Differential patterns of urban and rural cardiovascular emergency department utilization and clinical outcomes

Introduction

Cardiovascular disease (CVD) is the leading cause of death in the United States (US). Rural US residents (nearly 60 million people) have higher overall and CVD-specific mortality (1). Rural populations also have significant barriers in underinsurance, cardiovascular comorbidities, healthcare system differences, and social determinants of health (1).

Increased rural CVD mortality contributes to the 20% rural-urban disparity in reported all-cause mortality (2, 3). Recent studies found urban-rural differences in mortality among specific CVD diseases, including an increased mortality in patients at rural hospitals with acute myocardial infarction (4), heart failure (4, 5), and stroke (4, 6). However, limited data exist that compare patient characteristics and outcomes of CVD-specific emergency department (ED) visits by patient urbanicity.

The American Heart Association has recommended state-level solutions to rural cardiovascular disparities (1). States with a higher proportion of residents living in rural areas are more likely to have higher CVD morbidity and mortality compared to urban residents, driven by factors such as limited healthcare access, higher poverty rates, and greater social risk factors (7). Nebraska has one of the highest proportions of its population living in rural areas among US states. Nationally, about 15% of the US population resides in rural counties, whereas 30%−35% of Nebraska's population live in rural areas (8). Nebraska's high rural population share makes it a critical setting for targeted research and interventions to address rural CVD disparities. The aim of this study was to identify patterns in patient characteristics and outcomes of ED visits for five common CVD diagnoses: acute myocardial infarction, heart failure, hypertension, atrial fibrillation, and stroke. These data may inform healthcare delivery and population health initiatives and lead to future improvements in healthcare disparities in rural areas.

Methods

Data source

ED visits were identified in the 2016–2020 Nebraska State Inpatient Database (SID) and State Emergency Department Database (SEDD). The SID and SEDD are deidentified and Health Insurance Portability and Accountability Act of 1996 (HIPAA)-compliant administrative databases within the Healthcare Cost and Utilization Project (HCUP) (9). The SID and SEDD include discharge information on all ED visits; the SID captures ED visits that result in admission to the treating facility and the SEDD captures ED visits that do not result in admission. The Institutional Review Board at Creighton University (InfoEd Record Number: 2004218) acknowledged the study as Not Human Subjects Research.

Cohort identification

ED visits were identified using the primary discharge diagnosis. ED visits with a primary International Classification of Diseases, Tenth Revision diagnosis codes for one of five common CVD diagnoses in patients 18 years and older were included: myocardial infarction (I21), heart failure (I50), atrial fibrillation (I48), hypertension (I16, I10, I12, I13), and stroke (I63). Patients carrying only a secondary CVD diagnosis were not included in the study so as to ensure the primary reason for ED visit was consistent across the patient cohort and that outcomes reflected acute CVD presentations rather than secondary findings.

Outcomes

The primary aim was to assess whether ED visit outcomes differed by patient urbanicity (urban, micropolitan, rural). Outcomes included inpatient mortality, length of stay (LOS), total cost, and 30- and 90-day ED revisits. Cost was calculated using hospital charges and cost-to-charge ratios provided by the HCUP database. Additionally, cost was inflation-adjusted to July 2020 US dollars. Patient urbanicity was defined by the US Department of Agriculture's Rural Urban Continuum Codes (RUCC). Urban was defined as patient residing in a metropolitan county with a RUCC of 1–3. Micropolitan was defined as a patient residing in a non-metropolitan urbanized county with a RUCC of 4–7. Rural was defined as a nonmetropolitan, nonurbanized county with a RUCC of 8–9 (10).

Statistical analysis

Descriptive characteristics were stratified by patient urbanicity. Categorical descriptives were presented as percentages and compared using chi-squared tests. Continuous descriptives were presented as median and interquartile range and compared using regression. To evaluate whether the odds of inpatient mortality and 30- and 90-day ED revisits differed by patient urbanicity, logistic regression models were estimated. Similarly, to evaluate whether LOS and cost differed by patient urbanicity, lognormal models were estimated. For all outcomes, multivariable models were estimated to account for confounding. More specifically, multivariable models adjusted for CVD diagnosis, age, sex, insurance, income quartile, and comorbidity burden. Comorbidity burden was quantified using the Charlson Comorbidity Index (CCI) (11).

All models estimated a random intercept to account for the inherent correlation of patients treated at the same facility. Outcomes were assessed separately for ED visits that resulted in admission to the treating facility (i.e., SID) and ED visits that did not result in admission (i.e., SEDD). Relatedly, inpatient-specific outcomes (mortality and LOS) were only evaluated for ED visits where the patient was admitted to the presenting hospital. Inpatient-specific outcomes could not be assessed for patients who visited the ED but were not admitted to the presenting hospital. The SEDD and SID only follow patients for a calendar year; to allow for complete follow-up in 30- and 90-day revisit analyses, we excluded patients who died on the index visit or if the patient was discharged within 30 or 90 days of the end of the calendar year. The SEDD and SID do not assess out-of-hospital deaths, so we were unable to assess 30- and 90-day mortality.

Analyses were conducted used SAS version 9.4. Two-tailed p < 0.05 was considered statistically significant.

Results

From 2016 to 2020, there were 47,369 CVD ED visits that met inclusion criteria of a primary discharge diagnosis, of which 25,190 resulted in hospital admission. Of note, 2.38% of CVD ED visits meeting inclusion criteria were patients transferred from a different acute care hospital. Table 1 presents the frequency of ED visits by CVD diagnosis and urbanicity. Hypertension was the most common CVD. Urban patients were younger, carried a greater comorbidity burden, had more private insurance, were more frequently in the highest income quartile, and had fewer transfers from the ED (Table 2).

Table 1. Nebraska emergency departments visits for cardiovascular disease by patient urbanicity.

Table 2. Descriptives stratified by patient urbanicity.

IQR, interquartile range.

An ^* indicates the results could not be presented per the HCUP Data Use Agreement. Continuous descriptives were presented as median and interquartile range and compared using regression. Categorical descriptives were presented as percent and compared using chi-square tests. Discharge Dispostion “Other” included patients that left against medical advice, were discharged to home health care, died, or had an unknown discharge.

Outcomes for ED visits that resulted in admission to the presenting hospital are shown in Table 3A. Micropolitan and rural patients had 60 and 97% greater adjusted odds of inpatient mortality compared to urban patients, respectively [adjusted odds ratio (aOR): 1.60, 95% CI: 1.18–2.17, p = 0.003; aOR: 1.97, 95% CI: 1.34–2.89, p < 0.001]. Compared to urban patients, micropolitan patients had 10% increased cost, and rural patients had 13% increased cost (aOR: 1.10, 95% CI: 1.05–1.14, p < 0.001; aOR: 1.13, 95% CI: 1.07–1.20, p < 0.001). Micropolitan patients had 18% lower adjusted odds of an ED revisit within 30 days and 15% lower adjusted odds of an ED revisit within 90 days compared to urban patients (aOR: 0.82, 95% CI: 0.73–0.93, p = 0.002; aOR: 0.85, 95% CI: 0.76–0.96, p = 0.008). Rural patients had 25% lower adjusted odds of an ED revisit within 30 days and 40% lower adjusted odds of an ED revisit within 90 days (aOR: 0.75, 95% CI: 0.61–0.92, p = 0.006; aOR: 0.60, 95% CI: 0.50–0.74, p < 0.001) compared to urban patients. Compared to micropolitan patients, rural patients had 29% lower adjusted odds of an ED revisit within 90 days (aOR: 0.71, 95% CI: 0.58–0.86, p < 0.001).

Table 3. Outcomes for patients who visited the emergency department and were (A) admitted or (B) not admitted to their treating (presenting) facility.

A. 95% confidence intervals (95% CI) that omit 1 indicate statistical significance.

B. The group following “vs.” served as reference.

Outcomes for ED visits that did not result in admission are presented in Table 3B. Micropolitan and rural patients had 9 and 13% greater adjusted cost compared to urban patients, respectively (aOR: 1.09, 95% CI: 1.02–1.16, p = 0.013; aOR: 1.13, 95% CI: 1.06–1.22, p < 0.001). Rural patients had 24% lower odds of an ED revisit within 90 days compared to urban patients (aOR: 0.76, 95% CI: 0.64–0.90, p = 0.002). Rural patients had 15% lower odds of an ED revisit within 90 days compared to micropolitan patients (aOR: 0.85, 95% CI: 0.73–0.98, p = 0.029).

Discussion

Significant differences were identified in urban-rural CVD ED visit patient characteristics and outcomes in Nebraska. Most notably, for ED visits resulting in hospital admission, micropolitan and rural patients had 60 and 97% greater adjusted odds of inpatient mortality compared to urban patients, respectively. Urban patients had contrasting demographic, financial, and transfer rate findings compared to micropolitan and rural patients. Hospital cost was higher for micropolitan and rural patients, and rural patient revisits to the ED were lower for 30 and 90 days after the initial visit.

In the US, the rate of rural ED visits has increased, with nearly 20% of all ED visits occurring in rural settings (12). Specifically, from 2005 to 2016, rural ED visit rates increased by 50% (12). Furthermore, there was a 26.7% increase in rural EDs functioning as safety-net hospitals from 2005 to 2016 (12). The increasing proportion of rural EDs serving as safety-net hospitals for disadvantaged populations may be associated with further worsening of rural CVD outcomes and access to healthcare services. More than a third of the population in NE resides in rural designated areas, and the majority of counties in the state are rural. Less or inequitable access to healthcare in rural and micropolitan areas may explain the worse CVD outcomes, higher transfer rates, and lower 30- and 90-day ED revisit rates.

Inpatient mortality was greater in micropolitan (60%) and rural (97%) patients compared to urban patients, consistent with prior studies finding higher rural CVD mortality (4–6). Loccoh et al. (4) demonstrated higher mortality among patients presenting to rural hospitals for myocardial infarction, heart failure, and stroke. CVD mortality differences may be impacted by limited access and services in rural areas. The barrier of healthcare access worsened as rural hospital closure rates increased in the last decade (13). Since 2005, more than 190 US and two NE rural hospitals closed (14). A recent study showed that from 2010 to 2019 one NE rural hospital closure resulted in 1,648 residents being unable to access a hospital within 15 min and 357 residents unable to access a hospital within 30 min (13). This delay in hospital access could result in lower quality or less time-sensitive treatments due to delayed treatment. In fact, Loccoh et al. (4) showed that rural patients were less likely to undergo cardiac catheterization, percutaneous coronary intervention, coronary artery bypass graft, thrombolysis, and endovascular therapy. Another factor contributing to differential CVD outcomes by patient urbanicity is cardiologist density. Cardiologist density is lower in rural areas, potentially impacting the type, quality, and timing of treatment (15). The combination of increased rural hospital closures and increased barriers to receiving important acute treatments (4, 6, 16) in rural areas may explain differences in rural CVD mortality.

Of patients admitted to the hospital from the ED, micropolitan ($13,691) and rural ($14,719) patients had higher average costs than urban ($12,406) patients. Similarly, of patients who visited the ED and were not admitted, micropolitan ($991) and rural ($1,091) had higher average costs than urban patients ($823). These results were similar to a prior study that found that rural residents with incident atrial fibrillation had higher ED costs and fewer outpatient physician visits but lower inpatient costs (13). Notably, our study found greater overall cost for rural patients admitted to the presenting hospital from the ED (i.e., ED cost and inpatient cost). The combination of ED cost and inpatient cost provides a more comprehensive picture of the financial burden for admitted patients. Our study expands previous knowledge by finding this cost-related disparity among a broader CVD population. Moreover, the true economic disparity between urban patients and patients residing in micropolitan and rural areas may be even wider than reported in our study. In our study, cost represented charges incurred during the ED visit and subsequent admission (if admitted). Cost did not include transport/ambulance fees or other fees incurred outside the treatment facility.

Rural patients had lower odds of a 30- and 90-day ED revisit compared to micropolitan and urban patients. This contrasts with findings from Zeitler et al. (5) where the 30-day ED revisit rate was higher in rural patients with heart failure between 2008 and 2017. Importantly, this study included a broader CVD population and data from a different region. Recent changes in rural access to healthcare services for urgent conditions (1), increased rural hospital closures (13, 14), and increased time to get to hospitals in rural areas (13) may explain the difference in findings.

Rural EDs often serve as primary access points for care. The lower ED revisit rates observed in rural areas may reflect adverse factors, including the competing risk of out-of-hospital mortality, which is not captured in the SID/SEDD databases. If rural patients die at home due to barriers in accessing care, these events would not be recorded as revisits to the ED.

Limitations of this review include that CVD diagnoses were assessed as one cohort due to sample size limitations, thereby limiting information about any specific primary diagnosis. Importantly, all multivariable regression analyses adjusted for the type of CVD diagnosis. The exclusion of hospitalizations with any nonprimary CVD diagnosis resulted in a smaller sample size. For example, hypertension was noted as a diagnosis in 649,752 cases. Notably, the combined data provide generalizable conclusions about CVD ED urban-rural variation in NE. These NE-specific results may not be generalizable to other states and regions. The limited availability of stroke centers and specialty care in rural and micropolitan areas may have contributed to adverse stroke outcomes. Inpatient outcomes (mortality and LOS) could not be assessed for patients who visited the ED and were discharged or transferred to other facilities. Nevertheless, the higher inpatient mortality in more rural hospitals is noteworthy as transfers often occur when higher levels of care are needed, so transfers likely represent sicker patients. If death occurred after transfer, these events would not be reflected in the analysis. Since rural ED visits of critically ill patients are associated with more transfers than urban ED visits, rural mortality may be higher than reported in these data. Also, the CCI may not capture the severity of illness and burden of comorbidity in their entirety. Our data are from administrative claims databases (i.e., SID and SEDD) that use administrative billing codes to identify diagnoses (i.e., ICD-10-CM). These administrative billing codes often lack clinical information to assess illness severity. The data set available at the time of the project was only as recent as 2020.

Conclusion

Differences in mortality, ED revisits, and cost exist between urban, micropolitan, and rural patients presenting to the ED for major cardiovascular primary discharge diagnoses. These findings inform and provide insight into patient social risk factors and potential unmet needs. Future research should focus on assessing access to timely and quality care, care coordination, and overall improved CVD outcomes in rural populations.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Korakot Apiratwarakul, Khon Kaen University, Thailand

Reviewed by: Eashwar Krishna, University of Connecticut, United States

Adil Mohammed, Central Michigan University, United States