By Stacey Kusterbeck
If an ED patient is placed in observation status instead of admitted, it frees up an inpatient bed for another ED patient. “That does two things. It means that ED providers are not stuck caring for sick patients who require inpatient hospitalization while at the same time trying to care for arriving ED patients. Also, it shortens length of stay, because the turnaround time for observation is quicker than for hospitalized patients,” says Shamai A. Grossman, MD, MS, an associate professor of medicine and emergency medicine at Harvard Medical School, vice chair for healthcare quality, and director of observation medicine in the Department of Emergency Medicine at Beth Israel Deaconess Medical Center.
In the ED observation unit, patients are observed for 24 hours or less. For the most part, patients are very satisfied with ED observation, in Grossman’s experience. Many people are happy that their care was expedited, allowing them to get out of the hospital faster.
However, 13% of ED patients admitted to the hospital had a length of stay of only one day. This suggests that at least some of those patients could have safely been observed instead of being admitted. Grossman and colleagues conducted a study to find out if tracking and reviewing one-day hospital admissions would increase ED observation use.1 Senior emergency medicine faculty members reviewed charts of all 24-hour inpatient discharges from 2020 to 2021 (about 200 per month, for the duration of the study). “We then emailed the doctors that decided to admit the patient to ask them to consider, in retrospect: Could that patient have gone into ED observation?” says Grossman.
Some EPs defended their decision to admit the ED patients instead of placing them in observation and explained their reasoning in detail. Despite this, EPs started to place more patients in observation after receiving the email. The researchers analyzed ED observation volumes before and after the emails were sent and found that the percentage of patients placed into ED observation increased from 11.77% to 14.21%.
The researchers also analyzed admission rates for 1,448 patients put in ED observation during the study period. Overall admission rates for ED observation patients increased from 20.12% to 23.80%. “We found that it actually doesn’t take a lot to push doctors to change their practice. Just knowing someone was watching over them was enough to get them to put more patients in ED observation. The paradox was, we ended up admitting a disproportionate number of those patients,” explains Grossman.
The 21% increase in patients placed into observation resulted in an 18% increase in admissions from ED observation. “In some cases, the EP might decide that it’s appropriate to give ED observation a try, in the hopes of avoiding a hospital admission,” says Grossman. It is a reasonable decision for some ED patients with transient ischemic attacks to be put into observation, for example. The EP may have wanted to see if the patient’s symptoms would resolve, allowing the patient to avoid a hospital admission.
“What we think happened was, the clinical judgment of our physicians was pretty good. If they thought some patients were a little bit sicker and ended up admitting them, there was a good reason for that,” says Grossman.
Just because an admitted patient went home in 24 hours does not necessarily mean it was a mistake to admit that patient. In some cases, the EP admitted the patient specifically because of the need for services (such as hemodialysis and wound care teaching) that were unavailable in the ED observation unit. Some patients needed an operation, such as an appendectomy, yet still only required a day in the hospital. The study’s findings show that the decision to admit or observe is not always clear cut. Whether a one-day admit should have gone to ED observation instead “is still hard to pinpoint,” acknowledges Grossman.
- Grossman ES, Fradinho J, Chiu D, et al. The effect of increasing emergency department observation volumes on downstream admission rates. Am J Emerg Med 2024;77:17-20.
