decreased over time; however, the higher rates of long-term use following trauma, and medical admissions via ED, warrant further surveillance. Strategies supporting appropriate prescribing and access to multidisciplinary pain services will facilitate best practice care.

Clinical guideline highlights for the hospitalist: Pediatric direct hospital admission.

Direct Admission to the Hospital for Children in the United States RELEASE DATE: March 3, 2023 PRIOR VERSION(S): n/a DEVELOPER: American Academy of Pediatrics, Committee on Hospital Care FUNDING SOURCE: American Academy of Pediatrics TARGET POPULATION: Children who are potential candidates for direct hospital admission.

Difficult Intravenous Access Requiring Ultrasound in the Emergency Department: Associations With Delays in Care and Areas for Quality Improvement.

Background Patients with difficult intravenous access (DIVA) requiring ultrasound-guided intravenous (USGIV) access have been associated with delays in treatment, imaging, and disposition in academic emergency department (ED) patient populations. Our objective in this study was to characterize differences in time to intravenous access, imaging, and disposition between patients with DIVA versus those without DIVA requiring USGIV access in a community ED while also assessing for DIVA-associated comorbidities. Methods A cross-sectional, observational analysis was performed on admitted ED patients evaluated from September 2to September 31, 2022, at a community ED. Patients with DIVA were defined as patients with two failed attempts at traditional intravenous placement. These patients require USGIV placement per institutional protocol. Patients younger than 18 years of age, trauma admissions, repeated visits from the same patient, patients with missing data, and direct hospital admissions were excluded. Continuous variables were recorded with medians and included ED throughput measures of time to vascular access, contrast CT imaging, and disposition. Differences in median times between DIVA patients versus non-DIVA patients were assessed with the Mann-Whitney U-test. Categorical data involving comorbidities were reported as percentages, and differences in proportions between DIVA versus non-DIVA patients were assessed via chi-square tests. Multivariate logistic regression analysis evaluated for correlations between DIVA and times to access, contrast CT imaging, disposition, and significant covariates while adjusting for demographic information. Results A total of 1250 patients were included in this investigation (5.8% associated with DIVA requiring USGIV access). The median age of all subjects was 69 (interquartile range = 58, 79) with no significant difference between the DIVA and non-DIVA groups. Patients with DIVA were more likely to be female in comparison to patients without DIVA (65.3% and 51.2%, respectively, p < 0.05). Patients with a history of end-stage renal disease (ESRD) (p < 0.001), intravenous drug use (IVDU) (p < 0.001), and venous thromboembolism (p < 0.05) had statistically significant associations with DIVA. On regression analysis, patients with DIVA were more likely to have a history of ESRD with an odds ratio (OR) of 3.56 (95% confidence interval (CI): 1.62-7.81) and a history of IVDU with an OR of 14.29 (95% CI: 5.17-39.54). Patients with DIVA were associated with statistically significant greater median times to vascular access, contrast CT imaging, and disposition (p < 0.001 for time to access and disposition and p < 0.01 for time to contrast CT imaging). Conclusion In this study, DIVA cases requiring USGIV access were positively associated with significantly longer times to access, contrast CT imaging, and disposition compared to patients without DIVA at our community ED. Comorbidities such as IVDU and ESRD had statistically significant associations with DIVA requiring USGIV access.

Providing a Clear Path to the Direct Admission Process.

The aim of this study was to increase patient safety during care transitions through the development of a new process and electronic screening tool. Direct hospital admissions that are not clinically triaged can put the patient at an increased safety risk. Utilizing the electronic medical record and mnemonic situation, background, assessment, and recommendation (SBAR), an admission module was created to intake direct admission requests by transfer center nurses to ensure clinical triage and accurate bed placement. One hundred eighty-nine direct admissions met inclusion criteria. Thirteen patients were clinically screened, deemed not stable for the acute care setting, and sent to the emergency department. One direct admission safety event occurred involving a hypoglycemic patient upon arrival. Results indicate the new clinical screening program offered a safer way to directly admit patients to the inpatient setting. Implementation of the electronic SBAR handoff tool meets the Joint Commission standard of care transitions compliance. The significant workflow changes increased patient safety and will be expanded in the future to all service lines.

Using Virtual Care to Facilitate Direct Hospital Admissions in Outpatients with Worsening COVID-19 Infection.

Recognizing emergency department overcrowding during the COVID-19 pandemic, a pathway to facilitate direct admissions for outpatients with worsening COVID-19 infection was created using the COVID-19 expansion to outpatients (COVIDEO) virtual care program. Outpatients appropriate for direct admission had oxygen saturations consistently <92% without severe respiratory distress. Pulse oximeters were proactively delivered to high-risk patients, and patients contacted the program in the event of worsening symptoms or desaturation persistently <92%. Over a 15-month period, 9,116 outpatients were managed by the program, 164 of whom were hospitalized, and 83 of those hospitalized (50.6%) were directly admitted through this pathway. Of those directly admitted, 10 (12.0%) patients required ICU admission, occurring a median of 4 days from hospital admission. The mortality rate among directly admitted patients was 3.6% (3/83). Implementation of a virtual care program to facilitate direct admissions in outpatients with COVID-19 created a safe, efficient, and patient-centered pathway of care.

Home management of COVID-19 symptomatic patients: a safety study on COVID committed home medical teams.

To reduce the overburden in the hospital, during the COVID-19 pandemic, some "COVID Committed Home Medical Teams" (CCHTs) were created in Italy. These units consist of a small pool of general practitioners who aim to evaluate all patients with COVID-19 who require a medical examination directly at home. After the first visit (which can end with patient hospitalisation or home management), CCHTs periodically monitor the patients' clinical conditions and vital signs (usually a revaluation every 24-48 hours, except for a sudden worsening). However, this strategy - which reduces the pressure on hospitals - has never been evaluated for patient safety. Our study aims to determine whether a home-based monitoring and treatment strategy for non-severe COVID-19 patients was safe as direct hospital admission by the emergency department. We conducted a retrospective observational study about 1,182 patients admitted to the hospital for COVID-19 between September 2020 and April 2021, confronting in-hospital and 30-day mortality in both CCHT-referred (n=275) and directly admitted by emergency department (n=907). Patients assessed by the CCHT had lower in-hospital and 30-day mortality (18% vs 28%, p=0.001; and 20% vs 30%, p=0.002); but, in the propensity score matching comparison, there was no characteristic between the two groups turned out significantly different. CCHT did not correlate with in-hospital or 30-day mortality. CCHT is a safe strategy to reduce hospital overburden for COVID-19 during pandemic surges.

Coordinación entre atención primaria y un servicio de geriatría, 20 años después

ObjectiveTo describe the evolution of the care and coordination activity of a specialized hospital team (Geriatric Home Care-AGD team) in support of Primary Care (PC) for the care of geriatric patients. LocationHealth-5 area of the Community of Madrid. Participants, interventions and main measurementsEvaluation of patients attended at home at the request of PC by AGD, in 1997-1999 and twenty years later, in 2017-2019. Sociodemographic, clinical, functional and mental variables were recorded. Also care data such as time to first visit, average length of stay, origin and destination of patients, reason for consultation and main reason for disability. Results524 patients (58% of the total) requested from AP in 1997-1999 and 1196 (72.2% of the total) in 2017-2019 were attended. Currently we show an older patient, more physically and mentally disabled, with a higher prevalence of geriatric syndromes and greater social fragility. Dementia has increased as the main reason for incapacity, with an increase in Comprehensive Geriatric Assessment and clinical control as the main reasons for referral. Conclusionsin our setting, after 20 years, AGD continues to be an important support for PC in the complex approach to frail elderly patients confined to their homes, who are increasingly vulnerable from a clinical, functional, cognitive and social point of view. It contributes to the management of geriatric syndromes, deprescription, direct hospital admissions when needed without having to go to the emergency department, and a decrease in institutionalization, facilitating the maintenance of the patient at home.

Designing and describing an electronic referral system to facilitate direct hospital admissions

BackgroundIn France, the progressive use of emergency departments (EDs) by primary care providers (PCPs) as a point of access to hospitalization for nonurgent patients is one of the many causes of their overcrowding. To increase the proportion of direct hospital admissions, it is necessary to improve coordination between PCPs and hospital specialists. The objective of our work was to describe the design and implementation of an electronic referral system aimed at facilitating direct hospital admissions.MethodsThis initiative was conducted in a French area (Hauts-de-Seine Sud) through a partnership between the Antoine-Béclère University Hospital, the Paris-Saclay University Department of General Medicine and the local health care network. The implementation was carried out in 3 stages, namely, conducting a survey of PCPs in the territory about their communication methods with the hospital, designing and implementing a web-based application called “SIPILINK” (Système d'Information de la Plateforme d’Intermédiation Link) and an innovative organization for hospital management of the requests, and analysing through descriptive statistics the platform use 9 months after launch.ResultsThe e-referral platform was launched in November 2019. First, a PCP filled out an electronic form describing the reason for his or her request. Then, a hospital specialist worked to respond within 72 h. Nine months after the launch, 132 PCPs had registered for the SIPILINK platform, which represented 36.6% of PCPs in this area. Of the 124 requests made, 46.8% corresponded to a hospitalization request (conventional or day hospitalization). The most requested specialty was internal medicine (48.4% of requests). The median time to first response was 43 min, and 43.5% of these requests resulted in direct admission (conventional or day hospitalization).ConclusionsThis type of system responds to a need for coordination in the primary-secondary care direction, which is less often addressed than in the secondary-primary care direction. The first results show the potential of the system to facilitate direct admissions within a short time frame. To make the system sustainable, the next step is to extend its use to other hospitals in the territory.

Hospitalization Trends for Acute Appendicitis in Spain, 1998 to 2017.

The incidence of acute appendicitis decreased in Western countries from 1930 to at least the early 1990s, when epidemiological data started becoming scarcer. This study aimed to assess the trend in annual hospitalizations for acute appendicitis in all people Spain for a 20-year period between 1998 and 2017. This observational study analyzing direct age-standardized hospital admission rates by gender and age group (0–14 years, 15–34 years, 35–44 years, 45–64 years, and ≥65 years). Joinpoint regression models were fitted to evaluate changes in trends. There were 789,533 emergency hospital admissions for acute appendicitis between 1998 and 2017: 58.9% in boys and men and 41.1% in girls and women. Overall, there was a significant increase in admissions for this cause from 1998 to 2009, with an annual percent change (APC) of 0.6%. Following the peak in 2009, admission rates decreased by around 1.0% annually until 2017. The length of hospital stay gradually decreased from 4.5 days in 1998 to 3.4 days in 2017. The trends in hospital admissions for acute appendicitis in Spain changed over the study period, decreasing from 2009, especially in people younger than 35 years.

Machine learning-based approach to the risk assessment of potentially preventable outpatient cancer treatment-related emergency care and hospitalizations.

333 Background: Patients undergoing outpatient infusion chemotherapy for cancer are at risk for potentially preventable, unplanned acute care in the form of emergency department (ED) visits and hospital admissions. This can impact outcomes, patient decisions, and costs to the patient and healthcare system. To address this need, the Centers for Medicare &amp; Medicaid Services developed the Chemotherapy Measure (OP-35). Recent randomized controlled data indicate that electronic health record (EHR)-based machine learning (ML) approaches accurately direct supportive care to reduce acute care during radiotherapy. As this may extend to systemic therapy, this study aims to develop and evaluate ML approaches to predict the risk of OP-35 qualifying, potentially preventable acute care within 30 days of infusional systemic therapy. Methods: This study included data from UCSF cancer patients receiving infusional chemotherapy from July 1, 2017, to February 11, 2021, (total 7,068 patients over 84,174 treatments). The data incorporated into the ML included 430 EHR-derived variables, including cancer diagnosis, therapeutic agents, laboratory values, vital signs, medications, and encounter history. Three ML approaches were trained to predict an OP-35 acute care risk following a systemic therapy infusion with least absolute shrinkage selection operator (LASSO), random forest, and gradient boosted trees (GBT; XGBoost) approaches. The models were trained on a subset (75% of patients; before October 12, 2019) of the dataset and validated on a mutually exclusive subset (25% patients; after October 12, 2019) based on the receiver operating characteristic (ROC) curves and calibration plots. Results: There were 1,651 total acute care visits (244 ED visits and 1,407 ED visits converted into hospitalization); 1,310 infusions included a qualifying acute care visit (200 with ED visits only, 0 direct hospital admissions, and 1,110 with both ED visit and hospitalization). Each ML approach demonstrated good performance in the internal validation cohort, with GBT (AUC 0.805) outpacing the random forest (0.750) and LASSO logistic regression (0.755) approaches. Visualization of calibration plots verified concordance between predicted and observed rates of acute care. All three models shared patient age and days elapsed since last treatment as important contributors. Conclusions: EHR-based ML approaches demonstrate high predictive ability for OP-35 qualifying acute care rates on a per-infusion basis, identifying 30-day potentially preventable acute care risk for patients undergoing chemotherapy. Prospective validation of these models is ongoing. Early prediction can facilitate interventional strategies which may reduce acute care, improve health outcomes, and reduce costs.

Orthopedic fracture hospitalizations are revving up from E-Scooter related injuries

IntroductionDue to market expansion of electric-scooter companies, a significant rise of personal e-scooter use in dense, urban communities has been observed. No literature has specifically focused on e-scooter fracture epidemiology and risk factors associated with direct hospital admission. The aims of this study were to evaluate the 1) patterns of e-scooter related orthopaedic fractures 2) risk factors associated with direct hospital admission. Materials and methodsA retrospective review of National Electronic Injury Surveillance System (NEISS) from the United States between 2015 and 2019 was utilized to identify e-scooter fracture epidemiology. Uni/multivariable analyses were conducted to identify independent variables associated with direct hospital admission. Results5,016 patients were identified. The most common fracture location was the upper extremity (25.4%). Multiple distinct fractures diagnoses (p < 0.001), fracture of the upper arm (p = 0.01), metacarpal (p = 0.03), skull(p < 0.001), and associated internal organ injury (p = 0.02) all had a statistical increase over time. Fracture of the upper leg (OR 58.31), lower trunk (OR: 47.04), and associated internal organ damage (OR: 37.82) had the greatest association with direct hospital admission. DiscussionThis study highlights that e-scooter fracture related injuries continue to progress, and without appropriate educational and public health efforts, these injuries will continue to rise.

Cardiovascular health within the UK prison population

In England and Wales, there were around 83 000 people in prison at any one time in 2019. People entering prison are more likely than the general population to have...

Acute care utilization in pediatric sickle cell disease and sickle cell trait in the USA: prevalence, temporal trends, and cost.

The objective of this study was to analyze acute care utilization of sickle cell disease (SCD) and sickle cell trait (SCT) in children and identify trends in emergency department (ED) visits and inpatient admissions over a 10-year period. This is a retrospective population-based study of SCD- and SCT-related ED visits and admissions from 2006 to 2015. Data were acquired from the Healthcare Cost and Utilization Project (HCUP), National Inpatient Sample (NIS), and National Emergency Department Sample (NEDS) database. Cost-to-charge and estimated professional fee ratios were applied to approximate costs. Over 80% of medical expenditure on HbSS is through ED-based admissions. There is a statistically significant increase from 2006 to 2015 in the direct hospital admissions associated with patients less than 18years of age who have been diagnosed with SCT.Conclusion: Among patients less than 18years of age with HbSS, inpatient admissions through the emergency department accounted for the largest medical expenditure of the SCD subtypes. What is Known: • There are currently no multi-year, nationwide analyses of acute care utilization in sickle cell disease and sickle cell trait (SCT) in the pediatric population. • SCT is more common than SCD, affecting 1.5% of all infants born in the USA. What is New: • Comprehensive annual costs of acute care utilization of patients less than 18years of age with SCD and SCT in the USA which includes aggregated demographical patient care data and to illustrate temporal trends of acute care utilization in children less than 18years of age with SCD and SCT • Among patients less than 18years of age with HbSS, inpatient admissions through the emergency department accounted for the largest medical expenditure of the sickle cell disease subtypes.

Serious sports-related injury in England and Wales from 2012-2017: a study protocol

BackgroundPhysical activity is an important component of healthy lifestyles, with a central role in morbidity prevention. However, sporting and physical activity also involve an inherent injury risk. Some sports and activities have a higher injury risk, and may involve more severe injuries. Furthermore, injuries of a severe nature have substantial individual and societal consequences, including the burden of assessment, treatment, and potential on-going care costs. There are limited data on severe sports injury risk in England and Wales, and no national data describing risk across sports. The aims of this study are to identify the cases and incidence of: i) paediatric and ii) adult severe sports injury from 2012 to 2017; and to describe injury incidence in individual sports.MethodsThis study is an analysis of prospectively collected sport-related injuries, treated from January 2012 to December 2017. Incidents involving a severe injury (in-patient trauma care) in England and Wales, will be identified from the Trauma Audit Research Network registry. Data for patients who were: transfers or direct hospital admissions, with inpatient stays of ≥3 days, admissions to High Dependency areas, or in-hospital mortality after admission; and whose injury mechanism was sport, or incident description included one of 62 sporting activities, will be extracted. Data will be categorised by sport, and sports participation data will be derived from Sport England participation surveys. Descriptive statistics will be estimated for all demographic, incident, treatment and sport fields, and crude serious annual injury incidence proportions estimated. Poisson confidence intervals will be estimated for each sport and used to describe injury risk (incidence) across sporting activities.DiscussionThis study will be the first to describe the number of, and trends in severe sport-related injuries in England and Wales. These data are useful to monitor the number and burden of severe sports injury, and inform injury prevention efforts. The monitoring and mitigation of sports injury risk is essential for individuals, health services and policy, and to encourage physically active lifestyles and safer participation for adults and children.

Management of Heart Failure in Left Ventricular Assist Device (LVAD) Patients Utilizing an Outpatient Diuresis Clinic

Left ventricular assist device (LVAD) implantation and survival is increasing. With patients living longer on device support, LVAD-related healthcare costs and hospitalizations, including those for volume overload, are on the rise as well. The Johns Hopkins Heart Failure Bridge Clinic (HFBC), a HF-trained Nurse Practitioner (NP) led clinic with ability to administer outpatient intravenous (IV) diuretics, has resulted in significant reductions in HF readmission rates. We developed a protocol for LVAD patient assessment and management via the HFBC, and describe our initial experience regarding the role and efficacy of this unique program in outpatient management of HF in LVAD patients. LVAD patients were evaluated by an LVAD provider and subsequently referred to the diuresis clinic for volume overload thought not to be related to an LVAD issue. If seen in the diuresis clinic, patients then had repeat evaluation with the LVAD/HF team within one week. Baseline characteristics and HF outcomes were assessed. Nineteen LVAD patients were seen in the diuresis clinic a total of 65 times. IV furosemide was administered 31 times to 13 patients (mean dose 80 ± 23.5 mg). Urine output during first in clinic void after IV diuretic was 125 ± 53 mL. At 30 of 65 (46%) visits, oral diuretic dose was uptitrated. Potassium was given at 18 of the 65 (27.7%) visits. There were no significant changes in blood pressure, renal function, or electrolytes post diuresis. Only 1 of 65 patient visits resulted in direct hospital admission (hyponatremia). All other visits discharged to home in improved/stable condition and avoidance of hospitalization for heart failure. During follow up, only 6 of 19 patients were hospitalized for a primary diagnosis of volume overload. We describe successful implementation of an outpatient IV diuresis protocol for treatment of HF in LVAD patients. The strategy has major implications in reducing LVAD-related HF hospitalizations and healthcare utilization.

A Nurse Led Triage System in the Ambulatory Cellular Immunotherapy Setting

Topic Significance & Study Purpose/Background/Rationale Delivery of cellular immunotherapy (IMTX) in the ambulatory setting continues to evolve in an academic Comprehensive Cancer Center. A dedicated IMTX outpatient department (OPD) was opened in 2016 (patients treated=427), providing both FDA approved CAR T cell and phase l/ll immune effector cell treatments. Provision of care to acutely ill patients, ensuring early recognition of treatment related side effects, and triggering a prompt hospitalization when indicated, requires an intricate triage system. To address these complexities of care and support the institutional mission of state of the science therapy, a comprehensive triage infrastructure was developed. Methods, Intervention, & Analysis As the first line of contact, Nurses are instrumental in early assessment and navigation of the triage system to ensure management of care in the appropriate setting. The triage infrastructure is reliant on patient and caregiver education and the understanding of the structure which includes: symptom-based algorithms, 24-hour access to dedicated cellular IMTX professionals, triage in the ambulatory setting, access to after-hours in-patient triage, capacity for direct hospital admission, and collaboration between in-patient and out-patient attending Physicians. Education for patient/caregiver is comprised of fever and symptom management algorithms, a thermometer, and institutional contact information. To ensure safe patient outcomes, patients are advised to reside within 30 minutes of treatment facility and have a reliable caregiver. Findings & Interpretation The innovative and complex practice of cellular immunotherapy delivered in the OPD necessitates patient and caregiver understanding of the availability of providers and assessment intervention for effective symptom management. Continued reinforcement of education at crucial time points helps to safeguard against complications in care. Bridging communication gaps between in-patient and ambulatory settings are essential for effective patient hand-offs. Discussion & Implications A dynamic triage system is essential to monitor, assess, and deliver prompt patient interventions. Nurse triage skills are vital in early detection of Cytokine Release Syndrome (CRS), neurotoxicity, and responding to these potentially fatal complications. A highly developed and refined triage system in the IMTX OPD setting is crucial for patients and caregivers to achieve best outcomes.

ED RAPID: A Novel Children's Hospital Direct Admission Process Utilizing the Emergency Department.

Direct hospital admission of children without evaluation in the emergency department (ED) is common, but few guidelines exist to maximize safety by assessing patient stability. This report describes a novel approach to support patient safety. An interdisciplinary children's hospital team developed a brief ED-based evaluation process called the ED Rapid Assessment of Patients Intended for Inpatient Disposition (ED RAPID). It entails a brief evaluation of vital signs and clinical stability by the ED attending physician and nurse. Children deemed stable are admitted to inpatient wards, whereas those requiring immediate intervention undergo full ED evaluation and disposition. We assessed outcomes for all children evaluated through this process from March 2013 through February 2015. During the study period, we identified 715 patients undergoing ED RAPID evaluation. Of these, we directly admitted 691 (96.4%) to the hospital ward after ED RAPID evaluation; median ED treatment time was 4.0 minutes. We transitioned 24 (3.4%) to full ED evaluation, 14 (2.0%) because a ward bed was unavailable, and 10 (1.4%) for clinical reasons identified in the evaluation. We admitted four of the 10 stopped (40% of stops, 0.6% of total) to an intensive care unit, and 6 (60% of stops, 0.8% of total) to the hospital ward after ED care. Eight children (1.1%) admitted to the hospital ward after ED RAPID evaluation required a transfer to an intensive care unit within 12 hours. The ED RAPID evaluation process for children directly admitted to the hospital was feasible and effective in this setting.

Emergency department triage prediction of clinical outcomes using machine learning models

BackgroundDevelopment of emergency department (ED) triage systems that accurately differentiate and prioritize critically ill from stable patients remains challenging. We used machine learning models to predict clinical outcomes, and then compared their performance with that of a conventional approach—the Emergency Severity Index (ESI).MethodsUsing National Hospital and Ambulatory Medical Care Survey (NHAMCS) ED data, from 2007 through 2015, we identified all adult patients (aged ≥ 18 years). In the randomly sampled training set (70%), using routinely available triage data as predictors (e.g., demographics, triage vital signs, chief complaints, comorbidities), we developed four machine learning models: Lasso regression, random forest, gradient boosted decision tree, and deep neural network. As the reference model, we constructed a logistic regression model using the five-level ESI data. The clinical outcomes were critical care (admission to intensive care unit or in-hospital death) and hospitalization (direct hospital admission or transfer). In the test set (the remaining 30%), we measured the predictive performance, including area under the receiver-operating-characteristics curve (AUC) and net benefit (decision curves) for each model.ResultsOf 135,470 eligible ED visits, 2.1% had critical care outcome and 16.2% had hospitalization outcome. In the critical care outcome prediction, all four machine learning models outperformed the reference model (e.g., AUC, 0.86 [95%CI 0.85–0.87] in the deep neural network vs 0.74 [95%CI 0.72–0.75] in the reference model), with less under-triaged patients in ESI triage levels 3 to 5 (urgent to non-urgent). Likewise, in the hospitalization outcome prediction, all machine learning models outperformed the reference model (e.g., AUC, 0.82 [95%CI 0.82–0.83] in the deep neural network vs 0.69 [95%CI 0.68–0.69] in the reference model) with less over-triages in ESI triage levels 1 to 3 (immediate to urgent). In the decision curve analysis, all machine learning models consistently achieved a greater net benefit—a larger number of appropriate triages considering a trade-off with over-triages—across the range of clinical thresholds.ConclusionsCompared to the conventional approach, the machine learning models demonstrated a superior performance to predict critical care and hospitalization outcomes. The application of modern machine learning models may enhance clinicians’ triage decision making, thereby achieving better clinical care and optimal resource utilization.

Machine Learning–Based Prediction of Clinical Outcomes for Children During Emergency Department Triage

While machine learning approaches may enhance prediction ability, little is known about their utility in emergency department (ED) triage. To examine the performance of machine learning approaches to predict clinical outcomes and disposition in children in the ED and to compare their performance with conventional triage approaches. Prognostic study of ED data from the National Hospital Ambulatory Medical Care Survey from January 1, 2007, through December 31, 2015. A nationally representative sample of 52 037 children aged 18 years or younger who presented to the ED were included. Data analysis was performed in August 2018. The outcomes were critical care (admission to an intensive care unit and/or in-hospital death) and hospitalization (direct hospital admission or transfer). In the training set (70% random sample), using routinely available triage data as predictors (eg, demographic characteristics and vital signs), we derived 4 machine learning-based models: lasso regression, random forest, gradient-boosted decision tree, and deep neural network. In the test set (the remaining 30% of the sample), we measured the models' prediction performance by computing C statistics, prospective prediction results, and decision curves. These machine learning models were built for each outcome and compared with the reference model using the conventional triage classification information. Of 52 037 eligible ED visits by children (median [interquartile range] age, 6 [2-14] years; 24 929 [48.0%] female), 163 (0.3%) had the critical care outcome and 2352 (4.5%) had the hospitalization outcome. For the critical care prediction, all machine learning approaches had higher discriminative ability compared with the reference model, although the difference was not statistically significant (eg, C statistics of 0.85 [95% CI, 0.78-0.92] for the deep neural network vs 0.78 [95% CI, 0.71-0.85] for the reference; P = .16), and lower number of undertriaged critically ill children in the conventional triage levels 3 to 5 (urgent to nonurgent). For the hospitalization prediction, all machine learning approaches had significantly higher discrimination ability (eg, C statistic, 0.80 [95% CI, 0.78-0.81] for the deep neural network vs 0.73 [95% CI, 0.71-0.75] for the reference; P < .001) and fewer overtriaged children who did not require inpatient management in the conventional triage levels 1 to 3 (immediate to urgent). The decision curve analysis demonstrated a greater net benefit of machine learning models over ranges of clinical thresholds. Machine learning-based triage had better discrimination ability to predict clinical outcomes and disposition, with reduction in undertriaging critically ill children and overtriaging children who are less ill.