It has been well established that the change in patient-centered outcome measures such as dyspnea can be utilized to assess therapeutic success and improvement in symptoms. Dyspnea is the most common symptom in patients with HF that present to the ED [1, 5], and resolution of dyspnea remains the most common goal of treatment [1]. A high proportion of patients have improvement in their dyspnea during the ED stay due to standard therapy, and many have complete resolution of their dyspnea within their 24-h OU stay [10]. However, the subjective nature of this patient-centered measure had made it challenging to standardize. Furthermore, dyspnea has not been shown to correlate well with worsening HF during hospital admission or post-discharge events. Maneuvers that illicit cardiac stress such as changing patient’s position from sitting to supine or ambulation tests may help to identify the more subtle effects of HF on dyspnea and may serve as more definitive measures of clinical improvement. Other symptoms related to HF such as fatigue and body swelling may become equally important to explore as patient-centered outcomes [1]. Assessment of dyspnea is an integral component of disposition assessment.

Patients also should be able to ambulate without an increase in dyspnea from the baseline. Although there is no trial that has assessed this measure in an observation unit setting, it is effectively an inexpensive 6-min exercise test. The distance that a patient can ambulate in a 6-min period without excessive dyspnea and fatigue has been shown to correlate with long-term mortality [11, 12]. Unfortunately, many comorbid illnesses, such as obesity and lung disease, affect this outcome measure. It is important to assess a change from the baseline.

While the prior studies did not specifically evaluate the ED, one investigation tested the feasibility of a 3-min walk in the emergency department and found that 85 % of all patients were able to complete the walk and the ability to walk 3 min was associated with outcomes [13]. In addition, freedom from symptoms of congestion has also been associated with improved long-term outcomes, although orthopnea can persist even after subjective improvement in dyspnea [14, 15].

Another patient-centered measure that should be present at the time of discharge is the lack of ongoing chest pain. It has been reported that acute coronary syndrome (ACS) is a trigger for up to 25 % of patients with heart failure decompensation. Therefore, patients should be pain-free or have undergone an evaluation for ACS prior to discharge [16].

Finally, self-care barriers must be overcome in order to be able to discharge HF patients safely. Patients and/or family members must feel confident in caring for themselves at home and express full understanding of medication changes and follow-up appointments. Moreover, patients must not be at a high risk of falling at home and must have dependable transportation.

Hemodynamic and clinical parameters can be a part of the data used to assess suitability for discharge. These comprise measures of perfusion, volume status, and oxygenation-based physical exam findings, as well as automated measures. Systolic blood pressure (SBP) is a useful predictor of adverse events at the time of presentation and discharge [17]. In the initial presentation of patients with decompensated HF, a hypertensive response is adaptive, although persistent elevation of SBP can correlate with increased risk of worsening renal function. In HF, any deterioration of renal function clearly correlates with morbidity and mortality; therefore, adjustment of medications to prevent hypertension is essential prior to discharge. While the ideal blood pressure at the time of hospital discharge is not clearly elucidated, patients should at least have a SBP <160 mmHg [17]. Conversely, as medications are titrated, patients must be able to ambulate without symptoms of dizziness; therefore, the SBP should exceed 90 mmHg [18].

Clinical findings can also be used to assess adequacy of acute interventions. These include a combination of changes in physical exam and easily obtained values such as pulse oximetry, urine output, and weight. Of all the clinical examination findings, the presence of an S₃ is most suggestive of acute decompensation [19]. Serial exams that document the resolution of an S₃ by auscultation can be used as a discharge criterion [19]. However, the presence of a digitally recorded S₃ has not been shown to be associated with prognosis or improved diagnostic accuracy in one large clinical trial [20]. This physical exam finding, like an improvement in jugular venous distention, is dependent of physical attributes of the patient and careful physical exam assessment by the physician.

Another criterion, noted as part of the evaluation, is oxygen saturation. Patients should have an oxygen saturation greater than 90 % [18]. No data exist to support this value; however, it is reasonable to only discharge patients who are able to maintain their oxygen saturation. Transient nighttime drops in oxygen saturations are common because HF is associated with an increased prevalence of obstructive sleep apnea. Therefore, pulse oximetry as a discharge criterion should be assessed when the patient is awake.

Urine output assessment is another parameter that can be used as a surrogate to assess treatment efficacy. Although there are no studies that compare the amount of urine output with outcomes, intuitively, this makes sense. Clinically, 1 L appears to be a significant amount. Closely linked to urine output is the patient’s weight [18, 34]. Dry weight is often one of the only baseline parameters that is known in the ED. Theoretically, a decline in the patient’s weight can represent a resolution of the acute progression of the disease process; however, “overshooting” this parameter can lead to hypotension, hypoperfusion, and worsening renal function. Although not supported by clinical trials, it is reasonable to suggest that a patient’s weight should be declining at the time of discharge; however, additional assessment may be warranted in patients who are below their dry weight at the time of discharge assessment.

Improvement in laboratory parameters may also be used to assess patients at the time of discharge. B-type natriuretic peptide (BNP) levels are the most established diagnostic biomarkers for HF [1]. BNP levels have been associated with a decrease in hospitalization, intensive care unit utilization, hospital length of stay, and cost of treatment when used as a diagnostic strategy. The Rapid ED Heart Failure Outpatient Trial (REDHOT) results showed that BNP levels predicted early outcomes more accurately than physician’s impression. Patients in this study with BNP levels less than 200 pg/mL had 0 % mortality at 30 days, regardless of EF [4]. Another study involving Veterans Administration patients found that patients with BNP level less than 230 pg/mL demonstrate a very low risk of recidivism to the ED, readmission, and even death [4]. However, BNP levels can be affected by age, sex, weight, and renal function and can thus be misleading [1]. BNP levels are found to be lower in obese patients with HF and are higher in patients with renal dysfunction. For these reasons, comparing the ED BNP level to the patient’s baseline is of more value than standardizing the normal limits for all HF patients [4]. BNP levels that are more than 50 % changed from the baseline generally represent worsening HF [1]. Newer markers are also available, such as ST2 and galectin-3, and may aid with the prognostic assessment and diagnosis of HF [1].

Collins et al. comprised a STRATIFY decision tool to help identify HF patients in the ED who were at low risk for 30-day adverse events and could thus be safely discharged home. The decision tool included 13 variables that are readily available in the ED, including age, body mass index (BMI), BNP, diastolic blood pressure (DBP), blood urea nitrogen (BUN), serum sodium level, respiratory rate, oxygen saturation, troponin level, the use of dialysis, the outpatient use of supplemental O₂, the outpatient use of an ACEI, and QRS duration. While an elevated BNP was not shown to be a significant predictor of serious adverse events or death, its elevation did show a trend toward an association with adverse events [6].

Elevated troponin levels have been shown to be predictive of long-term prognosis in HF patients [21–23]. Patients with severe HF may have chronically elevated levels. The STRATIFY decision tool constructed by Collins et al. found that an elevated troponin level in the ED was found to be one of the two significant predictors of adverse events. The study also found that 13 % of patients with HF in the ED had a less than 5 % risk of death or serious complications in the first 30 days after ED presentation. Of note, adverse events related to ACS were most commonly seen in the first 5 days after ED presentation and were detected using troponin testing in the ED [6]. It is recommended that HF patients have no ischemic changes or elevated troponin in order to be candidates for the OU. However, patients with only minimally elevated troponin levels may still be appropriate for observation management if the levels are trended to rule out acute coronary syndrome. These patients are at risk for failing observation care and may need to be moved to inpatient management [1]. Therefore, patients with an elevated initial troponin level are probably not ideal candidates for an early discharge strategy.

Traditional chemistry labs that are routinely assessed daily in patients with decompensated heart failure can also be used in the assessment at the time of discharge. The ED measured BUN level is the other variable measured in the STRATIFY decision tool found to be a significant predictor of adverse events [6]. Appropriate candidates for the OU include patients whose BUN levels are less than 40 mg/dL and whose creatinine levels are less than 3.0 mg/dL [10]. Improvement in the BUN and serum creatinine in patients with initially abnormal values is a potential marker of treatment success and may be useful in determining disposition [24].

Attention has been placed on the significance of worsening serum creatinine in the setting of treatment for decompensated heart failure [17, 26, 35]. An increase in creatinine level of >0.3 mg/dL from hospital admission correlates with in-hospital death, complications, and length of stay. The presence of worsening renal insufficiency, as defined by a creatinine change of >0.3 mg/dL from prior values, is concerning, and patients may warrant further treatment until the creatinine improves or stabilizes [17]. Extrapolation from these studies suggests that an increase in serum creatinine identifies a high-risk group of patients.

In addition, studies have shown an association between worsening renal function after discharge and poor prognosis [25]. Gotsman et al. studied the significance of serum urea and renal function in patients with heart failure. They found that serum urea may independently have prognostic importance for patients beyond renal function [26]. It may be a more comprehensive data point to measure the clinical status because it encompasses parameters such as renal function, fluid volume balance, hemodynamics, and neurohormonal axis. Since serum admission and discharge urea are predictors of 1-year survival, admission serum urea may be used as possible data point for admission given its probable prognosis for both short-term and long-term survival.