Observation Unit Utilization in Decompensated Heart Failure and Clinical Predictors for Appropriate Triage

Document Type

Conference Proceeding

Publication Date


Publication Title

J Gen Intern Med


BACKGROUND: Acutely decompensated heart failure (ADHF) represents a significant financial and health burden in the United States with more than 1 million hospitalizations annually and 3% of annual health care expenditure. Currently, 75% of patients presenting to Emergency Departments (ED) are admitted to the hospital, however, estimates suggest as many as 50% may be safely treated in Observation Units (OU). The use of OUs is a cost-effective alternative to inpatient admission for ADHF, however, data to support appropriate triage of patients is sparse. We sought to isolate clinical data indicative of Length of Stay (LOS) greater than 48 hours that is readily available to the ED physician.

METHODS: We performed a retrospective cohort study with patients admitted from the ED with ADHF. Patients with end-stage renal disease, heart transplant, or ventricular-assist device were excluded. Demographic and clinical data pertinent to the evaluation of heart failure in the ED were collected. We compared patients with LOS < 48 hours to those with longer stays using Chi-Square analysis, Two-Sample T-Test, Cochran Armitage Trend Test, as well as Wilcoxon Rank Sum Test.

RESULTS: Of the 553 patients included in the study, 234 (42%) had LOS > 48 hours. The mean age was 70.3 years (standard deviation, 14.8 years) and 53% were female. Patients were more likely to require >48 hours of care if they had a higher Charlson Comorbidity Index (4.4 vs 3.9, p = 0.014), had systolic pressure <90 mmHg (3.8% vs 0.9%, p = 0.021) or had BNP >200 pg/mL (72.8% vs 60.9%, p = 0.006). Chest x-ray findings of pulmonary vascular congestion (21.6% vs 15.0%), pulmonary edema (6.3% vs 4.1%), or pleural effusion (21.2% vs 14.3%) were likewise associated with increased LOS (p = 0.003). Patients with LOS > 48 hours were also more likely to be anemic (hemoglobin 11.8 vs 12.2, p = 0.005), have a lower lymphocyte percentage count (19.7 vs 22.1, p = 0.009), and have higher troponin (0.07 vs 0.05, p = 0.042).

CONCLUSIONS: The Observation Unit is a growing care modality for ADHF. Our study identifies clinical characteristics readily available to the ED physician suggestive of the need for inpatient care. Multivariate analysis and prospective validation will be necessary to further develop our findings into a clinically useful triage tool.





First Page


This document is currently not available here.