Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19.
Ontology highlight
ABSTRACT:
SUBMITTER: Duckworth C
PROVIDER: S-EPMC8626460 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:ObjectiveThis study aimed to describe the impact of the coronavirus disease (COVID-19) pandemic on emergency department (ED) admissions for urgent diagnoses.MethodsFrom January 1, 2019, until December 31, 2020, patients older than 18 years who attended the ED at University Hospital of Leuven (UZ Leuven, Belgium) were included. Urgent diagnoses selected in the First Hour Quintet were collected. The periods of the pandemic waves in 2020 were analyzed and compared with the same time period in 2019.ResultsDuring the first wave of the pandemic, 16 075 patients attended the ED compared with 16 893 patients during the comparison period in 2019. The proportion of patients having one of the diagnoses of the First Hour Quintet was similar between the periods (4.4% vs 4.5%). During the second wave, 14 739 patients attended the ED compared with 18 704 patients during the same period in 2019; 5.6% of patients had a diagnosis of the First Hour Quintet compared with 4.3% of patients in the comparison period.ConclusionThis study showed a decrease in the number of patients attending the ED during the COVID-19 pandemic. Further studies are needed to determine for which conditions patients visited the ED less.
| S-EPMC9343343 | biostudies-literature
Project description:Study objectiveOur institution experienced a change in SARS-CoV-2 testing policy as well as substantial changes in local COVID-19 prevalence, allowing for a unique examination of the relationship between SARS-CoV-2 testing and emergency department (ED) length of stay.MethodsThis was an observational interrupted time series of all patients admitted to an academic health system between March 15, 2020, and September 30, 2020. Given testing limitations from March 15 to April 24, all patients receiving SARS-CoV-2 tests were symptomatic. On April 24, testing was expanded to all ED admissions. The primary and secondary outcomes were ED length of stay and number needed to test to obtain a positive, respectively.ResultsA total of 70,856 patients were cared for in the EDs during the 7-month period. The testing change increased admission length of stay by 1.89 hours (95% confidence interval 1.39 to 2.38). The number needed to test was 2.5 patients and was highest yield on April 1, 2020, when the state positivity rate was 39.7%; however, the number needed to test exceeded 170 patients by Sept 1, 2020, at which point the state positivity rate was 0.5%.ConclusionAlthough universal SARS-CoV-2 testing of ED admissions may meaningfully support mitigation and containment efforts, the clinical cost of testing all admissions amid low community positivity is notable. In our system, universal ED SARS-CoV-2 testing was associated with a 24% increase in admission length of stay alongside the detection of only 1 positive case every other day. Given the known harms and risks of ED boarding and crowding, solutions must be developed to support regular operational flow while balancing infection prevention needs.
| S-EPMC8424016 | biostudies-literature
Project description:Emergency department (ED) crowding is linked with poor quality of care and worse outcomes, including higher mortality. With the growing emphasis on hospital performance measures, there is additional concern whether inadequate care during crowded periods increases a patient's likelihood of subsequent inpatient admission. We sought to determine if ED crowding during the index visit was associated with these "bounceback" admissions.We used comprehensive, nonpublic, statewide ED and inpatient discharge data from the California Office of Statewide Health Planning and Development from 2007 to identify index outpatient ED visits and bounceback admissions within 7 days. We further used ambulance diversion data collected from California local emergency medical services agencies to identify crowded days using intrahospital daily diversion hour quartiles. Using a hierarchical logistic regression model, we then determined if patients visiting on crowded days were more likely to have a subsequent bounceback admission.We analyzed 3,368,527 index visits across 202 hospitals, of which 596,471 (17.7%) observations were on crowded days. We found no association between ED crowding and bounceback admissions. This lack of relationship persisted in both a discrete (high/low) model (OR, 1.01; 95% CI, 0.99, 1.02) and a secondary model using ambulance diversion hours as a continuous predictor (OR, 1.00; 95% CI, 1.00, 1.00).Crowding as measured by ambulance diversion does not have an association with hospitalization within 7 days of an ED visit discharge. Therefore, bounceback admission may be a poor measure of delayed or worsened quality of care due to crowding.
| S-EPMC3839788 | biostudies-literature
Project description:BackgroundFor a number of fiscal and practical reasons, data on heroin use have been of poor quality, which has hampered the ability to halt the growing epidemic. Internet search data, such as those made available by Google Trends, have been used as a low-cost, real-time data source for monitoring and predicting a variety of public health outcomes. We aimed to determine whether data on opioid-related internet searches might predict future heroin-related admissions to emergency departments (ED).MethodsAcross nine metropolitan statistical areas (MSAs) in the United States, we obtained data on Google searches for prescription and non-prescription opioids, as well as Substance Abuse and Mental Health Services Administration (SAMHSA) data on heroin-related ED visits from 2004 to 2011. A linear mixed model assessed the relationship between opioid-related Internet searches and following year heroin-related visits, controlling for MSA GINI index and total number of ED visits.ResultsThe best-fitting model explained 72% of the variance in heroin-related ED visits. The final model included the search keywords "Avinza," "Brown Sugar," "China White," "Codeine," "Kadian," "Methadone," and "Oxymorphone." We found regional differences in where and how people searched for opioid-related information.ConclusionsInternet search-based modeling should be explored as a new source of insights for predicting heroin-related admissions. In geographic regions where no current heroin-related data exist, Internet search modeling might be a particularly valuable and inexpensive tool for estimating changing heroin use trends. We discuss the immediate implications for using this approach to assist in managing opioid-related morbidity and mortality in the United States.
| S-EPMC6436381 | biostudies-literature
Project description:Introduction: Patients boarding in the Emergency Department can contribute to overcrowding, leading to longer waiting times and patients leaving without being seen or completing their treatment. The early identification of potential admissions could act as an additional decision support tool to alert clinicians that a patient needs to be reviewed for admission and would also be of benefit to bed managers in advance bed planning for the patient. We aim to create a low-dimensional model predicting admissions early from the paediatric Emergency Department. Methods and Analysis: The methodology Cross Industry Standard Process for Data Mining (CRISP-DM) will be followed. The dataset will comprise of 2 years of data, ~76,000 records. Potential predictors were identified from previous research, comprising of demographics, registration details, triage assessment, hospital usage and past medical history. Fifteen models will be developed comprised of 3 machine learning algorithms (Logistic regression, naïve Bayes and gradient boosting machine) and 5 sampling methods, 4 of which are aimed at addressing class imbalance (undersampling, oversampling, and synthetic oversampling techniques). The variables of importance will then be identified from the optimal model (selected based on the highest Area under the curve) and used to develop an additional low-dimensional model for deployment. Discussion: A low-dimensional model comprised of routinely collected data, captured up to post triage assessment would benefit many hospitals without data rich platforms for the development of models with a high number of predictors. Novel to the planned study is the use of data from the Republic of Ireland and the application of sampling techniques aimed at improving model performance impacted by an imbalance between admissions and discharges in the outcome variable.
| S-EPMC8085432 | biostudies-literature
Project description:BACKGROUND:People with cancer experience symptoms related to the disease and treatments. Symptom distress has a negative impact on quality of life (QoL). Attending to symptoms and side effects of treatment promotes safe and effective delivery of therapies and may prevent or reduce emergency department visits (EDVs) and unplanned hospital admissions (HAs). There is limited evidence examining symptom-related EDVs or HAs (sx-EDV/HAs) and interventions in ambulatory oncology patients. OBJECTIVE:To examine factors associated with sx-EDV/HAs in ambulatory oncology patients receiving chemotherapy and/or radiation. METHODS:This secondary analysis used data from a randomized controlled trial of ambulatory oncology patients (n = 663) who received the web-based Electronic Self-Report Assessment – Cancer intervention (symptom self-monitoring, tailored education, and communication coaching) or usual care with symptom self-monitoring alone. Group differences were described by summary statistics and compared by t test. Factors associated with the odds of at least 1 sx-EDV/HA were modeled using logistic regression. RESULTS:98 patients had a total of 171 sx-EDV/HAs with no difference between groups. Higher odds of at least 1 sx-EDV/HA were associated with socioeconomic and clinical factors. The multivariable model indicated that work status, education level, treatment modality, and on-treatment Symptom Distress Scale-15 scores were signifcantly associated with having at least 1 sx-EDV/HA. LIMITATIONS:This is a secondary analysis not sized to determine cause and effect. The results have limited generalizability. CONCLUSION:Most patients did not experience a sx-EDV/HA. Demographic and clinical factors predicted a sx-EDV/HA.
| S-EPMC4849131 | biostudies-literature
Project description:BackgroundEmergency departments (EDs) see a rising number of patients, but only a small fraction of ED patients need immediate intensive care. The characteristics of these patients are mostly unknown and there is reason to believe that there are large inter-hospital differences in thresholds for intensive care admissions from the ED. The purpose of this study was to give a nationwide overview of ED admissions directly to intensive care units.MethodsWe used the Swedish Intensive care Registry to identify all patients admitted to intensive care from the ED (January 1, 2013 until June 7, 2018). The primary outcome was discharge diagnosis after intensive care (primary ICU diagnosis code). ICU mortality and" ICU admission due to only observation" were analyzed as secondary outcomes.Results110,072 ICU admissions were included, representing 94,546 unique patients. Intoxication, trauma and neurological conditions were the most common causes for intensive care, but large variations according to age, sex and hospital type were seen. Intoxication was the most prevalent diagnosis in young adults (46.8% of admissions in 18-29 years old), whereas infectious diseases predominated in the elderly (17.0% in 65-79 years old). Overall, ICU mortality was 7.2%, but varied substantially with age, sex, type of hospital and medical condition. Cardiac conditions had the highest mortality rates, reaching 32.9%. The mortality was higher in academic centers compared to rural hospitals (9.3% vs 5.0%). It was more common to be admitted to ICU for only observation in rural hospitals than in academic centers (20.1% vs 7.8%). Being admitted to ICU only for observation was most common in patients with intoxication (30.6%).ConclusionsOverall, intoxication was the most common cause for ICU admission from the ED. However, causes of ED ICU admissions differ substantially according to age, sex and hospital type. Being admitted to the ICU only for observation was most common in intoxicated patients.Trial registrationNot applicable (no interventions).
| S-EPMC8540137 | biostudies-literature
Project description:ObjectiveHeadaches represent 0.9% to 2.6% of visits to a pediatric emergency department (PED). We noted a trend of increasing visits for headache in our tertiary care PED and sought to further characterize this trend.MethodsWe identified PED visits with International Classification of Disease, Ninth Revision, Clinical Modification diagnoses for headache at 25 hospitals in Pediatric Health Information System between 2003 and 2013. To further characterize demographics and treatment trends over time we used the electronic health record in our emergency department to identify children ages four to 18 between January 2007 and December 2014 with International Classification of Disease, Ninth Revision codes for headache: a random sample of 50 visits per year were chosen for chart review.ResultsPediatric Health Information System visits for headache increased by 166% (18,041 in 2003 and 48,020 in 2013); by comparison, total PED visits increased by 57.6%. The percent admission increased by 300% (2020 admissions in 2003 and 8087 admissions in 2013). At our hospital, headache visits increased 111% from 896 visits in 2007 to 1887 visits in 2014; total PED visits increased 30.2%. The admission percentage for headache increased 187% with 156 admissions in 2007 and 448 in 2014. Management over time differed in the frequency of head computed tomography which decreased 3.7% per year (r = -0.93, 95% CI -0.99, -0.64) from 34% in 2007 to 18% in 2014.ConclusionPediatric emergency department visits for headache are increasing and a growing proportion of these patients are admitted. This finding identifies a potential patient population to target for interventions to improve outpatient management and reduce pediatric emergency department utilization.
| S-EPMC8485652 | biostudies-literature
Project description:Heat waves have been linked to increases in emergency-related morbidity, but more research is needed on the demographic and disease-specific aspects of these morbidities. Using a case-crossover approach, over 700,000 daily emergency department hospital admissions in Charlottesville, Virginia, U.S.A. from 2005⁻2016 are compared between warm season heat wave and non-heat wave periods. Heat waves are defined based on the exceedance, for at least three consecutive days, of two apparent temperature thresholds (35 °C and 37 °C) that account for 3 and 6% of the period of record. Total admissions and admissions for whites, blacks, males, females, and 20⁻49 years old are significantly elevated during heat waves, as are admissions related to a variety of diagnostic categories, including diabetes, pregnancy complications, and injuries and poisoning. Evidence that heat waves raise emergency department admissions across numerous demographic and disease categories suggests that heat exerts comorbidity influences that extend beyond the more well-studied direct relationships such as heat strokes and cardiac arrest.
| S-EPMC6068980 | biostudies-literature
Project description:IntroductionIntrahospital transports (IHTs) are high-risk activities with the potential for adverse outcomes. Suboptimal care of a patient in our emergency department (ED) needing IHT to the pediatric intensive care unit (ICU) identified improvement opportunities. We describe implementing a novel checklist (Briefing ED-to-ICU Transport To Exit Ready: BETTER) for improving the IHT safety of pediatric ED patients admitted to the pediatric ICU.MethodsA multidisciplinary team used the Model for Improvement to create a key driver diagram and process map. An evidence-based IHT checklist was implemented on July 23, 2019 after multiple plan-do-study-act checklist revisions. The specific aim was a ≥80% checklist completion rate for 6 months and maintaining that rate for 6 months. An anonymous, voluntary survey of ED nurses and physicians, 9 months postimplementation, evaluated perceived improvements in IHT safety. The outcome measure was the proportion of IHT-related incident reports, per ED-to-pediatric ICU admission, comparing baseline (2-year preimplementation) and intervention (1-year postimplementation) periods. Balancing measures included a quantitative assessment for any throughput measure delays and a survey question on perceived delays.ResultsFrom July 23, 2019 to July 22, 2020, 335 (84%) of 400 ED-to-ICU admissions had completed IHT checklists. Ninety percent of survey respondents (84% response rate) agreed that the checklist improved IHT safety. The incident report rate was lower in the intervention period (0.5% versus 2.3%; P = 0.03), with special cause improvement on T-chart analysis. Balancing measures did not indicate any delays secondary to checklist implementation.ConclusionsThis IHT checklist was feasible and associated with improvements in perceived safety and incident event reporting. Further studies are needed to assess generalizability.
| S-EPMC8225371 | biostudies-literature