Background:   Observation care has evolved significantly since its original inception by the Centers for Medicare and Medicaid Services (CMS).  Originally intended to span less than 48 hours in the majority of cases, the Office of Inspector General (OIG) reported that 11% of all observation stays lasted three or more nights in 2012.[1]Having a higher length of stay (LOS) increases costs to the patient and the hospital, impacts patient satisfaction and safety, and can affect patient flow. In our institution, despite a short stay Emergency Department Observation Unit and a Cardiac Short Stay Unit, 22 % of all ED medical admissions (2644 patients) were placed on medical units under observation status.  Average LOS for these patients in calendar year 2014 was 51.3 hours and reflects an observation patient population with multiple comorbidities and complicated observation diagnoses that did not meet criteria for the other two units.

Methods:

An inter-disciplinary team of administrators, physicians, nurses, case managers, were convened to decrease observation length of stay. Using Lean Six Sigma methodology, several processes were identified that were contributing to higher observation LOS. These were 1) lack of awareness of observation status on inpatient units 2) delays in testing for observation patients 3) variability in care delivery.  We implemented several strategies to achieve the goal of reducing LOS. First, a dedicated 12-bed Medical Observation Unit (MOU) was identified to cohort medical observation patients. Next, the new unit was staffed with a dedicated on-site nurse practitioners available to expedite discharges and deliver standardized care. Next, clinical pathways were designed for the top 12 diagnoses in the previous calendar year.  Lastly, prioritized workflows were designed for laboratory, radiology, and non-invasive testing to accommodate patients from the MOU.

Results:

Between 7/15/2015 and 7/17/16, observation status patients that were admitted to the MOU had an average LOS of 33.3.  This represents a 35% decrease in LOS compared to the pre-intervention metric.  For the concurrent time period, the LOS for medical observation patients outside of the MOU was 54.4 hours.

Conclusions:

The results indicate that when observation status patients are placed in a cohorted unit they have a substantially lower LOS when compared to the same patient population placed on general medicine units. LOS reduction is likely a result of cohorting leading to awareness of observation status, dedicated staffing, and protocol based care, and priority workflows for ancillary testing. Further, large academic medical centers may benefit from more than one observation unit and a reduction in LOS is possible even for more medically complicated observation patients.  Future interventions should be directed at identifying appropriate patients for observation level care, and increasing awareness of observation care among the healthcare delivery team.