ED Observation Units and Their Efficacy
- Sep 4th, 2014
- William Fox
I had concerns for admitting a patient to the observation unit for suspected CHF exacerbation. The specific protocols for CHF exacerbation are available on EPIC and list a very specific set of criteria that allows for admission to the 2SS observation unit. I was initially curious about the development of these criteria, and more broadly, the efficacy of observation units and their sustainability as the face of a changing environment in Emergency Medicine. My main question was: do observation units lead to equal or better outcomes for patients versus conventional admission? Are they cost-effective or do they merely delay the costs associated with admission and full work-up?
A brief literature search presented an article published last October (now in print in the August 2014 issue) in Annals of Emergency Medicine titled “Randomized clinical trial of an Emergency Department observation syncope protocol versus routine inpatient admission.” The article examined intermediate-risk syncope patients who were either admitted to an observation unit or admitted to the hospital. The paper examined inpatient admission rate and length of stay, 30 day and 6 month serious outcomes after discharge, index and hospital costs, quality of life scores, and patient satisfaction all at 30 days. Five Emergency Departments were chosen to participate from March 2010 to October 2011, with full follow-up completed in April 2012. They were able to randomize 124 patients from an original screening group of 2,724. Many patients were excluded due to risk stratification (the paper looked at medium risk patients only), presence of exclusion criteria (study only included patients ≥50 years old, without seizures/intoxication/head trauma/confusion/languages other than English and Spanish/DNR/DNI/chemotherapy), or refusal of consent. After exclusion criteria were met, the remaining 124 patients were randomized to either the routine admission group or observation group.
The study had 62 patients in the observation group and 62 in the admission group. A comparison of the groups reveals the only significant difference between the two groups is the increased presence of abnormal initial ECG results in the “admission” control group when compared to the observation test group. Upon completion of the study, researchers found that there was a notable difference in inpatient length of stay (29 hours in obs and 47 hours in routine admission groups) and frequency of inpatient admission (9 patients in obs vs. 57 patients in admission). In terms of secondary admissions, the researchers found differences in the mean and median hospital cost at the index visit and at 30 days and an absolute cost reduction of $629 in the ED observation syncope protocol. The paper goes on to state that there were “no differences in diagnostic testing rates” and thus the cost differences are related to length of stay. Finally, they found “no significant differences in general health utility, syncope-specific quality of life, or patient satisfaction.”
The study, despite its findings, does discuss its limitations in detail. It specifically comments on the use of “intermediate risk criteria” to include patients, and that a number of the excluded “high risk” patients were due to treating physician judgment. This can significantly interfere with the study’s ability to make broad assessments of the efficacy of observation units in multiple groups of patients. Additionally, the study did not take into account outpatient facility costs or costs directly passed to the patient (i.e. co-pays). This could potentially limit the assessment of cost-effectiveness in this study. Despite these and other limitations, I felt this paper served as a fine introduction to the implementation of observation units and their potential benefits, in addition to their limitations. The paper demonstrates the importance of selecting appropriate patients for observation care in addition to the benefit of having clear admission/observation criteria for certain disorders.