Abstract
Objectives
The goal of this study was to examine 2006 to 2010 emergency department (ED) admission rates, hospital procedures, lengths of stay, and costs for acute heart failure (AHF).
Background
Patients with AHF are often admitted and are associated with high readmissions and cost.
Methods
We utilized Nationwide Emergency Department Sample AHF data from 2006 to 2010 to describe admission proportion, hospital length of stay (LOS), and ED charges as a surrogate for resource utilization. Results were compared across U.S. regions, patient insurance status, and hospital characteristics.
Results
There were 958,167 mean yearly ED visits for AHF in the United States. Fifty-one percent of the patients were female, and the median age was 75.1 years (interquartile range [IQR]: 62.5 to 83.7 years). Overall, 83.7% (95% confidence interval: 83.1% to 84.2%) were admitted; the median LOS was 3.4 days (IQR: 1.9 to 5.8 days). Comparing 2006 with 2010, there was a small decrease in median LOS (0.09 days), but the proportion admitted did not change. Odds of admission, adjusting for age, sex, hospital characteristic (academic and safety net status), and insurance (Medicare, Medicaid, private, self-pay/no charge) were highest in the Northeast. Median ED charges were $1,075 (IQR: $679 to $1,665) in 2006 and $1,558 (IQR: $1,018 to $2,335) in 2010. Patients without insurance were more likely to be discharged from the ED, but when admitted, were more likely to receive a major diagnostic or therapeutic procedure.
Conclusions
A very high proportion of ED patients with AHF are admitted nationally, with significant variation in disposition and procedural decisions based on region of the country and type of insurance, even after adjusting for potential confounding.
Keywords: emergency medicine, heart failure, resource utilization
Heart failure affects approximately 5 million Americans (1), results in nearly 1 million annual hospital stays (1,2), and is the top reason for Medicare hospital readmissions (3,4). The vast majority of patients hospitalized for acute heart failure (AHF) are originally evaluated and managed in the emergency department (ED). Prior data suggest more than 80% of ED patients with AHF are admitted to the hospital and have a median inpatient length of stay (LOS) of approximately 3.4 days. Of the $39.2 billion dollars spent on heart failure care in the United States in 2010, hospital stay was the single largest proportion of this expenditure (5,6). Among Medicare beneficiaries, hospital stay accounts for more than 50% of all heart failure costs in the last 6 months of life (7). Despite a small decline in the AHF hospital stay rate among Medicare beneficiaries over the last decade (1,2), mortality remains high (2,8,9) and uneven across states (1).
Although heart failure has largely been defined as chronic debilitation, most patients will, at some point, require emergency care for acute symptoms (10). Unfortunately, the U.S. emergency care system faces compelling challenges as patient visits rise and the number of EDs decrease (11–16). In 2010, American Heart Association statistics reported nearly 668,000 annual ED encounters for AHF (1), representing approximately 20% of the total heart failure–specific ambulatory care delivered each year (17). Few settings beyond the ED can provide the care intensity required to initially stabilize, manage, and risk stratify patients with AHF. The ED plays a crucial role in the continuum of heart failure care, is frequently a patient's point of first contact when worsening symptoms require urgent or emergent care, and is largely where admission or discharge decisions are made (18).
It is important to quantify the burden of AHF on U.S. emergency care in order to focus individual and health system strategies to address the challenges before they become overwhelming. We describe national as well as regional-, hospital-, and patient-level variations with regard to ED admission and discharge tendencies, frequency of major diagnostic and therapeutic procedures, and overall healthcare resource utilization. We explored these questions using the Nationwide Emergency Department Sample (NEDS) database, the largest publicly available all-payer ED database in the United States.
Methods
Study design and data sources
We conducted a retrospective cohort study using data from the NEDS database from 2006 through 2010. NEDS is a component of the Healthcare Cost and Utilization Project (HCUP) sponsored by the Agency for Healthcare Research and Quality (19). It represents the largest all-payer ED database in the United States and was created to enable analyses of ED utilization and support health professionals, administrators, policymakers, and clinicians in their decision making regarding this source of care. The institutional review board at Vanderbilt University approved this as a nonhuman study.
NEDS was constructed using records from both the HCUP State Emergency Department Databases and the State Inpatient Databases. The State Emergency Department Databases capture information on ED visits that do not result in an admission (i.e., treat-and-release visits and transfers to another hospital). The State Inpatient Databases contain information on patients initially seen in the ED and then admitted to the same hospital. The NEDS contains between 25 and 30 million (unweighted) records of ED visits for more than 950 hospitals and approximates a 20% stratified sample of U.S. hospital–based EDs. Taken in total, NEDS thus represents all types of ED visits, regardless of ultimate ED disposition.
NEDS stratifies data from participating states by geographic region (Northeast, Midwest, South, and West), trauma center designation, urban/rural status, teaching hospital status (academic vs. nonacademic), and ownership (public vs. private). The NEDS database includes weights for calculating national estimates. For each ED encounter, NEDS reports up to 15 diagnoses coded according to their respective International Classification of Diseases-Ninth Revision-Clinical Modification (ICD-9 CM) codes.
Study setting and population
ED visits were included in our analysis if patients were at least 18 years old and carried a primary ED diagnosis of AHF (ICD-9 CM code 428.x). Patients were excluded if they had cardiogenic shock (ICD-9-CM code 785.51) or unspecified shock (ICD-9-CM code 785.5), were intubated (Current Procedural Terminology [CPT] code 31500), on noninvasive ventilation (CPT code 94660), or had an acute myocardial infarction (ICD-9CM codes 410.0 to 410.9). We excluded these patients because they would not be eligible for ED discharge, one of the main outcomes in this study.
Patient- and ED-level variables
The NEDS contains information on month and year of evaluation, patient demographics, patient disposition from the ED, ICD-9CM and CPT procedures, median household income for a patient's zip code, expected payment source (e.g., Medicare, Medicaid, private insurance, self-pay), total ED and hospital charges, and hospital characteristics (region, trauma center indicator, urban/rural location, teaching status). Geographic regions were defined according to U.S. census rule boundaries (20).
Insured patients were defined as those with any type of insurance (private, Medicare, or Medicaid). Uninsured patients were those classified as self-pay or no charge. An academic hospital was defined as a metropolitan teaching institution, whereas a community hospital was defined as a metropolitan nonteaching or nonmetropolitan institution. A safety net hospital was defined based on previous definitions (21) as any of the following: 1) >30% of ED visits are Medicaid; 2) >30% of visits are with self-pay as the source of payment; or 3) combined Medicaid and uninsured pool is >40%.
Major diagnostic and therapeutic procedures were defined using Clinical Classification Software (CCS) codes. CCS codes were developed at the Agency for Healthcare Research and Quality as a tool for clustering patient diagnoses and procedures into a manageable number of clinically meaningful categories, offering researchers the ability to group conditions and procedures without having to sort through thousands of codes. CCS collapses diagnosis and procedure codes from the ICD-9-CM, which contains more than 14,000 diagnosis codes and 3,900 procedure codes. Major diagnostic procedures considered in this study included CCS codes 193, 47, 204, 48, and 201 (echocardiography, diagnostic cardiac catheterization, pulmonary artery catheterization for monitoring, cardiac pacemaker or cardioverter/defibrillator, cardiac stress test); similarly, major therapeutic procedures included CCS codes 45 and 46 (percutaneous transluminal coronary angioplasty, coronary thrombolysis).
Outcome measures
We utilized data across 5 years (2006 to 2010) to describe and compare: 1) proportion of ED patients admitted to the hospital; 2) proportion of major diagnostic and therapeutic procedures in those patients who were hospitalized; 3) LOS during hospital stay; and 4) total ED and hospital charges, as a surrogate for resource utilization.
Analysis
Proportions and weighted frequencies, and medians and interquartile ranges (IQRs) were computed from a weighted analysis accounting for the NEDS sampling design (22). In addition, 2 multivariable logistic regression models were fit to assess the national and regional odds of admission with several a priori–selected patient- and hospital-level characteristics. Both models accounted for the NEDS sampling design for appropriate estimation of the SE and adjusted for age, sex, hospital characteristics (academic status and safety net status), and insurance status (Medicare, Medicaid, private, self-pay/no charge) (23). The primary model was fit, treating region as a subpopulation, to estimate regional odds of admission for each of the covariates of interest. A secondary model included hospital region as a covariate in order to obtain national estimates of the odds of admission for each of the covariates of interest plus an estimate of the odds of admission for region. For completeness, an additional model similar to the secondary model was also fit with the same covariates, except insurance status was dichotomized to insured (Medicare, Medicaid, private) versus uninsured (self-pay, no charge). All analyses were done using the statistical programming language SAS for Linux (24).
Results
There were 958,167 average yearly weighted ED visits for AHF meeting our inclusion criteria, representing 0.77% of all ED visits. Characteristics for these visits are presented in Table 1. Ignoring our exclusion criteria slightly increased the average yearly weighted ED visits to 963,770 (28,220 total were excluded, or 0.58% of the entire cohort). Overall, 83.7% (95% confidence interval [CI]: 83.1% to 84.2%) were admitted, and the median hospital LOS was 3.4 days (IQR: 1.9 to 5.8 days). Comparing 2006 with 2010, there was a small increase in the proportion of ED visits that resulted in hospital admission (0.13%), a small decrease in median hospital LOS (0.09 days), as well as an increase in median ED charges ($1,075 to $1,558) and combined median ED and hospital charges ($16,990 to $21,088).
Table 1. ED Visits for AHF 2006 to 2010.
| 2006 | 2007 | 2008 | 2009 | 2010 | |
|---|---|---|---|---|---|
| ED visits | 988,915 (0.82) | 941,682 (0.77) | 938,823 (0.75) | 959,167 (0.74) | 962,250 (0.75) |
|
| |||||
| ED disposition | |||||
| Discharged | 161,703 (16.35) | 149,693 (15.90) | 152,563 (16.25) | 152,313 (15.88) | 146,020 (15.17) |
| Admitted | 795,799 (80.47) | 760,889 (80.80) | 751,709 (80.07) | 775,852 (80.89) | 775,552 (80.60) |
| Transferred | 23,711 (2.40) | 27,184 (2.89) | 31,244 (3.33) | 28,524 (2.97) | 37,957 (3.94) |
| Died in ED | 1,165 (0.12) | 1,086 (0.12) | 1,251 (0.13) | 1,497 (0.16) | 1,299 (0.14) |
| Unknown | 6,536 (0.66) | 2,831 (0.30) | 2,056 (0.22) | 981 (0.10) | 14,22 (0.15) |
|
| |||||
| Hospital disposition | |||||
| Routine | 430,545 (54.16) | 404,261 (53.24) | 396,241 (52.72) | 407,220 (52.50) | 394,521 (50.96) |
| Transfer to acute care facility | 26,422 (3.32) | 24,764 (3.26) | 25,158 (3.35) | 25,081 (3.23) | 24,599 (3.18) |
| SNF/rehabilitation | 163,181 (20.53) | 154,820 (20.39) | 153,214 (20.39) | 159,504 (20.57) | 159,911 (20.66) |
| Home health care | 139,385 (17.53) | 141,423 (18.63) | 144,906 (19.28) | 150,705 (19.43) | 163,947 (21.18) |
| Against medical advice | 9,761 (1.23) | 10,034 (1.32) | 8,741 (1.16) | 9,221 (1.19) | 9,085 (1.17) |
| Died in hospital | 25,303 (3.18) | 23,542 (3.10) | 22,838 (3.04) | 23,536 (3.03) | 21,607 (2.79) |
| Discharged alive, destination unknown | 403 (0.05) | 463 (0.06) | 461 (0.06) | 345 (0.04) | 449 (0.06) |
|
| |||||
| ED charges, US$ | 1,075 (679–1,665) | 1,131 (685–1,778) | 1,283 (836–2,001) | 1,427 (885–2,138) | 1,558 (1,018–2,335) |
|
| |||||
| Total charges for admitted patients, US$ | 16,990 (10,009–31,006) | 17,370 (10,236–31,572) | 18,945 11,188–34,136) | 19,926 (11,766–35,923) | 21,088 (12,598–37,395) |
|
| |||||
| Primary payer | |||||
| Medicare | 746,992 (75.63) | 701,386 (74.63) | 701,611 (74.86) | 716,224 (74.75) | 719,805 (74.91) |
| Medicaid | 75,918 (7.69) | 76,636 (8.15) | 74,378 (7.94) | 78,312 (8.17) | 81,439 (8.48) |
| Private insurance | 105,876 (10.72) | 104,938 (11.17) | 110,107 (11.75) | 112,241 (11.71) | 104,067 (10.83) |
| Self-pay | 37,171 (3.76) | 36,978 (3.93) | 33,326 (3.56) | 32,976 (3.44) | 36,584 (3.81) |
| No charge | 4,224 (0.43) | 2,575 (0.27) | 2,353 (0.25) | 3,804 (0.40) | 2,309 (0.24) |
| Other | 17,456 (1.77) | 17,355 (1.85) | 15,442 (1.65) | 14,586 (1.52) | 16,648 (1.73) |
|
| |||||
| Length of stay for admitted patients, days | 3.5 (2.0–5.9) | 3.5 (1.9–5.9) | 3.5 (2.0–5.9) | 3.4 (1.9–5.8) | 3.4 (1.9–5.7) |
|
| |||||
| Location of care | |||||
| Academic medical center | 384,409 (38.87) | 344,764 (36.61) | 333,060 (35.48) | 355,656 (37.08) | 378,357 (39.32) |
| Safety net | 531,598 (53.76) | 502,558 (53.37) | 510,734 (54.40) | 567,330 (59.15) | 635,489 (66.04) |
|
| |||||
| Age, yrs | 74.9 (62.4–83.3) | 75.2 (62.6–83.6) | 75.2 (62.7–83.7) | 75.2 (62.7–83.8) | 74.9 (62.2–83.9) |
|
| |||||
| Female | 511,764 (51.75) | 486,906 (51.72) | 478,571 (50.98) | 488,560 (50.95) | 486,959 (50.61) |
|
| |||||
| Median household income | |||||
| 1st quartile | 323,215 (33.30) | 314,681 (34.23) | 297,472 (32.51) | 302,351 (32.25) | 309,258 (32.84) |
| 2nd quartile | 248,265 (25.58) | 241,218 (26.24) | 262,928 (28.73) | 265,012 (28.27) | 257,292 (27.32) |
| 3rd quartile | 222,059 (22.88) | 206,041 (22.41) | 186,794 (20.41) | 207,172 (22.10) | 202,511 (21.51) |
| 4th quartile | 177,142 (18.25) | 157,377 (17.12) | 167,912 (18.35) | 162,895 (17.38) | 172,606 (18.33) |
|
| |||||
| Clinical characteristics | |||||
| Chronic hypertension | 220,982 (22.35) | 206,776 (21.96) | 198,821 (21.18) | 184,744 (19.26) | 175,820 (18.27) |
| Chronic diabetes | 159,932 (16.17) | 147,393 (15.65) | 143,688 (15.31) | 138,446 (14.43) | 130,624 (13.57) |
| Chronic coronary artery disease | 207,853 (21.02) | 196,410 (20.86) | 213,209 (22.71) | 199,975 (20.85) | 184,171 (19.14) |
| Chronic CKD | 3,249 (0.33) | 2,675 (0.28) | 2,688 (0.29) | 2,327 (0.24) | 2,282 (0.24) |
Values are weighted n (%) or median (interquartile range).
AHF = acute heart failure; CKD = chronic kidney disease; ED = emergency department; SNF = skilled nursing facility; US$ = U.S. dollars.
Regional variations
Hospital stay proportions (Fig. 1) were higher in the Northeast (89.7%, 95% CI: 89.0% to 90.4%) than in the South (83.2%, 95% CI: 82.2% to 84.2%), Midwest (82.6%, 95% CI: 81.6% to 83.6%), and West (79.2%, 95% CI: 77.9% to 80.6%). However, even though ED patients with AHF in the Northeast were consistently slightly older (median 77.9 years vs. 74.3 years) than the other regions, the odds of admission by region followed similar trends in the secondary multivariable model, even after adjusting for age, hospital characteristic (academic and safety net status), and insurance type (Medicare, Medicaid, private, self-pay/no charge) (Table 2). ED patients with AHF in the Northeast underwent a longer inpatient median LOS (3.9 days, IQR: 2.2 to 6.6 days vs. 3.3 days, IQR: 1.9 to 5.6 days) when compared with other regions.
Figure 1. Hospital Admission (%) by Region, Academic Status, and Safety Net Status.
Table 2. Results From the Primary and Secondary Logistic Regression Models.
| Covariate | Region | Primary Model | Secondary Model | ||||
|---|---|---|---|---|---|---|---|
|
|
|
||||||
| Adjusted OR | (95% CI) | p Value | Adjusted OR | (95% CI) | p Value | ||
| Insurance status | |||||||
| Medicare (ref) | 1.00 | 1.00 | |||||
| Medicaid | National | 1.00 | (0.96–1.05) | 0.8818 | |||
| Northeast | 1.02 | (0.90–1.15) | 0.7575 | ||||
| Midwest | 0.89 | (0.81–0.98) | 0.0227 | ||||
| South | 0.96 | (0.89–1.02) | 0.1853 | ||||
| West | 1.15 | (1.04–1.27) | 0.0063 | ||||
| Private insurance | National | 0.81 | (0.78–0.85) | <0.0001 | |||
| Northeast | 0.61 | (0.53–0.71) | <0.0001 | ||||
| Midwest | 0.91 | (0.85–0.96) | 0.0015 | ||||
| South | 0.82 | (0.78–0.86) | <0.0001 | ||||
| West | 0.87 | (0.79–0.95) | 0.0032 | ||||
| Self-pay/no charge | National | 0.61 | (0.57–0.65) | <0.0001 | |||
| Northeast | 0.61 | (0.50–0.73) | <0.0001 | ||||
| Midwest | 0.73 | (0.65–0.81) | <0.0001 | ||||
| South | 0.63 | (0.57–0.70) | <0.0001 | ||||
| West | 0.52 | (0.44–0.61) | <0.0001 | ||||
|
| |||||||
| Year (per year) | National | 1.02 | (1.00–1.04) | 0.0530 | |||
| Northeast | 1.09 | (1.03–1.14) | 0.0013 | ||||
| Midwest | 1.06 | (1.02–1.10) | 0.0016 | ||||
| South | 0.99 | (0.96–1.02) | 0.6239 | ||||
| West | 1.00 | (0.96–1.03) | 0.7968 | ||||
|
| |||||||
| Age (per year) | National | 1.01 | (1.009–1.011) | <0.0001 | |||
| Northeast | 1.01 | (1.01–1.01) | <0.0001 | ||||
| Midwest | 1.01 | (1.00–1.01) | <0.0001 | ||||
| South | 1.01 | (1.01–1.02) | <0.0001 | ||||
| West | 1.01 | (1.01–1.01) | <0.0001 | ||||
|
| |||||||
| Academic hospital | National | 1.58 | (1.44–1.74) | <0.0001 | |||
| Northeast | 1.97 | (1.65–2.34) | <0.0001 | ||||
| Midwest | 1.82 | (1.56–2.14) | <0.0001 | ||||
| South | 1.53 | (1.28–1.83) | <0.0001 | ||||
| West | 1.20 | (1.00–1.44) | 0.0529 | ||||
|
| |||||||
| Safety net hospital | National | 0.91 | (0.83–0.99) | 0.0290 | |||
| Northeast | 0.74 | (0.62–0.90) | 0.0018 | ||||
| Midwest | 0.94 | (0.83–1.06) | 0.3106 | ||||
| South | 0.82 | (0.67–0.99) | 0.0386 | ||||
| West | 1.07 | (0.93–1.24) | 0.3581 | ||||
|
| |||||||
| Region | |||||||
| West (ref) | 1.00 | ||||||
| Northeast | 1.95 | (1.73–2.20) | <0.0001 | ||||
| Midwest | 1.16 | (1.04–1.30) | 0.0070 | ||||
| South | 1.35 | (1.20–1.52) | <0.0001 | ||||
The primary model reports the associations between the covariates and admission versus discharge from the ED at the regional level. The secondary model reports the associations between the covariates and the outcome at the national level, adjusting for hospital region.
CI = confidence interval; OR = odds ratio; ref = reference variable; other abbreviations as in Table 1.
Variability in admission, LOS, and procedures based on hospital- and patient-level variables
Compared with nonacademic hospitals, patients at academic hospitals tended to be younger (academic: 73 years, IQR: 59 to 83 years; nonacademic: 76 years, IQR: 64 to 84 years), but the proportion admitted was significantly higher (87.34% vs. 81.47%, p < 0.0001). This observation was retained in the multivariable model at the national and regional levels after adjusting for age, hospital characteristics (academic and safety net status), and insurance status (Medicare, Medicaid, private, self-pay/no charge). The national-level adjusted odds of admission at an academic hospital were 1.58 times those at a nonacademic hospital (95% CI: 1.44 to 1.74). Median hospital LOS at academic hospitals was slightly longer than at nonacademic hospitals throughout all 5 years of analysis (Table 3). Likewise, overall hospital charges tended to be higher at academic institutions (Table 4). Compared with non-safety net hospitals, safety net hospitals had lower proportions of ED patients with AHF admitted; however, this difference was not significant in the multivariable model (adjusted odds ratio: 0.91, 95% CI: 0.83 to 0.99). Finally, similar hospital LOS, and ED and total hospital charges were seen in safety net and non-safety net hospitals.
Table 3. LOS Based on Academic Hospital and Insurance Status.
| Covariate | 2006 | 2007 | 2008 | 2009 | 2010 |
|---|---|---|---|---|---|
| Academic status | |||||
|
| |||||
| Academic | 3.6 (2.0–6.2) | 3.6 (2.0–6.3) | 3.6 (2.0–6.2) | 3.6 (2.0–6.2) | 3.5 (2.0–6.1) |
| Nonacademic | 3.4 (1.9–5.7) | 3.4 (1.9–5.6) | 3.4 (2.0–5.7) | 3.3 (1.9–5.6) | 3.2 (1.9–5.5) |
|
| |||||
| Insurance status | |||||
|
| |||||
| Insured | 3.5 (2.0–5.9) | 3.5 (2.0–5.9) | 3.5 (2.0–5.9) | 3.5 (2.0–5.8) | 3.4 (1.9–5.8) |
| Noninsured | 2.9 (1.5–5.0) | 2.9 (1.5–5.0) | 2.9 (1.6–5.0) | 2.9 (1.6–5.0) | 2.8 (1.6–4.9) |
Values are median (interquartile range). Length of stay (LOS) is for admitted patients.
Table 4. Total Charges Based on Academic Hospital.
| Academic Status | 2006 | 2007 | 2008 | 2009 | 2010 |
|---|---|---|---|---|---|
| Academic | 17,810 (10,486–33,307) | 18,625 (11,044–34,243) | 20,482 (12,024–37309) | 21,715 (12,757–39,459) | 22,544 (13,659–40,397) |
| Nonacademic | 16,467 (9,685–29,644) | 16585 (9,792–29,942) | 18,071 (10,732–32,215) | 18,904 (11,179–33,753) | 20,054 (11,893–35,503) |
Values are median (interquartile range). The charges are for admitted patients and are reported as U.S. dollars.
Compared with insured patients across all 5 years of analysis, those who were uninsured were considerably younger (median 52.1 years vs. 76.0 years), more often male (65.3% vs. 47.8%), more often had chronic hypertension (33.6% vs. 20.0%) and diabetes mellitus (16.5% vs. 14.9%), and were less likely to have coronary artery disease (15.7% vs. 21.1%) (p < 0.0001 across all comparisons). Patients without insurance (Fig. 2) were significantly more likely to be discharged from the ED (25.4% vs. 15.4%, p < 0.001), even after adjusting for potential confounders in the multivariable model (adjusted odds ratio for admission among uninsured compared with insured patients 0.641, 95% CI: 0.601 to 0.683). When admitted, uninsured patients were more likely to be the recipients of a major diagnostic (21.6% vs. 12.6%) or major therapeutic (0.91% vs. 0.74%) procedure (p < 0.001 and p < 0.01 for major diagnostic and major therapeutic procedures, respectively) (Fig. 2). Of the major diagnostic procedures reported, 45.8% were for echocardiography.
Figure 2. Hospital Admission, Major Diagnostic Procedures, and Major Therapeutic Procedures by Insurance Status.
Discussion
Our results suggest the ED burden of AHF remains compelling. Although there is regional variability, the vast majority of ED patients with AHF are admitted to the hospital, and the median LOS remains more than 3 days. The proportion of patients hospitalized and the LOS remained relatively unchanged between 2006 and 2010.
Regional variability in admission decisions may have important implications for healthcare expenditures. Although the Northeast has increased odds of admissions compared with other regions, their cumulative bed days over a 30-day period and resource utilization should be compared with other regions. Perhaps it is more efficient to manage a higher proportion of patients in the hospital rather than having them present to the ED multiple times, or experience a downstream complication and admission in close proximity to an ED discharge. Discrepancies in admissions between academic and nonacademic hospitals can likely be understood based on the increased complexity of the case mix at tertiary care hospitals. Further scrutiny of these issues will be especially important with the Affordable Care Act (25) and as penalties are enforced for 30-day readmissions (26).
A decrease in the admission rate plus significant state-to-state variability was reported in Medicare beneficiaries with heart failure from 1998 to 2008 (2). As the U.S. population becomes older, hospital stay rates for AHF would be expected to rise (2). Mitigating factors may be decreases in or better treatment of precursors to heart failure such as ischemic heart disease (27,28) and hypertension (29,30), as well as increased use of secondary prevention therapies (31,32), better overall management of risk factors (2), and changes in the threshold of admission in EDs (33). Although we found variability geographically, we did not see a significant decrease in admission rates. Our results are based on 5 years of data and include all ED patients with AHF, regardless of age and insurance status. An earlier 1993 to 2006 Medicare analysis reported reductions in mean LOS, lower in-hospital mortality, and to a lesser extent, reductions in 30-day mortality, unfortunately accompanied by increased 30-day readmission rates (17.2% to 201.1%) (34). Although Medicare inpatients cannot be directly compared with our population, this change in 30-day readmissions, paralleling a decrease in LOS, potentially has substantial ED ramifications. An increase in readmissions is likely to place an additional burden upon the emergency care system and the need for more attention to the early transition period from hospital to home.
Despite a significantly greater proportion of some comorbidities, uninsured patients, although younger, are admitted less frequently than those with insurance. Once admitted, they undergo more diagnostic and therapeutic procedures. These findings have several implications as we implement national quality-of-care initiatives and set our priorities for AHF research in the ED (35). Perhaps a large proportion of uninsured patients may be utilizing the ED for primary care and may therefore be less clinically ill when they present with AHF. If this were true, the uninsured may be more amenable to discharge following a period of treatment and observation in the ED. Alternatively, there also may be a concerted effort to manage the uninsured as outpatients, both as a result of lower disease acuity and increased amount of unfunded care as a result of an inpatient admission when compared with ED discharge. Patient-specific variables and results of subsequent testing and clinical course over the 30-day period after ED evaluation would be necessary to further distill these issues.
The U.S. emergency care system affects every American and faces broad challenges as patient visits rise and the number of EDs decrease (11–16). The total number of ED visits increased 34% between 1995 and 2010 (from 97 to 130 million). Accounting for changes in population, this is a 16% increase, from 37 to 43 visits per 100 persons (36). In addition, the total number of U.S. EDs declined from 1995 to 2010 by 11%, to 3,700. Despite U.S. inpatient admissions between 2003 and 2009 growing at a slower rate than the population overall, nearly all the growth in admissions was due to a 17% increase from EDs (36). This growth more than offsets a 10% decrease in admissions from outpatient settings. Similarly, analysis of the 1993 to 2006 Nationwide Inpatient Sample (18) revealed a 15.0% increase in total U.S. admissions, but a striking 50.4% increase from the ED. The proportion of all admissions originating from the ED increased from 33.5% to 43.8%.
These data strongly suggest the ED is increasingly being relied on to evaluate more complex patients, some of whom previously received care as outpatients, or were directly admitted from an outpatient setting. The easy availability of quick advanced diagnostics, such as point-of care testing and imaging, as well as the availability of high-quality care 24 h a day, may all play a role (18). As a result, emergency physicians increasingly serve as the major decision makers for approximately one-half of all U.S. inpatient admissions (36). In addition, emergency physician's tolerance of “risk” in relation to ED discharge decisions, especially for patients they do not have an ongoing relationship with, is likely lower than for other caregivers. For AHF, a <0.5% risk of 30-day death or serious nonfatal complication has been suggested as an emergency physician's threshold for ED discharge, a proportion likely substantially lower than for cardiologists or other primary care physicians (37). Because admissions generate the majority of revenue for hospitals, and inpatient care accounts for 31% of national healthcare spending (36), all of these observations have profound clinical and financial implications.
Alternatives to hospital stay in a subset of ED patients must be considered (38). Those patients without high-risk features after ED evaluation could be discharged home, or managed for 24 to 48 h in an ED-based observation unit (38,39). Both strategies would require seamless transition to outpatient care, including early follow-up (40). Exploring these alternative strategies is crucial because the prevalence of chronic heart failure, and ED presentations for AHF are expected to increase over the next decade (1).
Unfortunately, risk-prediction tools in AHF have been largely unsuccessful when attempting to define a cohort of patients safe for early discharge and at low risk of 30-day mortality and readmission (41,42). Further, developed tools have largely not affected disposition decision making. ED patients with AHF are complex and heterogeneous, and may need both an objective evaluation of physiological risk as well as an evaluation of barriers to ideal self-care (e.g., symptom monitoring, medication access, transportation, diet, exercise facilities, caregiver support), along with strategies to overcome these barriers (43). Social, behavioral, and environmental factors strongly influence one's ability to implement a healthy lifestyle required to optimally manage a chronic illness (44,45). Further, patient engagement through shared decision making (46–48) has been successful in other ED-based treatment strategies (49–54) and is an important area to investigate in AHF. Combining comprehensive risk stratification and strategies to overcome self-care barriers, along with shared decision making, may increase the likelihood of successful outpatient transition after either an ED stay or brief management in an observation unit.
Study limitations
This investigation used a large federal database and is subject to several limitations inherent to these databases. There is the potential for underestimating the true AHF ED burden because patients with AHF-related ED visits might have an alternative diagnosis (e.g., pneumonia, chronic obstructive pulmonary disease, asthma, pulmonary embolus) listed as the primary diagnosis and AHF listed as a supporting diagnosis. NEDS also does not have ED-centric measures such as vital signs and results from cardiac biomarkers and chest radiographs. The geographic variability in admission patterns could be explained by differences in illness severity among ED patients with AHF in different regions. Our inclusion and exclusion criteria purposefully omitted patients who would typically not be eligible for ED discharge (cardiogenic shock, mechanical ventilation, acute myocardial infarction), thus admission proportion, LOS, procedures, and charge estimates are likely underestimates. NEDS data likely reflects an undercounting of comorbid conditions; we expected higher percentages. The accuracy of the data may be limited by the potential undercounting of procedures. Although procedures are identified by specific CCS codes, there is the potential that the NEDS hospital medical records staff misreported the procedure or classified an ED procedure as occurring in the hospital. Lastly, HCUP recognizes a limitation on the coding for observation services, although a small proportion of patients with AHF are managed in this way. Observation could be coded as inpatient admission or as ED discharges, likely based on whether the service was provided by the ED or inpatient team.
Conclusions
This large ED cohort analysis suggests a very high admission proportion with significant regional variation in disposition decisions for patients with a primary diagnosis of AHF. Further, although uninsured patients were more likely to be discharged from the ED, when admitted, they received a disproportionally higher rate of major diagnostic and therapeutic procedures. These data suggest AHF is a major ED challenge and consumes significant resources; strategies to reduce this clinical and economic burden are needed.
Acknowledgments
Research reported in this publication was supported by the National Heart, Lung, and Blood Institute (NHLBI) under award number K23HL085387 and grant K12HL1090. The project described was supported by Clinical and Translational Science Award No. UL1TR000445 from the National Center for Advancing Translational Sciences. Its contents are solely the responsibility of the authors and do not necessarily represent official views of the National Center for Advancing Translational Sciences or the National Institutes of Health (NIH). Dr. Storrow receives grant support from Abbott Diagnostics, NIH/NHLBI (K23HL085387 and K12HL1090), National Center for Advancing Translational Sciences (UL1TR000445), Centers for Disease Control, and Roche Diagnostics; and is a consultant for Roche Diagnostics, Abbott Point-of-Care, Astellas, and Novartis Pharmaceuticals Corp, USA. Dr. Self has received research funding from the Centers for Disease Control and Prevention, NIH, Affinium Pharmaceuticals, bioMerieux, Astute Medical, and CareFusion Inc. Dr. Barrett is supported by grant K23HL102069; and serves as a consultant for Red Bull GmbH. Dr. Han is supported by funding from the National Institutes of Health K23AG032355 and the National Center for Advancing Translational Sciences UL1 TR000445. Dr. McNaughton has served on an advisory committee for Cornerstone Therapeutics, Inc. Dr. Gheorghiade has received consulting fees from Bayer, Novartis, Sigma Tau, Johnson & Johnson, Takeda, Otsuka, Stemedica, and Medtronic. Dr. Collins receives research funding from NIH/NHLBI, Medtronic, Cardiorentis, Abbott Point-of-Care, Novartis, The Medicines Company, and Radiometer; and is a consultant for Trevena, Novartis, Otsuka, Radiometer, The Medicines Company, Medtronic, and Astellas.
Abbreviations and Acronyms
- AHF
acute heart failure
- CCS
Clinical Classification Software
- CI
confidence interval
- CPT
Current Procedural Terminology
- ED
emergency department
- HCUP
Healthcare Cost and Utilization Project
- ICD-9 CM
International Classification of Diseases-Ninth Revision-Clinical Modification
- IQR
interquartile range
- LOS
length of stay
- NEDS
Nationwide Emergency Department Sample
Footnotes
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
References
- 1.Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics—2013 update: a report from the American Heart Association. Circulation. 2013;127:e6–245. doi: 10.1161/CIR.0b013e31828124ad. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chen J, Normand SL, Wang Y, Krumholz HM. National and regional trends in heart failure hospital stay and mortality rates for Medicare beneficiaries, 1998–2008. JAMA. 2011;306:1669–78. doi: 10.1001/jama.2011.1474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med. 2009;360:1418–28. doi: 10.1056/NEJMsa0803563. [DOI] [PubMed] [Google Scholar]
- 4.Dharmarajan K, Hsieh AF, Lin Z, et al. Diagnoses and timing of 30-day readmissions after hospital stay for heart failure, acute myocardial infarction, or pneumonia. JAMA. 2013;309:355–63. doi: 10.1001/jama.2012.216476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Writing Group Members. Lloyd-Jones D, Adams RJ, Brown TM, et al. Heart disease and stroke statisticsd2010 update: a report from the American Heart Association. Circulation. 2010;121:e46–215. doi: 10.1161/CIRCULATIONAHA.109.192667. [DOI] [PubMed] [Google Scholar]
- 6.Heidenreich PA, Trogdon JG, Khavjou OA, et al. Forecasting the future of cardiovascular disease in the United States: a policy statement from the American Heart Association. Circulation. 2011;123:933–44. doi: 10.1161/CIR.0b013e31820a55f5. [DOI] [PubMed] [Google Scholar]
- 7.Blecker S, Herbert R, Brancati FL. Comorbid diabetes and end-of-life expenditures among Medicare beneficiaries with heart failure. J Card Fail. 2012;18:41–6. doi: 10.1016/j.cardfail.2011.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Richardson LD, Asplin BR, Lowe RA. Emergency department crowding as a health policy issue: past development, future directions. Ann Emerg Med. 2002;40:388–93. doi: 10.1067/mem.2002.128012. [DOI] [PubMed] [Google Scholar]
- 9.Roger VL, Weston SA, Redfield MM, et al. Trends in heart failure incidence and survival in a community-based population. JAMA. 2004;292:344–50. doi: 10.1001/jama.292.3.344. [DOI] [PubMed] [Google Scholar]
- 10.Weintraub NL, Collins SP, Pang PS, et al. Acute heart failure syndromes: emergency department presentation, treatment, and disposition: current approaches and future aims: a scientific statement from the American Heart Association. Circulation. 2010;122:1975–96. doi: 10.1161/CIR.0b013e3181f9a223. [DOI] [PubMed] [Google Scholar]
- 11.Tang N, Stein J, Hsia RY, Maselli JH, Gonzales R. Trends and characteristics of US emergency department visits, 1997–2007. JAMA. 2010;304:664–70. doi: 10.1001/jama.2010.1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hsia RY, Kellermann AL, Shen YC. Factors associated with closures of emergency departmentsin the United States. JAMA. 2011;305:1978–85. doi: 10.1001/jama.2011.620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pitts SR, Pines JM, Handrigan MT, Kellermann AL. National trends in emergency department occupancy, 2001 to 2008: effect of inpatient admissions versus emergency department practice intensity. Ann Emerg Med. 2012;60:679–686.e3. doi: 10.1016/j.annemergmed.2012.05.014. [DOI] [PubMed] [Google Scholar]
- 14.Morganti KG, Bauhoff S, Blanchard JC, et al. The Evolving Role of Emergency Departments in the United States. Santa Monica, CA: RAND Corporation; 2013. [Accessed July 7, 2013]. Available at: http://www.rand.org/pubs/research_reports/RR280. [PMC free article] [PubMed] [Google Scholar]
- 15.Kellermann AL. Crisis in the emergency department. N Engl J Med. 2006;355:1300–3. doi: 10.1056/NEJMp068194. [DOI] [PubMed] [Google Scholar]
- 16.Committee on the Future of Emergency Care in the United States Health System. Future of Emergency Care Report Series. Washington, DC: National Academies Press; 2006. Hospital-Based Emergency Care: At the Breaking Point. [Google Scholar]
- 17.Schappert SM, Rechtsteiner EA. Ambulatory medical care utilization estimates for 2006. Natl Health Stat Report. 2008;8:1–29. [PubMed] [Google Scholar]
- 18.Schuur JD, Venkatesh AK. The growing role of emergency departments in hospital admissions. N Engl J Med. 2012;367:391–3. doi: 10.1056/NEJMp1204431. [DOI] [PubMed] [Google Scholar]
- 19.AHRQ. [Accessed March 8, 2014];AHRQ Healthcare Cost and Utilization Project Overview of the Nationwide Emergency Department Sample (NEDS) 2013 Available at https://www.hcup-us.ahrq.gov/nedsoverview.jsp.
- 20.U.S. Census Bureau. Census regions and divisions of the United States. US Department of Commerce Economics, and Statistics Administration, Geography Division; [Accessed March 8, 2014]. Available at: http://www.census.gov/geo/maps-data/maps/pdfs/reference/us_regdiv.pdf. [Google Scholar]
- 21.Burt CW, Arispe IE. Characteristics of emergency departments serving high volumes of safety-net patients: United States, 2000. Vital Health Stat 13. 2004;155:1–16. [PubMed] [Google Scholar]
- 22.SAS Institute. SAS/STAT 9.2 User's Guide: The SURVEYFREQ Procedure (Book Excerpt) Cary, NC: SAS Institute; 2008. [Accessed March 8, 2014]. Available at: https://support.sas.com/documentation/cdl/en/statugsurveyfreq/61835/PDF/default/statugsurveyfreq.pdf. [Google Scholar]
- 23.SAS Institute. SAS/STAT 9.2 User's Guide: The SURVEYLOGISTIC Procedure (Book Excerpt) Cary, NC: SAS Institute; 2008. [Accessed March 8, 2014]. Available at: http://support.sas.com/documentation/cdl/en/statugsurveylogistic/61836/PDF/default/statugsurveylogistic.pdf. [Google Scholar]
- 24.SAS Version 9.2 of the SAS System for Linux (software) Cary, NC: SAS Institute; 2009. [Google Scholar]
- 25.Public Law 111-148. U.S. Government Printing Office; 2010. 111th Congress. Patient Protection and Affordable Care Act. [Google Scholar]
- 26.Vaduganathan M, Bonow RO, Gheorghiade M. Thirty-day read-missions: the clock is ticking. JAMA. 2013;309:345–6. doi: 10.1001/jama.2012.205110. [DOI] [PubMed] [Google Scholar]
- 27.Chen J, Normand SL, Wang Y, Drye EE, Schreiner GC, Krumholz HM. Recent declines in hospital stayhospital stays for acute myocardial infarction for Medicare fee-for-service beneficiaries: progress and continuing challenges. Circulation. 2010;121:1322–8. doi: 10.1161/CIRCULATIONAHA.109.862094. [DOI] [PubMed] [Google Scholar]
- 28.Yeh RW, Sidney S, Chandra M, Sorel M, Selby JV, Go AS. Population trends in the incidence and outcomes of acute myocardial infarction. N Engl J Med. 2010;362:2155–65. doi: 10.1056/NEJMoa0908610. [DOI] [PubMed] [Google Scholar]
- 29.Hajjar I, Kotchen JM, Kotchen TA. Hypertension: trends in prevalence, incidence, and control. Annu Rev Public Health. 2006;27:465–90. doi: 10.1146/annurev.publhealth.27.021405.102132. [DOI] [PubMed] [Google Scholar]
- 30.Wang TJ, Vasan RS. Epidemiology of uncontrolled hypertension in the United States. Circulation. 2005;112:1651–62. doi: 10.1161/CIRCULATIONAHA.104.490599. [DOI] [PubMed] [Google Scholar]
- 31.Fonarow GC, Albert NM, Curtis AB, et al. Improving evidence-based care for heart failure in outpatient cardiology practices: primary results of the Registry to Improve the Use of Evidence-Based Heart Failure Therapies in the Outpatient Setting (IMPROVE HF) Circulation. 2010;122:585–96. doi: 10.1161/CIRCULATIONAHA.109.934471. [DOI] [PubMed] [Google Scholar]
- 32.Fonarow GC, Heywood JT, Heidenreich PA, Lopatin M, Yancy CW. Temporal trends in clinical characteristics, treatments, and outcomes for heart failure hospital stayhospital stays, 2002 to 2004: findings from Acute Decompensated Heart Failure National Registry (ADHERE) Am Heart J. 2007;153:1021–8. doi: 10.1016/j.ahj.2007.03.012. [DOI] [PubMed] [Google Scholar]
- 33.Lee DS, Schull MJ, Alter DA, et al. Early deaths in patients with heart failure discharged from the emergency department: a population-based analysis. Circ Heart Fail. 2010;3:228–35. doi: 10.1161/CIRCHEARTFAILURE.109.885285. [DOI] [PubMed] [Google Scholar]
- 34.Bueno H, Ross JS, Wang Y, et al. Trends in length of stay and short-term outcomes among Medicare patients hospitalized for heart failure, 1993–2006. JAMA. 2010;303:2141–7. doi: 10.1001/jama.2010.748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Peacock WF, Braunwald E, Abraham W, et al. National Heart, Lung, and Blood Institute working group on emergency department management of acute heart failure: research challenges and opportunities. J Am Coll Cardiol. 2010;56:343–51. doi: 10.1016/j.jacc.2010.03.051. [DOI] [PubMed] [Google Scholar]
- 36.Health, United States, 2012: With a Special Feature on Emergency Care. National Center for Health Statistics; Hyattsville, MD: National Center for Health Statistics; 2013. [PubMed] [Google Scholar]
- 37.McCausland JB, Machi MS, Yealy DM. Emergency physicians' risk attitudes in acute decompensated heart failure patients. Acad Emerg Med. 2010;17:108–10. doi: 10.1111/j.1553-2712.2009.00623.x. [DOI] [PubMed] [Google Scholar]
- 38.Collins SP, Pang PS, Fonarow GC, Yancy CW, Bonow RO, Gheorghiade M. Is hospital admission for heart failure really necessary? The role of the emergency department and observation unitin preventing hospital stay and rehospital stay. J Am Coll Cardiol. 2013;61:121–6. doi: 10.1016/j.jacc.2012.08.1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Pang PS, Jesse R, Collins SP, Maisel A. Patients with acute heart failure in the emergency department: do they all need to be admitted? J Card Fail. 2012;18:900–3. doi: 10.1016/j.cardfail.2012.10.014. [DOI] [PubMed] [Google Scholar]
- 40.Hernandez AF, Greiner MA, Fonarow GC, et al. Relationship between early physician follow-up and 30-day readmission among Medicare beneficiaries hospitalized for heart failure. JAMA. 2010;303:1716–22. doi: 10.1001/jama.2010.533. [DOI] [PubMed] [Google Scholar]
- 41.Collins SP, Lindsell CJ, Jenkins CA, et al. Risk stratification in acute heart failure: rationale and design of the STRATIFY and DECIDE studies. Am Heart J. 2012;164:825–34. doi: 10.1016/j.ahj.2012.07.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Collins SP, Storrow AB. Acute heart failure risk stratification: can we define low risk? Heart Fail Clin. 2009;5:75–83, vii. doi: 10.1016/j.hfc.2008.08.010. [DOI] [PubMed] [Google Scholar]
- 43.Collins SP, Storrow AB. Moving toward comprehensive acute heart failure risk assessment in the emergency department. J Am Coll Cardiol HF. 2013;1:273–80. doi: 10.1016/j.jchf.2013.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Wu JR, Moser DK, Lennie TA, Burkhart PV. Medication adherence in patients who have heart failure: a review of the literature. Nurs Clin North Am. 2008;43:133–53. vii–viii. doi: 10.1016/j.cnur.2007.10.006. [DOI] [PubMed] [Google Scholar]
- 45.van der Wal MH, Jaarsma T. Adherence in heart failure in the elderly: problem and possible solutions. Int J Cardiol. 2008;125:203–8. doi: 10.1016/j.ijcard.2007.10.011. [DOI] [PubMed] [Google Scholar]
- 46.Barry MJ, Edgman-Levitan S. Shared decision making—pinnacle of patient-centered care. N Engl J Med. 2012;366:780–1. doi: 10.1056/NEJMp1109283. [DOI] [PubMed] [Google Scholar]
- 47.Gerteis M, Edgman-Levitan S, Daley J, Delbanco T. Through the Patient's Eyes. San Francisco, CA: Jossey-Bass; 1993. [Google Scholar]
- 48.Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Patient Educ Couns. 2006;60:301–12. doi: 10.1016/j.pec.2005.06.010. [DOI] [PubMed] [Google Scholar]
- 49.Yamamoto LG. Application of informed consent principles in the emergency department evaluation of febrile children at risk for occult bacteremia. Hawaii Med J. 1997;56:313–7. 320–2. [PubMed] [Google Scholar]
- 50.Yamamoto LG, Young LL, Roberts JL. Informed consent and parental choice of anesthesia and sedation for the repair of small lacerations in children. Am J Emerg Med. 1997;15:285–9. doi: 10.1016/s0735-6757(97)90017-6. [DOI] [PubMed] [Google Scholar]
- 51.Karpas A, Finkelstein M, Reid S. Parental preference for rehydration method for children in the emergency department. Pediatr Emerg Care. 2009;25:301–6. doi: 10.1097/PEC.0b013e3181a34144. [DOI] [PubMed] [Google Scholar]
- 52.Hess EP, Knoedler MA, Shah ND, et al. The chest pain choice decision aid: a randomized trial. Circ Cardiovasc Qual Outcomes. 2012;5:251–9. doi: 10.1161/CIRCOUTCOMES.111.964791. [DOI] [PubMed] [Google Scholar]
- 53.Kline JA, Zeitouni RA, Hernandez-Nino J, Jones AE. Randomized trial of computerized quantitative pretest probability in low-risk chest pain patients: effect on safety and resource use. Ann Emerg Med. 2009;53:727–735.e1. doi: 10.1016/j.annemergmed.2008.09.034. [DOI] [PubMed] [Google Scholar]
- 54.Isaacs CG, Kistler C, Hunold KM, et al. Shared decision-making in the selection of outpatient analgesics for older individuals in the emergency department. J Am Geriatr Soc. 2013;61:793–8. doi: 10.1111/jgs.12207. [DOI] [PMC free article] [PubMed] [Google Scholar]