Patient Ethnicity Affects Triage Assessments and Patient Prioritization in U.S. Department of Veterans Affairs Emergency Departments

Adaption and National Validation of a Tool for Predicting Mortality from Other Causes Among Men with Nonmetastatic Prostate Cancer

BACKGROUND

An electronic health record-based tool could improve accuracy and eliminate bias in provider estimation of the risk of death from other causes among men with nonmetastatic cancer.

OBJECTIVE

To recalibrate and validate the Veterans Aging Cohort Study Charlson Comorbidity Index (VACS-CCI) to predict non-prostate cancer mortality (non-PCM) and to compare it with a tool predicting prostate cancer mortality (PCM).

DESIGN, SETTING, AND PARTICIPANTS

An observational cohort of men with biopsy-confirmed nonmetastatic prostate cancer, enrolled from 2001 to 2018 in the national US Veterans Health Administration (VA), was divided by the year of diagnosis into the development (2001-2006 and 2008-2018) and validation (2007) sets.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Mortality (all cause, non-PCM, and PCM) was evaluated. Accuracy was assessed using calibration curves and C statistic in the development, validation, and combined sets; overall; and by age (<65 and 65+ yr), race (White and Black), Hispanic ethnicity, and treatment groups.

RESULTS AND LIMITATIONS

Among 107 370 individuals, we observed 24 977 deaths (86% non-PCM). The median age was 65 yr, 4947 were Black, and 5010 were Hispanic. Compared with CCI and age alone (C statistic 0.67, 95% confidence interval [CI] 0.67-0.68), VACS-CCI demonstrated improved validated discrimination (C statistic 0.75, 95% CI 0.74-0.75 for non-PCM). The prostate cancer mortality tool also discriminated well in validation (C statistic 0.81, 95% CI 0.78-0.83). Both were well calibrated overall and within subgroups. Owing to missing data, 18 009/125 379 (14%) were excluded, and VACS-CCI should be validated outside the VA prior to outside application.

CONCLUSIONS

VACS-CCI is ready for implementation within the VA. Electronic health record-assisted calculation is feasible, improves accuracy over age and CCI alone, and could mitigate inaccuracy and bias in provider estimation.

PATIENT SUMMARY

Veterans Aging Cohort Study Charlson Comorbidity Index is ready for application within the Veterans Health Administration. Electronic health record-assisted calculation is feasible, improves accuracy over age and Charlson Comorbidity Index alone, and might help mitigate inaccuracy and bias in provider estimation of the risk of non-prostate cancer mortality.

Applications of natural language processing at emergency department triage: A narrative review

INTRODUCTION

Natural language processing (NLP) uses various computational methods to analyse and understand human language, and has been applied to data acquired at Emergency Department (ED) triage to predict various outcomes. The objective of this scoping review is to evaluate how NLP has been applied to data acquired at ED triage, assess if NLP based models outperform humans or current risk stratification techniques when predicting outcomes, and assess if incorporating free-text improve predictive performance of models when compared to predictive models that use only structured data.

METHODS

All English language peer-reviewed research that applied an NLP technique to free-text obtained at ED triage was eligible for inclusion. We excluded studies focusing solely on disease surveillance, and studies that used information obtained after triage. We searched the electronic databases MEDLINE, Embase, Cochrane Database of Systematic Reviews, Web of Science, and Scopus for medical subject headings and text keywords related to NLP and triage. Databases were last searched on 01/01/2022. Risk of bias in studies was assessed using the Prediction model Risk of Bias Assessment Tool (PROBAST). Due to the high level of heterogeneity between studies and high risk of bias, a metanalysis was not conducted. Instead, a narrative synthesis is provided.

RESULTS

In total, 3730 studies were screened, and 20 studies were included. The population size varied greatly between studies ranging from 1.8 million patients to 598 triage notes. The most common outcomes assessed were prediction of triage score, prediction of admission, and prediction of critical illness. NLP models achieved high accuracy in predicting need for admission, triage score, critical illness, and mapping free-text chief complaints to structured fields. Incorporating both structured data and free-text data improved results when compared to models that used only structured data. However, the majority of studies (80%) were assessed to have a high risk of bias, and only one study reported the deployment of an NLP model into clinical practice.

CONCLUSION

Unstructured free-text triage notes have been used by NLP models to predict clinically relevant outcomes. However, the majority of studies have a high risk of bias, most research is retrospective, and there are few examples of implementation into clinical practice. Future work is needed to prospectively assess if applying NLP to data acquired at ED triage improves ED outcomes when compared to usual clinical practice.

County-level social vulnerability and nonfatal drug overdose emergency department visits and hospitalizations, January 2018-December 2020

BACKGROUND

Nonfatal drug overdoses (NFODs) are often attributed to individual behaviors and risk factors; however, identifying community-level social determinants of health (SDOH) associated with increased NFOD rates may allow public health and clinical providers to develop more targeted interventions to address substance use and overdose health disparities. CDC's Social Vulnerability Index (SVI), which aggregates social vulnerability data from the American Community Survey to produce ranked county-level vulnerability scores, can help identify community factors associated with NFOD rates. This study aims to describe associations between county-level social vulnerability, urbanicity, and NFOD rates.

METHODS

We analyzed county-level 2018-2020 emergency department (ED) and hospitalization discharge data submitted to CDC's Drug Overdose Surveillance and Epidemiology system. Counties were ranked in vulnerability quartiles based on SVI data. We used crude and adjusted negative binomial regression models, by drug category, to calculate rate ratios and 95% confidence intervals comparing NFOD rates by vulnerability.

RESULTS

Generally, as social vulnerability scores increased, ED and hospitalization NFOD rates increased; however, the magnitude of the association varied across drugs, visit type, and urbanicity. SVI-related theme and individual variable analyses highlighted specific community characteristics associated with NFOD rates.

CONCLUSIONS

The SVI can help identify associations between social vulnerabilities and NFOD rates. Development of an overdose-specific validated index could improve translation of findings to public health action. The development and implementation of overdose prevention strategies should consider a socioecological perspective and address health inequities and structural barriers associated with increased risk of NFODs at all levels of the social ecology.

Disparities in Care Among Gunshot Victims: A Nationwide Analysis

INTRODUCTION

Given the well-known healthcare disparities most pronounced in racial and ethnic minorities, trauma healthcare in underrepresented patients should be examined, as in-hospital bias may influence the care rendered to patients. This study seeks to examine racial differences in outcomes and resource utilization among victims of gunshot wounds in the United States.

METHODS

This is a retrospective review of the National Trauma Data Bank (NTDB) conducted from 2007 to 2017. The NTDB was queried for patients who suffered a gunshot wound not related to accidental injury or suicide. Patients were stratified according to race. The primary outcome for this study was mortality. Secondary outcomes included racial differences in resource utilization including air transport and discharge to rehabilitation centers. Univariate and multivariate analyses were used to compare differences in outcomes between the groups.

RESULTS

A total of 250,675 patients were included in the analysis. After regression analysis, Black patients were noted to have greater odds of death compared to White patients (odds ratio [OR] 1.14, confidence interval [CI] 1.037-1.244; P = 0.006) and decreased odds of admission to the intensive care unit (ICU) (OR 0.76, CI 0.732-0.794; P < 0.001). Hispanic patients were significantly less likely to be discharged to rehabilitation centers (Hispanic: 0.78, CI 0.715-0.856; P < 0.001). Black patients had the shortest time to death (median time in minutes: White 49 interquartile range [IQR] [9-437] versus Black 24 IQR [7-205] versus Hispanic 39 IQR [8-379] versus Asian 60 [9-753], P < 0.001).

CONCLUSIONS

As society carefully examines major institutions for implicit bias, healthcare should not be exempt. Greater mortality among Black patients, along with differences in other important outcome measures, demonstrate disparities that encourage further analysis of causes and solutions to these issues.

Racial and Ethnic Differences in Ondansetron Use for Acute Gastroenteritis in Children

OBJECTIVES

There is limited research examining racial/ethnic disparities in antiemetic use for acute gastroenteritis (AGE). We assessed racial/ethnic differences in the care of children with AGE.

METHODS

The Pediatric Health Information System was used to conduct a retrospective cohort study of children 6 months to 6 years old with AGE seen in participating emergency departments from 2016 to 2018. Cases were identified using International Classification of Diseases, Tenth Revision codes. The primary outcome was administration of ondansetron, secondary outcomes were administration of intravenous (IV) fluids and hospitalization, and primary predictor was race/ethnicity. Multivariable logistic regression followed by a mixed model adjusted for sex, age, insurance, and hospital to examine the association of race/ethnicity with each outcome.

RESULTS

There were 78,019 encounters included; 24.8% of patients were non-Hispanic White (NHW), 29.0% non-Hispanic Black (NHB), 37.3% Hispanic, and 8.9% other non-Hispanic (NH) race/ethnicity. Compared with NHW patients, minority children were more likely to receive ondansetron (NHB: adjusted odds ratio, 1.36 [95% confidence interval, 1.2-1.55]; Hispanic: 1.26 [1.1-1.44]; other NH: 1.22 [1.07-1.4]). However, minority children were less likely to receive IV fluids (NHB: 0.38 [0.33-0.43]; Hispanic: 0.44 [0.36-0.53]; other NH: 0.51 [0.44-0.61]) or hospital admission (NHB: 0.37 [0.29-0.48]; Hispanic: 0.41 [0.33-0.5]; other NH: 0.52 [0.41-0.66]). Ondansetron use by hospital ranged from 73% to 95%.

CONCLUSIONS

This large database analysis of emergency departments around the nation found that NHW patients were less likely to receive ondansetron but more likely to receive IV fluids and hospital admission than minority patients. These findings are likely multifactorial and may represent bias, social determinants of health, access to care, or illness severity among other possible causes.

Development and Structure of an Accurate Machine Learning Algorithm to Predict Inpatient Mortality and Hospice Outcomes in the Coronavirus Disease 2019 Era

BACKGROUND

The coronavirus disease 2019 (COVID-19) pandemic has challenged the accuracy and racial biases present in traditional mortality scores. An accurate prognostic model that can be applied to hospitalized patients irrespective of race or COVID-19 status may benefit patient care.

RESEARCH DESIGN

This cohort study utilized historical and ongoing electronic health record features to develop and validate a deep-learning model applied on the second day of admission predicting a composite outcome of in-hospital mortality, discharge to hospice, or death within 30 days of admission. Model features included patient demographics, diagnoses, procedures, inpatient medications, laboratory values, vital signs, and substance use history. Conventional performance metrics were assessed, and subgroup analysis was performed based on race, COVID-19 status, and intensive care unit admission.

SUBJECTS

A total of 35,521 patients hospitalized between April 2020 and October 2020 at a single health care system including a tertiary academic referral center and 9 community hospitals.

RESULTS

Of 35,521 patients, including 9831 non-White patients and 2020 COVID-19 patients, 2838 (8.0%) met the composite outcome. Patients who experienced the composite outcome were older (73 vs. 61 y old) with similar sex and race distributions between groups. The model achieved an area under the receiver operating characteristic curve of 0.89 (95% confidence interval: 0.88, 0.91) and an average positive predictive value of 0.46 (0.40, 0.52). Model performance did not differ significantly in White (0.89) and non-White (0.90) subgroups or when grouping by COVID-19 status and intensive care unit admission.

CONCLUSION

A deep-learning model using large-volume, structured electronic health record data can effectively predict short-term mortality or hospice outcomes on the second day of admission in the general inpatient population without significant racial bias.

Disparities in Pediatric Emergency Department Length of Stay and Utilization Associated With Primary Language

OBJECTIVE

The aim of the study was to investigate the association between primary language and length of stay (LOS) in the pediatric emergency department (ED) within the context of known disparities impacting healthcare experiences and outcomes for patients with language barriers.

METHODS

We conducted a retrospective cohort study of consecutive encounters of patients presenting to, and discharged from, an urban pediatric ED from May 2015 through April 2018. Encounters were grouped into English primary language (EPL), Spanish (SPL), and other (OPL). Mean LOS comparisons were stratified by Emergency Severity Index (ESI). Bivariate and multivariate analyses were used to examine the relationship between LOS and variables, including age, sex, race/ethnicity, insurance, and time of presentation.

RESULTS

A total of 139,163 encounters were included. A higher proportion of SPL and OPL encounters were characterized as lower ESI acuity compared with EPL. Significantly longer LOS for SPL and OPL encounters was observed in the 2 lower acuity strata. The ESI 4-5 stratum demonstrated the greatest LOS disparity between EPL, SPL, and OPL (94 vs 103 vs 103 minutes, respectively, P < 0.001). In the highest acuity stratum, ESI 1-2, there was a nonsignificant trend toward longer LOS among EPL encounters (P = 0.08). The multivariate model accounted for 24% of LOS variance, but effect sizes were small for all variables except for ESI and age.

CONCLUSIONS

Patients with Spanish or other non-EPL who were triaged to lower acuity ESI levels experienced longer LOS in the pediatric ED than English-speaking counterparts. They also used the ED more frequently for low acuity issues, possibly reflecting disparities in access to primary care.

Racial and Language Disparities in Pediatric Emergency Department Triage

OBJECTIVE

The aim of this study was to assess the impact race and language have on emergency department (ED) triage scores while accounting for illness severity. We hypothesized that non-White and non-English-speaking patients were assigned lower-acuity triage scores compared with White and English-speaking patients, respectively.

METHODS

We used a chart review-based retrospective cohort study design, examining patients aged 0 to 17 years at our pediatric ED from July 2015 through June 2016. Illness severity was measured using a truncated Modified Pediatric Early Warning Score calculated from patient vital signs. We used univariate and multivariate multinomial logistic regression to assess the association between race and language with Emergency Severity Index scores.

RESULTS

Our final data set consisted of 10,815 visits from 8928 patients. Non-Hispanic (NH) White patients accounted for 34.6% of patients. In the adjusted analyses, non-White patients had significantly reduced odds of receiving a score of 2 (emergency) (odds ratio [OR], 0.4; 95% confidence interval [CI], 0.33-0.49) or 3 (urgent) (OR, 0.5; 95% CI, 0.45-0.56) and significantly higher odds of receiving a score of 5 (minor) (OR, 1.34; 95% CI, 1.07-1.69) versus a score of 4 (nonurgent). We did not find a consistent disparity in Emergency Severity Index scores when comparing English- and non-English-speaking patients.

CONCLUSIONS

We confirm that non-White patients receive lower triage scores than White patients. A more robust tool is required to account for illness severity and will be critical to understanding whether the relationship we describe reflects bias within the triage system or differences in ED utilization by racial groups.

Understanding Contributors to Racial/Ethnic Disparities in Emergency Department Throughput Times: a Sequential Mixed Methods Analysis

BACKGROUND

Ensuring equitable care remains a critical issue for healthcare systems. Nationwide evidence highlights the persistence of healthcare disparities and the need for research-informed approaches for reducing them at the local level.

OBJECTIVE

To characterize key contributors in racial/ethnic disparities in emergency department (ED) throughput times.

DESIGN

We conducted a sequential mixed methods analysis to understand variations in ED care throughput times for patients eventually admitted to an emergency department at a single academic medical center from November 2017 to May 2018 (n=3152). We detailed patient progression from ED arrival to decision to admit and compared racial/ethnic differences in time intervals from electronic medical record time-stamp data. We then estimated the relationships between race/ethnicity and ED throughput times, adjusting for several patient-level variables and ED-level covariates. These quantitative analyses informed our qualitative study design, which included observations and semi-structured interviews with patients and physicians.

KEY RESULTS

Non-Hispanic Black as compared to non-Hispanic White patients waited significantly longer during the time interval from arrival to the physician's decision to admit, even after adjustment for several ED-level and patient demographic, clinical, and socioeconomic variables (Beta (average minutes) (SE): 16.35 (5.8); p value=.005). Qualitative findings suggest that the manner in which providers communicate, advocate, and prioritize patients may contribute to such disparities. When the race/ethnicity of provider and patient differed, providers were more likely to interrupt patients, ignore their requests, and make less eye contact. Conversely, if the race/ethnicity of provider and patient were similar, providers exhibited a greater level of advocacy, such as tracking down patient labs or consultants. Physicians with no significant ED throughput disparities articulated objective criteria such as triage scores for prioritizing patients.

CONCLUSIONS

Our findings suggest the importance of (1) understanding how our communication style and care may differ by race/ethnicity; and (2) taking advantage of structured processes designed to equalize care.

Breaking through barriers: the need for effective research to promote language-concordant communication as a facilitator of equitable emergency care

Individuals with limited English proficiency (LEP) are at high risk for adverse outcomes in the US health care system. This is particularly true for patients with LEP seeking care in the emergency department (ED). Although professional language interpretation improves the quality of care for these patients, it remains underused. The dynamic, discontinuous nature of an ED visit poses distinct challenges and opportunities for providing equitable, high-quality care for patients with LEP. Evidence-based best practices for identifying patients with LEP and using professional interpretation are well described but inadequately implemented. There are few examples in the literature of rigorous interventions to improve quality of care and outcomes for patients with LEP. There is an urgent need for high-quality research to improve communication with patients with LEP along the continuum of emergency care in order to achieve equity in outcomes.

Examining emergency department inequities in Aotearoa New Zealand: Findings from a national retrospective observational study examining Indigenous emergency care outcomes

OBJECTIVE

There is increasing evidence that EDs may not operate equitably for all patients, with Indigenous and minoritised ethnicity patients experiencing longer wait times for assessment, differential pain management and less evaluation and treatment of acute conditions.

METHODS

This retrospective observational study used a Kaupapa Māori framework to investigate ED admissions into 18/20 District Health Boards in Aotearoa New Zealand (2006-2012). Key pre-admission variable was ethnicity (Māori:non-Māori), and outcome variables included: ED self-discharge; ED arrival to assessment time; hospital re-admission within 72 h; ED re-presentation within 72 h; ED length of stay; ward length of stay; access block and mortality (in ED or within 10 days of ED departure). Generalised linear regression models controlled for year of presentation, sex, age, deprivation, triage category and comorbidity.

RESULTS

Despite some ED process measures favouring Māori, for example arrival to assessment time (mean difference -2.14 min; 95% confidence interval [CI] -2.42 to -1.86) and access block (odds ratio [OR] 0.89, 95% CI 0.87-0.91), others showed no difference, for example self-discharge (OR 0.98, 95% CI 0.97-1.00). Despite this, Māori mortality (OR 1.60, 95% CI 1.50-1.71) and ED re-presentation (OR 1.11, 95% CI 1.09-1.12) were higher than non-Māori.

CONCLUSION

To our knowledge, this is the most comprehensive investigation of acute outcomes by ethnicity to date in New Zealand. We found ED mortality inequities that are unlikely to be explained by ED process measures or comorbidities. Our findings reinforce the need to investigate health professional bias and institutional racism within an acute care context.

First Nations status and emergency department triage scores in Alberta: a retrospective cohort study

BACKGROUND

Previous studies have found that race is associated with emergency department triage scores, raising concerns about potential health care inequity. As part of a project on quality of care for First Nations people in Alberta, we sought to understand the relation between First Nations status and triage scores.

METHODS

We conducted a population-based retrospective cohort study of health administrative data from April 2012 to March 2017 to evaluate acuity of triage scores, categorized as a binary outcome of higher or lower acuity score. We developed multivariable multilevel logistic mixed-effects regression models using the levels of emergency department visit, patient (for patients with multiple visits) and facility. We further evaluated the triage of visits related to 5 disease categories and 5 specific diagnoses to better compare triage outcomes of First Nations and non-First Nations patients.

RESULTS

First Nations status was associated with lower odds of receiving higher acuity triage scores (odds ratio [OR] 0.93, 95% confidence interval [CI] 0.92-0.94) compared with non-First Nations patients in adjusted models. First Nations patients had lower odds of acute triage for all 5 disease categories and for 3 of 5 diagnoses, including long bone fractures (OR 0.82, 95% CI 0.76-0.88), acute upper respiratory infection (OR 0.90, 95% CI 0.84-0.98) and anxiety disorder (OR 0.67, 95% CI 0.60-0.74).

INTERPRETATION

First Nations status was associated with lower odds of higher acuity triage scores across a number of conditions and diagnoses. This may reflect systemic racism, stereotyping and potentially other factors that affected triage assessments.

Machine learning methods applied to triage in emergency services: A systematic review

BACKGROUND

In emergency services is important to accurately assess and classify symptoms, which may be improved with the help of technology. One mechanism that could help and improve predictions from health records or patient flow is machine learning (ML).

AIM

To analyse the effectiveness of ML systems in triage for making predictions at the emergency department in comparison with other triage scales/scores.

METHODS

Following the PRISMA recommendations, a systematic review was conducted using CINAHL, Cochrane, Cuiden, Medline and Scopus databases with the search equation "Machine learning AND triage AND emergency".

RESULTS

Eleven studies were identified. The studies show that the use of ML methods consistently predict important outcomes like mortality, critical care outcomes and admission, and the need for hospitalization in comparison with scales like Emergency Severity Index or others. Among the ML models considered, XGBoost and Deep Neural Networks obtained the highest levels of prediction accuracy, while Logistic Regression performed obtained the worst values.

CONCLUSIONS

Machine learning methods can be a good instrument for helping triage process with the prediction of important emergency variables like mortality or the need for critical care or hospitalization.

Patient race, ethnicity, and care in the emergency department: A scoping review

OBJECTIVES

Health disparities between racial and ethnic groups have been documented in Canada, the United States, and Australia. Despite evidence that differences in emergency department (ED) care based on patient race and ethnicity exist, there are no comprehensive literature reviews in this area. The objective of this review is to provide an overview of the literature on the impact of patient ethnicity and race on the processes of ED care.

METHODS

A scoping review was conducted to capture the broad nature of the literature. A database search was conducted in MEDLINE/PubMed, EMBASE, CINAHL Plus, Social Sciences Citation Index, SCOPUS, and JSTOR. Five journals and reference lists of included articles were hand searched. Inclusion and exclusion criteria were defined iteratively to ensure literature captured was relevant to our research question. Data were extracted using predetermined variables, and additional extraction variables were added as familiarity with the literature developed.

RESULTS

Searching yielded 1,157 citations, reduced to 153 following removal of duplicates, and title and abstract screening. After full-text screening, 83 articles were included. Included articles report that, in EDs, patient race and ethnicity impact analgesia, triage scores, wait times, treatments, diagnostic procedure utilization, rates of patients leaving without being seen, and patient subjective experiences. Authors of included studies propose a variety of possible causes for these disparities.

CONCLUSIONS

Further research on the existence of disparities in care within EDs is warranted to explore the causes behind observed disparities for particular health conditions and population groups in specific contexts.

Assessment of Disparities Associated With a Crisis Standards of Care Resource Allocation Algorithm for Patients in 2 US Hospitals During the COVID-19 Pandemic

IMPORTANCE

Significant concern has been raised that crisis standards of care policies aimed at guiding resource allocation may be biased against people based on race/ethnicity.

OBJECTIVE

To evaluate whether unanticipated disparities by race or ethnicity arise from a single institution's resource allocation policy.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study included adults (aged ≥18 years) who were cared for on a coronavirus disease 2019 (COVID-19) ward or in a monitored unit requiring invasive or noninvasive ventilation or high-flow nasal cannula between May 26 and July 14, 2020, at 2 academic hospitals in Miami, Florida.

EXPOSURES

Race (ie, White, Black, Asian, multiracial) and ethnicity (ie, non-Hispanic, Hispanic).

MAIN OUTCOMES AND MEASURES

The primary outcome was based on a resource allocation priority score (range, 1-8, with 1 indicating highest and 8 indicating lowest priority) that was assigned daily based on both estimated short-term (using Sequential Organ Failure Assessment score) and longer-term (using comorbidities) mortality. There were 2 coprimary outcomes: maximum and minimum score for each patient over all eligible patient-days. Standard summary statistics were used to describe the cohort, and multivariable Poisson regression was used to identify associations of race and ethnicity with each outcome.

RESULTS

The cohort consisted of 5613 patient-days of data from 1127 patients (median [interquartile range {IQR}] age, 62.7 [51.7-73.7]; 607 [53.9%] men). Of these, 711 (63.1%) were White patients, 323 (28.7%) were Black patients, 8 (0.7%) were Asian patients, and 31 (2.8%) were multiracial patients; 480 (42.6%) were non-Hispanic patients, and 611 (54.2%) were Hispanic patients. The median (IQR) maximum priority score for the cohort was 3 (1-4); the median (IQR) minimum score was 2 (1-3). After adjustment, there was no association of race with maximum priority score using White patients as the reference group (Black patients: incidence rate ratio [IRR], 1.00; 95% CI, 0.89-1.12; Asian patients: IRR, 0.95; 95% CI. 0.62-1.45; multiracial patients: IRR, 0.93; 95% CI, 0.72-1.19) or of ethnicity using non-Hispanic patients as the reference group (Hispanic patients: IRR, 0.98; 95% CI, 0.88-1.10); similarly, no association was found with minimum score for race, again with White patients as the reference group (Black patients: IRR, 1.01; 95% CI, 0.90-1.14; Asian patients: IRR, 0.96; 95% CI, 0.62-1.49; multiracial patients: IRR, 0.81; 95% CI, 0.61-1.07) or ethnicity, again with non-Hispanic patients as the reference group (Hispanic patients: IRR, 1.00; 95% CI, 0.89-1.13).

CONCLUSIONS AND RELEVANCE

In this cohort study of adult patients admitted to a COVID-19 unit at 2 US hospitals, there was no association of race or ethnicity with the priority score underpinning the resource allocation policy. Despite this finding, any policy to guide altered standards of care during a crisis should be monitored to ensure equitable distribution of resources.

Examining emergency department inequities: Descriptive analysis of national data (2006-2012)

OBJECTIVE

Internationally, Indigenous and minoritised ethnic groups experience longer wait times, differential pain management and less evaluation and treatment for acute conditions within emergency medicine care. Examining ED Inequities (EEDI) aims to investigate whether inequities between Māori and non-Māori exist within EDs in Aotearoa New Zealand (NZ). This article presents the descriptive findings for the present study.

METHODS

A retrospective observational study framed from a Kaupapa Māori positioning, EEDI uses secondary data from emergency medicine admissions into 18/20 District Health Boards in NZ between 2006 and 2012. Data sources include variables from the Shorter Stays in ED National Research Project database and comorbidity data from NZ's National Minimum Dataset. The key predictor of interest is patient ethnicity with descriptive variables, including sex, age group, area deprivation, mode of presentation, referral method, Australasian Triage Scale and trauma status.

RESULTS

There were a total of 5 972 102 ED events (1 168 944 Māori, 4 803 158 non-Māori). We found an increasing proportion of ED events per year, with a higher proportion of Māori from younger age groups and areas of high deprivation compared to non-Māori events. Māori also had a higher proportion of self-referral and were triaged to be seen within a longer time frame compared to non-Māori.

CONCLUSION

Our findings show that there are different patterns of ED usage when comparing Māori and non-Māori events. The next level of analysis of the EEDI dataset will be to examine whether there are any associations between ethnicity and ED outcomes for Māori and non-Māori patients.

Racial and ethnic disparities in hospital observation in Maryland

OBJECTIVES

Hospital observation is a key disposition option from the emergency department (ED) and encompasses up to one third of patients requiring post-ED care. Observation has been associated with higher incidence of catastrophic financial costs and has downstream effects on post-discharge clinical services. Yet little is known about the non-clinical determinants of observation assignment. We sought to evaluate the impact of patient-level demographic factors on observation designation among Maryland patients.

METHODS

We conducted a retrospective analysis of all ED encounters in Maryland between July 2012 and January 2017 for four priority diagnoses (heart failure, chronic obstructive pulmonary disease [COPD], pneumonia, and acute chest pain) using multilevel logistic models allowing for heterogeneity of the effects across hospitals. The primary exposure was self-reported race and ethnicity. The primary outcome was the initial status assignment from the ED: hospital observation versus inpatient admission.

RESULTS

Across 46 Maryland hospitals, 259,788 patient encounters resulted in a disposition of inpatient admission (65%) or observation designation (35%). Black (adjusted odds ratio [aOR]: 1.19; 95% confidence interval [CI]: 1.16-1.23) and Hispanic (aOR: 1.11; 95% CI: 1.01-1.21) patients were significantly more likely to be placed in observation than white, non-Hispanic patients. These differences were consistent across the majority of acute-care hospitals in Maryland (27/46).

CONCLUSION

Black and Hispanic patients in Maryland are more likely to be treated under the observation designation than white, non-Hispanic patients independent of clinical presentation. Race agnostic, time-based status assignments may be key in eliminating these disparities.

A Racially Unbiased, Machine Learning Approach to Prediction of Mortality: Algorithm Development Study

Background

Racial disparities in health care are well documented in the United States. As machine learning methods become more common in health care settings, it is important to ensure that these methods do not contribute to racial disparities through biased predictions or differential accuracy across racial groups.

Objective

The goal of the research was to assess a machine learning algorithm intentionally developed to minimize bias in in-hospital mortality predictions between white and nonwhite patient groups.

Methods

Bias was minimized through preprocessing of algorithm training data. We performed a retrospective analysis of electronic health record data from patients admitted to the intensive care unit (ICU) at a large academic health center between 2001 and 2012, drawing data from the Medical Information Mart for Intensive Care–III database. Patients were included if they had at least 10 hours of available measurements after ICU admission, had at least one of every measurement used for model prediction, and had recorded race/ethnicity data. Bias was assessed through the equal opportunity difference. Model performance in terms of bias and accuracy was compared with the Modified Early Warning Score (MEWS), the Simplified Acute Physiology Score II (SAPS II), and the Acute Physiologic Assessment and Chronic Health Evaluation (APACHE).

Results

The machine learning algorithm was found to be more accurate than all comparators, with a higher sensitivity, specificity, and area under the receiver operating characteristic. The machine learning algorithm was found to be unbiased (equal opportunity difference 0.016, P=.20). APACHE was also found to be unbiased (equal opportunity difference 0.019, P=.11), while SAPS II and MEWS were found to have significant bias (equal opportunity difference 0.038, P=.006 and equal opportunity difference 0.074, P<.001, respectively).

Conclusions

This study indicates there may be significant racial bias in commonly used severity scoring systems and that machine learning algorithms may reduce bias while improving on the accuracy of these methods.

Weight-for-Age Percentile as a Pediatric Predictor of Emergency Department Outcome

This study is a retrospective cohort study that examines the association between weight-for-age percentile and pediatric admission incidence from the emergency department (ED) for all diagnoses. The charts of 1432 pediatric patients under 18 years with ED visits from 2013 to 2015 at a tertiary children’s hospital were reviewed. Analyses of subject age/weight stratifications were performed, along with ED disposition, reason for visit, and Emergency Severity Index (ESI). Multivariable logistic regression models were used to evaluate the independent effect of weight-for-age percentile on ED disposition while controlling for age, ESI, and reason for visit. Underweight subjects were more likely to be admitted than their normal weight counterparts when analyzed overall (odds ratio [OR] = 2.58, P < .01) and by age: less than 2.0 years of age (OR = 2.04, P = .033), between 2.01 and 6.0 years of age (OR = 8.60, P = .004), and between 6.01 and 13.0 years of age (OR = 3.83, P = .053). Younger age (OR = 0.935, P < .001) and higher acuity (OR = 3.49, P < .001) were also significant predictors of admission. No significant associations were found between weight and likelihood of admission for patients older than 13.01 years or between overweight/obese weight categories and admission for any age subgroups. This study suggests that underweight children younger than 13 years are at higher risk to be admitted from the ED than their normal weight, overweight, and obese counterparts. Even when controlling for other key factors, such as the ESI, a lower weight-for-age percentile was a reliable predictor of hospitalization.

Utilization of Ophthalmologist Consultation for Emergency Care at a University Hospital

Importance

Nearly 2 million patients visit emergency departments (EDs) because of eye concerns annually in the United States. How hospitals currently assign these patients to treatment is important for designing systems that equitably allocate resources for eye care in urgent settings.

Objective

To investigate factors associated with ophthalmology consultation for eye-related adult ED encounters to assess possible disparities by sex, race/ethnicity, language preference, or residential distance from the medical center.

Design, Setting, and Participants

Retrospective observational study of 13 361 adult ED encounters associated with an eye-related billing diagnosis between January 1, 2010, and September 30, 2015, at the University of Michigan Medical Center in Ann Arbor.

Exposures

Measures available from the University of Michigan clinical data warehouse included age, sex, race/ethnicity, preferred language, home distance from the ED, calendar year of encounter, and Charlson-Deyo Comorbidity Index score.

Main Outcomes and Measures

Association of the ED encounter with ophthalmology consultation. An ophthalmology consultation was identified by cross-referencing ophthalmology faculty and clinical instructors from 2010 to 2015 against billing providers for consultations using the Charlson-Deyo Comorbidity Index score and billing codes. Measures included patient age, sex, race/ethnicity, home address, preferred language (English vs non-English), and calendar year of encounter.

Results

Among the 13 361 encounters, 6840 (51.2%) involved a female patient. Mean (SD) age at encounter was 50.7 (19.3) years; 10 033 patients (75.1%) were of white and 1969 (14.7%) of black race/ethnicity. English was the preferred language for 13 022 patients (97.5%). The ophthalmology service was consulted in 5289 encounters (39.6%). Black patients had significantly lower odds of an ophthalmology consultation than white patients (odds ratio [OR], 0.85; 95% CI, 0.75-0.96). Patients who preferred a non-English language had significantly lower odds of receiving an ophthalmology consultation (OR, 0.73; 95% CI, 0.55-0.98).

Conclusions and Relevance

Many of the 13 361 eye-related ED encounters were managed by ED clinicians with no ophthalmology consultation. Patients who were black or who preferred a language other than English were less likely to have an ophthalmologist involved in their care. The associations found in this observational study do not imply causation but suggest disparities in care that should be further investigated.

Emergency Department Utilization among Underserved African American Older Adults in South Los Angeles

Objectives: Using the Andersen’s Behavioral Model of Health Services Use, we explored social, behavioral, and health factors that are associated with emergency department (ED) utilization among underserved African American (AA) older adults in one of the most economically disadvantaged urban areas in South Los Angeles, California. Methods: This cross-sectional study recruited a convenience sample of 609 non-institutionalized AA older adults (age ≥ 65 years) from South Los Angeles, California. Participants were interviewed for demographic factors, self-rated health, chronic medication conditions (CMCs), pain, depressive symptoms, access to care, and continuity of care. Outcomes included 1 or 2+ ED visits in the last 12 months. Polynomial regression was used for data analysis. Results: Almost 41% of participants were treated at an ED during the last 12 months. In all, 27% of participants attended an ED once and 14% two or more times. Half of those with 6+ chronic conditions reported being treated at an ED once; one quarter at least twice. Factors that predicted no ED visit were male gender (OR = 0.50, 95% CI = 0.29–0.85), higher continuity of medical care (OR = 1.55, 95% CI = 1.04–2.31), individuals with two CMCs or less (OR = 2.61 (1.03–6.59), second tertile of pain severity (OR = 2.80, 95% CI = 1.36–5.73). Factors that predicted only one ED visit were male gender (OR = 0.45, 95% CI = 0.25–0.82), higher continuity of medical care (OR = 1.39, 95% CI = 1.01–2.15) and second tertile of pain severity (OR = 2.42, 95% CI = 1.13–5.19). Conclusions: This study documented that a lack of continuity of care for individuals with multiple chronic conditions leads to a higher rate of ED presentations. The results are significant given that ED visits may contribute to health disparities among AA older adults. Future research should examine whether case management decreases ED utilization among underserved AA older adults with multiple chronic conditions and/or severe pain. To explore the generalizability of these findings, the study should be repeated in other settings.

Leaving the emergency department without complete care: disparities in American Indian children

BACKGROUND

Children who leave the emergency department (ED) without complete evaluation or care (LWCET) have poorer outcomes in general. Previous studies have found that American Indian (AI) children have higher rates of LWCET than other racial or ethnic groups. Therefore, this study aims to examine LWCET in AI children by exploring differences by ED location and utilization patterns.

METHODS

This is a retrospective cohort study of five EDs in the upper Midwest between June 2011 and May 2012. We included all visits by children aged 0-17 who identified as African American (AA), AI or White. Logistic regression was used to determine differences in LWCET by race and ED location controlling for other possible confounding factors including sex, age, insurance type, triage level, distance from ED, timing of visit, and ED activity level.

RESULTS

LWCET occurred in 1.73% of 68,461 visits made by 47,228 children. The multivariate model revealed that AIs were more likely to LWCET compared to White children (Odds Ratio (OR) = 1.62, 95% Confidence Interval (CI) = 1.30-2.03). There was no significant difference in LWCET between AA and White children. Other factors significantly associated with LWCET included triage level, distance from the ED, timing of visit, and ED activity level.

CONCLUSION

Our results show that AI children have higher rates of LWCET compared to White children; this association is different from other racial minority groups. There are likely complex factors affecting LWCET in AI children throughout the upper Midwest, which necessitates further exploration.

Based Real Time Remote Health Monitoring Systems: A Review on Patients Prioritization and Related "Big Data" Using Body Sensors information and Communication Technology

The growing worldwide population has increased the need for technologies, computerised software algorithms and smart devices that can monitor and assist patients anytime and anywhere and thus enable them to lead independent lives. The real-time remote monitoring of patients is an important issue in telemedicine. In the provision of healthcare services, patient prioritisation poses a significant challenge because of the complex decision-making process it involves when patients are considered 'big data'. To our knowledge, no study has highlighted the link between 'big data' characteristics and real-time remote healthcare monitoring in the patient prioritisation process, as well as the inherent challenges involved. Thus, we present comprehensive insights into the elements of big data characteristics according to the six 'Vs': volume, velocity, variety, veracity, value and variability. Each of these elements is presented and connected to a related part in the study of the connection between patient prioritisation and real-time remote healthcare monitoring systems. Then, we determine the weak points and recommend solutions as potential future work. This study makes the following contributions. (1) The link between big data characteristics and real-time remote healthcare monitoring in the patient prioritisation process is described. (2) The open issues and challenges for big data used in the patient prioritisation process are emphasised. (3) As a recommended solution, decision making using multiple criteria, such as vital signs and chief complaints, is utilised to prioritise the big data of patients with chronic diseases on the basis of the most urgent cases.

Community Perspectives on Emergency Department Use and Care for American Indian Children

Emergency department (ED) utilization by American Indian (AI) children is among the highest in the nation. Numerous health disparities have been well documented in AI children, but limited information is available on parental experiences of care for AI children in the ED. Our objective was to understand parental attitudes towards ED care for AI children. Focus groups were held with AI parents/caregivers at five sites in the Upper Midwest. Traditional content analysis was used to identify themes. A total of 70 parents participated in ten focus groups. Three main themes were identified: healthcare environment, access to care, and interaction with providers. Healthcare environment issues included availability of specialists, wait times, and child-friendly areas. Transportation and financial considerations were major topics in access to care. Issues in interaction with providers included discrimination, stereotyping, and trust. This is one of the first studies to assess parent perspectives on ED use for AI children. Obtaining parental perspectives on ED experiences is critical to improve patient care and provide important information for ED providers.