1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References

There is little information on the effectiveness of emergency department (ED) observation unit (OU) heart failure (HF) therapy. The authors' objective was to evaluate outcomes after implementation of an ED-OU treatment protocol for HF exacerbation. Unblinded assessment of the effectiveness of an HF protocol was performed, controlled by outcome for 9 months prior to implementation. This included diagnostic and therapeutic algorithms, cardiology consultation, close monitoring, patient education, and discharge planning. Adverse outcomes were defined as the 90-day rates of ED HF revisits, hospital HF readmissions, or death, as determined by chart review, computer database search, and phone follow-up. One hundred fifty-four patients were enrolled; 50 entered before, and 104 after protocol implementation. Only six (12%) in the preprotocol and one (1%) of the post protocol group were lost to follow-up. After an OU visit, post protocol 90-day ED HF revisit rates declined 56% (0.90–0.51; p<0.0000) and the 90-day HF rehospitalization rate decreased 64% (0.77–0.50; p=0.007). The 90-day rates of death and OU HF readmission decreased from 4% to 1% (p=0.096) and 18% to 11% (p=0.099), respectively. An intensive outpatient ED OU HF management protocol safely decreases 90-day rates of emergency department visits and inpatient hospitalizations.

Heart failure (HF) is the only cardiovascular disease increasing in incidence and prevalence.1,2 Disproportionally affecting the elderly, it is the leading cause of hospitalization if older than 65.2,3 Furthermore, it is the most expensive diagnosis for the Medicare health system, with annual costs exceeding $24 billion, as estimated by the Heart Failure Society of America in 1999. Annual US HF hospitalization costs are more than double those of all forms of cancer,3 and HF patients have an overall poor prognosis.4,5 After developing symptoms, 2-year mortality is about 35%, becoming 80% for men and 65% for women over the next 6 years.6 Frequently recurring symptoms result in multiple emergency department (ED) visits, which are followed by many inpatient hospitalizations.7 Methods to decrease ED visits and inpatient hospitalizations may not only decrease hospital costs,8,9 but also have the potential to improve quality of life for HF patients.10

It is not routine for the ED observation unit (OU) to admit HF patients. As a consequence, many symptomatic ED HF patients are admitted for inpatient care. There are few reports of the results of HF management in the ED OU. We report the rates of ED revisits, OU repeat admissions, and repeat inpatient hospitalizations for patients discharged after HF exacerbation management in an ED OU. We also report the changes in these outcome measures as a result of implementation of an aggressive ED OU HF management protocol.11


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References

This was a 20-month, unblinded, prospective study of the effect of a treatment protocol11 for ED OU HF patients. It was performed at a 1000-bed, university-affiliated, tertiary care, urban teaching hospital, in an ED that has an annual census of 45, 204, and a 20-bed OU. All patients were managed by board-prepared emergency physicians throughout their stay. The OU is staffed 24 hours per day with dedicated registered nurses and physician personnel.

All patients with an exacerbation of HF presenting to the ED during the study period were eligible for entry, but had to fulfill the entry criteria listed in Table I. Protocol candidates were initially evaluated in the ED for possible OU admission. Patients were excluded if they had any of the specific OU HF protocol exclusions (Table II). Patients excluded from OU care are not included in the analysis.

Table I.  Observation Unit Heart Failure Protocol Entry Guidelines*
Entry criteria: Must have at least one from each category
   Dyspnea on exertion
   Paroxysmal nocturnal dyspnea
   Shortness of breath
   Edema of legs or abdomen
   Weight gain
   Jugular venous distention or elevation in pulsation
   Positive abdominal jugular reflux
   Inspiratory rales
   Peripheral edema
Chest x-ray
   Pulmonary vascular congestion
   Kerley B lines
   Pulmonary edema
Pleural effusion
*If the ejection fraction is >40%, only certain facets of the protocol may be appropriate. Treatment should always be tailored to the individual patient.
Table II.  Observation Unit Heart Failure Protocol Exclusion Criteria*
• Unstable vital signs (blood pressure 220/120 mm Hg, respiratory rate >25, heart rate >130 beats/min)
• Temperature >38.5°C
• Unstable airway, or need for >4 L/min supplemental O2 by nasal cannula to keep O2 saturation at >90%
• Peak flow <50% of predicted, with wheezing
• Clinically significant arrhythmia or sustained ventricular tachycardia
• Any EKG with diagnostic criteria for acute myocardial infarction or ischemia
• Chest x-ray with pulmonary infiltrates
• Any CK-MB >8.8 ng/mL
• Any troponin T >0.1 μg/L if creatinine <2.0 μg/dL; >0.5 mg/L if creatinine >2.0 mg/dL
• Requirement for continuous vasoactive medication (e.g., nitroglycerin, nitroprusside, dobutamine, milrinone) to stabilize hemodynamics
• Complex decompensation: concomitant end-organ hypoperfusion, volume overload, and systemic vasoconstriction
• Requirement for hemodynamically guided care by pulmonary artery catheter
• Severe electrolyte imbalance
• Chronic renal failure requiring dialysis
• Acute mental status abnormality
EKG=electrocardiogram; CK-MB=creatine kinase MB; *if the ejection fraction is >40%, only certain facets of the protocol may be appropriate. Treatment should always be tailored to the individual patient.

The length of stay in the OU was limited to 24 hours. At 24 hours, the physicians were required to make an admission or discharge decision, based on their own clinical evaluation. Since this was a safety and efficacy evaluation of OU care, this analysis examines outcomes only in those patients discharged home after OU therapy, excluding those requiring hospital admission.

The primary end point was the occurrence of ED HF revisits, hospital HF admissions, or death within 90 days. All cardiorespiratory complaints were considered to be HF-related, including repeat visits for chest pain or syncope. Outcomes were determined by blinded reviews of the patient's chart, computer database, and by phone follow-up. Patients were considered to receive their care primarily at our facility if, in the year prior, they were seen in multiple outpatient departments (two or more non-HF departments) of our hospital system or exceeded three HF department visits.

In the first 10 months of the study, emergency physicians managed OU HF patients as per their usual practice. No specific therapy or guidelines were mandated. We undertook implementation of an aggressive OU HF management protocol in November, 1997. This included preprinted physician order sheets with diagnostic and therapeutic algorithms11 and listed options for consultations, patient education, and follow-up arrangements. Outcome data were collected for 10 months after the protocol was implemented.

In our protocol, the management focus is broadened from managing the presenting symptoms of HF to a global management approach. This includes determining the type of HF; assessing factors leading to worsening symptoms; optimizing medication management, with emphasis on treating clinical and subclinical volume overload; optimizing angiotensin-converting enzyme (ACE) inhibitor therapy toward target doses; providing HF education; and giving appropriate post-discharge follow-up outpatient care (Table III).

Table III.  Observation Unit Protocol
     Continuous EKG and pulse oximetry
     Strict input and output, 1800-cc fluid restriction, no added-salt diet
     Patient weight
     Angiotensin-converting enzyme inhibitor recommended
     Furosemide algorithm
          Double daily 24-hour dose, given as single i.v. bolus (180 mg maximum)
          Double previously administered dose and repeat if fail to meet 2-hour urine output goal
          Two-hour urine output goals: 500 cc if creatinine <2.5 mg/dL; 250 cc if creatinine >2.5 mg/dL
Diagnostic procedures
     Ejection fraction, measured by echocardiography, unless systolic HF is known, or
     diastolic HF was diagnosed within 1 year
     CK-MB and troponin T measured every 6 hours, for 12 hours
     HF specialist consult in all; social work, home health care, and dietary, as indicated
     View 15-minute HF video and receive personalized discharge instruction packet
EKG=electrocardiography; HF=heart failure; CK-MB=creatine kinase MB

Discharge determination was at the discretion of the emergency physician assigned to the OU. Guidelines were developed to assist in this evaluation (Table IV). Patients failing OU management were admitted to the appropriate inpatient unit.

Table IV.  Observation Unit Heart Failure Discharge Guidelines
The intent of this document is to establish guidelines for the discharge of heart failure patients from the observation unit and should not supersede good clinical judgment. Patients not meeting all of the following should be considered for inpatient treatment (except as appropriate in the end-stage palliative-care cohort).
   Patient reports subjective improvement
   Patient is ambulatory, without long-term orthostasis
   Resting heart rate <100 beats/min
   Systolic blood pressure >80 mm Hg
   Total urine output >1000 cc and no new decrease in urine output below 30 cc/hr (or <0.5 cc/kg/hr)
   Room air O2 saturation >90% (unless on home O2)
   All CK-MB <8.8 ng/mL, and troponin T <0.1 μg/L
   No ischemic-type chest pain
   No new clinically significant arrhythmia
Stable electrolyte profile
CK-MB=creatine kinase MB

Statistical analysis was performed using the Student t test for continuous data, and chi-square for discontinuous data. Significance was defined as p<0.05. This study was approved by the Institutional Review Board.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References

One hundred fifty-four patients were enrolled during the 20-month study period. Males comprised 60% (92 of 154) of the study population. The average age was 67.9 years (31–96; SD=13.2). As defined previously, 149 (97%) were cared for primarily at our facility. In the 10 months before the management protocol was implemented, 50 patients were entered. In the 10-month post protocol period, 104 patients were entered.

The final OU electrocardiogram (EKG), as interpreted by board-prepared emergency physicians, was coded as no acute ischemic changes, atrial fibrillation or flutter, diagnostic for acute ischemic changes, or presence of an electronic ventricular pacemaker with a paced rhythm. There was no significant difference in the EKG results between the pre- and post protocol groups (p=0.51).

Serial creatine kinase (CK) testing demonstrated no differences between the pre- and post protocol groups at any of the three collection times (p=0.59, p=0.54, and p=0.24), nor did serial CK-MB testing (p=0.12, p=0.63, and p=0.60). No patient undergoing serial cardiac enzyme testing whose initial results were negative developed positive results. Troponin T testing was performed only at the initial presentation. It was positive in no preprotocol and three post protocol patients. There was no significant difference between pre- and post protocol groups in troponin T results (p=0.16).

The chest radiograph of the OU admission film was interpreted by board-certified radiologists and categorized as normal; cardiomegaly; increased pulmonary vascularity, congestive heart failure, or pulmonary effusion; pulmonary edema; or non-HF pathology. The non-HF pathology category included pulmonary infiltrate suggestive of pneumonia and parenchymal lung mass. There was no significant difference in the pre- and post protocol groups in chest x-ray findings (p=0.295).

Outcome data were available for the majority of patients during the 90-day post-OU discharge follow-up period. Only six (12%) of the preprotocol group, and one (1%) of the post protocol cohort, were lost to follow-up.

Following the index OU HF admission, preprotocol 90-day rates of ED HF return visits were 90%. After implementation of the management protocol, only 51% required a return to the ED for management of HF in the ensuing 90 days. This represents a decline in ED HF return visits by 43% (p<0.0000). The return visit rate for OU HF management, during the 90 days after the index OU HF visit, was 18% before implementation of the management protocol. This decreased to 11% in the post protocol period (p=0.099).

Before the management protocol was implemented, 77% of patients returned within 90 days and required inpatient hospitalization for HF exacerbation. This decreased to 50% following implementation of the OU management protocol (p=0.007). Similarly, the 90 day post-OU index visit death rate was initially 4% and was 1% following implementation of the management protocol (p=0.096). No patients died during their stay in the OU.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References

This is the first study to examine outcomes of HF patients after a short course of therapy in an ED OU. We demonstrate that protocol-driven OU treatment of decompensated HF is safe and effectively manages patients while at the same time decreasing the inpatient hospitalization rate. Patients presenting to the ED with HF and managed in the OU before implementation of the HF management protocol had high rates of return visits to the ED and OU, and rehospitalizations for HF were frequent. Implementation of the management protocol nearly halved return ED visits for HF, from an average of almost one per patient during the 90-day follow-up period to a mean of 0.51 per patient. This beneficial effect was not the consequence of patients “bypassing” the ED, as HF rehospitalizations dropped by an even greater percentage (64%) during the post-protocol period.

Our protocol incorporates previously published HF treatment guidelines1,6,12–14 known to provide benefit, and these results are not the effect of novel HF medication or treatment. Rather, we suspect our patients benefit from the institution of an algorithm that insures aggressive diuresis. This prevents delays and omissions that may occur without an algorithm, and focuses attention on details of HF management. Furthermore, standardization of care insures that important interactions, such as bedside education, occur for all patients.

Inpatient multidisciplinary interventions for HF decrease 90-day hospital readmissions, improve quality of life, and decrease costs.12,15,16 It is reasonable to expect that a similar result could be duplicated in the OU environment. While management protocols are not mandatory for HF management, they insure the timely interplay of the many different interdisciplinary groups (e.g., nursing, social work, dietetics, etc.) required for successful management. Standardized treatment insures that details, individually insignificant but important in aggregate, are not omitted and that there is no delay in their performance.

Multiple re-visits to the hospital are common in HF patients.7,14–16 Medication and dietary noncompliance account for a significant percentage of rehospitalizations in this group.12,15,17 Frequent hospitalizations erode the quality of life for the HF patient. Education, while the patient is symptomatic, may be more effective in inducing lifestyle changes with negative consequences with respect to HF. Lifestyle recommendations and patient education with demonstrable near-term beneficial effects are more likely to be followed.

While there are no comprehensive studies reporting 90-day ED revisit rates, our high rate of preprotocol re-admission is compatible with common experience. By the time patients bypass routine care and present to the ED, their cardiovascular disease has advanced to such a state that frequent repeat visits are likely. Furthermore, ED visits are a marker of social situations that increase the likelihood of future ED presentations. Consequently, the combination of advanced disease and a poor social milieu conspire to result in frequent ED visits. These factors must be considered when designing treatment algorithms.

Ejection fraction measurement is the standard of care for HF patients,1,6,13 but it is not easily available in the ED environment. The OU allows for more complete evaluation and is especially helpful for those with social problems that make it difficult to complete their work-up. Non-compliance is a common reason for HF decompensation. The multiple outpatient visits required for evaluation, up-titration of medications, and other requirements, are frequently unmet by the individual most likely to present for ED therapy. Transportation issues can be problematic for the elderly, and this is particularly true for the disabled HF patient dependent on others to keep appointments and visit the pharmacist to fill prescriptions. In the OU, medications are initiated or up-titrated in a controlled environment for those in whom it would otherwise be difficult. The “one-stop shopping” of OU HF management can provide benefits in patients with confounding social issues.

OU patients are less acutely ill than those requiring immediate hospitalization; as such, they represent a subcategory of more stable patients. It is appropriate that they should have a shorter length of stay than patients with more symptomatic disease. The OU has a limited stay, which serves to prompt early decision-making by medical staff. Shorter stays result in treatment costs that are significantly lower than inpatient hospitalization. The financial impact is even greater when an OU admission results in avoidance of future hospitalizations.8

While death rates declined after the study period, the reduction did not reach statistical significance. Our absolute number of deaths is low for a HF population. However, these patients were preselected to be at the lowest risk for adverse outcomes, and should have low mortality rates. A larger study may identify a mortality benefit from OU HF treatment.

In this study, we report only on patients treated and successfully discharged from an ED OU. We have studied, and reported elsewhere,13 data on patients requiring inpatient hospitalization after OU treatment. Hospitalized patients represent a higher risk group, and are certainly not homogeneous with the population who can be adequately treated and safely discharged home within 24 hours.

In the analysis of our data, we considered that patients could have been hospitalized elsewhere, and would not appear in our outcome data. We controlled for this by determining where patients received the majority of their health care. If patients were seen multiple times by our HF specialists, or had many subspecialty visits at our facility, we defined them as receiving their care primarily within our hospital system. Since 97% of our patients met this definition, we feel it is likely that this relationship would continue during our 3-month follow-up period. If they sought hospitalization, it most probably would have been where they had historically received their medical care.

Limitations of our study include the possibility that the OU HF population acuity could have been diluted by ED admissions of patients who otherwise would have been discharged. We did not control for admission acuity during our study period. Future studies should focus on refinements in treatment algorithms using brain natriuretic peptide levels as a predefined entry criterion and examine the role of alternative medication regimens (e.g., continuous i. v. diuretic therapy or nesiritide) and the value of non-invasive hemodynamic monitoring technologies (e.g., thoracic bioimpedance).

While the numbers of patients managed for HF increased during the second half of the study period, this was associated with a decreased rate of inpatient rehospitalization and a decreased rate of return for OU HF management. Increases in the size of the post protocol cohort (104 patients), compared to the preprotocol group (50 patients), are likely due to several factors: 1) the global ED census increased during the 20 months of this study, and the HF population increased in parallel; 2) increased comfort in managing HF on the part of emergency physicians increased the probability of an OU HF admission; and 3) knowledge of the protocol by referral physicians increased the rate of HF patients sent to the ED, rather than directly admitted to the inpatient floor. Our study was not designed to directly test these hypotheses.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References

We conclude that the ED OU management protocol is safe and effective. We also present an effective strategy to decrease 90-day ED revisits and rehospitalizations in decompensated HF.

Acknowledgments: This paper was presented at The American Heart Association 72nd Scientific Session, Atlanta, GA; November, 1999.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. References
  • 1
    Packer M, Cohn JN. Consensus recommendations for the management of chronic heart failure. Am J Cardiol. 1999;83:2A, 1A38A.
  • 2
    Massie BM, Shah NB. Evolving trends in epidemiologic factors of heart failure: rationale for preventative strategies and comprehensive disease management. Am Heart J. 1997;133 703712.
  • 3
    O'Connell JB, Bristow MR. Economic impact of heart failure in the United States: time for a different approach. J Heart Lung Transplant. 1994;13:S107S112.
  • 4
    Rich MW, Beckham V, Wittenberg C, et al. A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med. 1995;333: 11901195.
  • 5
    Rich MW, Vinson JM, Sperry JC, et al. Prevention of read mission in elderly patients with congestive heart failure. J Gen Int Med. 1993;8:585590.
  • 6
    Vinson JM, Rich MW, Sperry JC, et al. Early readmission of elderly patients with congestive heart failure. J Am Geriatr Soc 1990;38:12901295.
  • 7
    Krumholz HM, Parent EM, Tu N, et al. Readmission after hospitalization for congestive heart failure among Medicare beneficiaries. Arch Intern Med. 1997; 157:99104.
  • 8
    Ashton CM, Kuykendall DH, Johnson ML, et al. The association between the quality of inpatient care and early readmission. Ann Intern Med. 1995;122(6):415421.
  • 9
    Van Lente F, McErlean ES, DeLuca SA, et al. Ability of troponins to predict adverse outcomes in patients with renal in sufficiency and suspected acute coronary syndromes: a case matched study. J Am Coll Cardiol. 1999;33(2):471478.
  • 10
    Guidelines for the evaluation and management of heart failure: Report of the American College of Cardiology/American Heart Association task force on practice guidelines (Committee of Evaluation and Management of Heart Failure). Circulation. 1995;92:27642784.
  • 11
    McKee PA, Castelli WP, McNamara PM, et al. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971;285:14411446.
  • 12
    Ho KK, Anderson KM, Kannel WB, et al. Congestive heart failure/myocardial responses/valvular heart disease: survival after the onset of congestive heart failure in Framingham Heart Study subjects. Circulation. 1993;88(1): 107115.
  • 13
    Albert NM, Peacock WF. Patient outcome and costs after implementation of an acute heart failure management program in an emergency department observation unit [abstract]. J Heart Lung Transplant. 1999;18(1):92.
  • 14
    Peacock WF, Albert NM. Observation unit management of heart failure. Emerg Med Clin North Am. 2001;19(1):209232.
  • 15
    Peacock WF, Albert NM, Kies P, et al. Emergency department observation unit heart failure protocol decreases adverse outcome rates. J Card Fail. 1999;5(3 suppl 1):77.
  • 16
    Peacock W, Aponte J, Craig M, et al. Predictors of unsuccessful treatment for congestive heart failure in the emergency department observation unit [abstract]. Acad Emerg Med. 1997;4(5):494.
  • 17
    Clinical Practice Guideline, Number 11. Heart Failure: Evaluation and Care of Patients With Left-Ventricular Systolic Dysfunction. US Department of Health and Human Services. June, 1994. AHCPR Publication No. 94-0612.