How University of Colorado Health used AI-driven operations to grow inpatient capacity

0
105

When directing inpatient flow, even outside times of crisis, hospital personnel are constantly challenged to place the right patient in the right bed and the right unit for the right level of care.

The number of patients who need beds is inherently unpredictable. Front line staff often lack the real-time information they need about which units across the hospital are open, which can shift rapidly by the time a patient actually needs a specific bed. The necessary steps involved in patient admissions or discharges, such as prepping rooms or clearing lab results, add further complicating factors.

Navigating this process is an ongoing burden for clinicians and staff alike, contributing to already high levels of stress and frustration. The process also impacts the experience of patients and their families, as they may spend excessive time waiting for admission, discharge, or transfer to the right unit. Such waits are fatiguing, frustrating, and do not support patient health and healing.

Like other organizations, the University of Colorado Health (UCHealth) had identified this challenge and tried to support providers and staff to direct patient flow successfully and ensure timely patient placement. But the manual processes and tools they used did not mitigate the ongoing chaos, confusion, and frustration.

UCHealth adopts a single source of truth for managing inpatient flow

The 12-hospital UCHealth system serves Colorado, southern Wyoming, and areas of Nebraska. The entire network consists of almost 2,000 beds, while the flagship University of Colorado Hospital in Denver currently has 700 beds and performs over 50,000 admissions and observation visits annually. The University of Colorado Hospital also sees about 140,000 annual emergency visits and serves a clinically diverse patient population.

With such an extensive network and high level of patient flow, UCHealth captured a large amount of historical and current data on key metrics like unit usage, lengths of stay, and surges in the intensive care or post-anaesthesia care units. Hospitals in the system also offered staff a range of tools, including those built into the EHR, to access and report the data they needed. But these reports, dashboards, worklists, and spreadsheets were either labor-intensive, requiring manual updates throughout the day, or simply revealed bottlenecks and problems without offering staff direction on how to solve them or predictions on what was likely to happen in the future. As a result, patient care teams had to rely on instinct to make decisions and were unable to plan proactively ahead of potential surges.

In order to equip staff to successfully and efficiently direct patient flow and manage inpatient capacity, UCHealth needed a single source of truth to share in real time across departments, clinical disciplines, and the health system as a whole. To begin providing this, University of Colorado Hospital implemented LeanTaaS’ AI-based iQueue for Inpatient Beds solution in early 2020 – just before the COVID pandemic. The whole system adopted the solution as a virtual distributed command center later that year.

iQueue for Inpatient Beds provided the much needed single, automated source of real-time data, as well as predictive and prescriptive analytics that helped operational teams proactively manage upcoming patient surges and capacity needs. This in turn reduced wait times at key steps along the patient journey and mitigated the chaos that unfolded throughout 2020, as ICU units filled with COVID patients during surges and other critically vulnerable patients had to be housed at a safe distance.

UCHealth continues to grow capacity within its existing bed resources

UCHealth uses iQueue to help run daily bed meetings, perform hourly administrative management, and drive capacity protocol standardization, including surge planning. Staff are now able to predict future admissions and discharges, conduct network level protocol assessments, and confidently make strategic decisions to get the right patient in the right bed at the right time.

UCHealth has adopted a few key tactics to continue optimally managing beds. One of these is utilizing the Transfer Tool in iQueue for Inpatient Beds. Prior to the use of iQueue, transfers would occur when specified beds (i.e. those in the ICU) were needed immediately, resulting in reactive movements that could likely have occurred at more optimal times. Now with iQueue, capacity leadership and bed assignment teams can utilize recommendations that prioritize transfer requests. This in turn has enabled them to be able to meet forecasted demand in critical areas throughout the hospital exactly when it is needed.

In addition, the iQueue discharge toolkit has also been instrumental in empowering inpatient services to focus only on the units and patients that need more support to help achieve safe and efficient discharges. Staff no longer need to attend huddles floor-by-floor to determine the state of discharges, nor are they waiting to be paged or called to certain floors when significant backup occurs. Teams can use the worklist within the tool to view and efficiently tackle outflow challenges, which has yielded increased bed availability and improved patient throughput.

These changes, along with embedding the use of iQueue for Inpatient Beds into the work culture of the organization, have driven powerful results. Throughout 2020, despite a COVID-driven 18+% increase in census, UCHealth noted a 10% decrease in time to admit patients from the Emergency Department and an overall 16% decrease in time to admit. As of February 2022, University of Colorado Hospital has also now seen a 65% reduction in time to complete transfers from the ICU. Leadership reports that the optimization of inpatient capacity, driven by AI-based operations and a newly empowered front line staff, has also dramatically improved the patient experience.

 

Source