Accounting for capacity: A real‐time optimization approach to managing observation unit utilization

Abstract

Hospital observation units (OUs) are meant to host patients for relatively short periods of time, during which healthcare providers can observe a patient (usually following an emergency department encounter) and assess the need for an inpatient hospital admission. The initial placement of a patient in a unit (inpatient or observation) can have consequences for the care the patient receives and the patient's length of stay in the hospital, as well as financial implications for both the patient and the hospital. We propose a practical real-time optimization algorithm that balances the need to place a patient in the OU with considerations for available capacity. Through computational experiments with inputs derived from real data, we study the benefit of using the algorithm for placing the right type of patients in the OU. We discuss the role of data-driven decision support tools like the proposed algorithm for anchoring the decision-making process in a hospital setting and for incorporating the viewpoints of multiple stakeholders by defining appropriate "rewards" for available alternatives.

Case Article—Machine Learning, Ethics, and Change Management: A Data-Driven Approach to Improving Hospital Observation Unit Operations

“…Case misclassifications, however, can be a substantial problem for OU operations, leading to increased average length of stay and decreased hospital capacity (Crenshaw et al 2006, Conley et al 2017. OU patients who "flip" to inpatient status and are transferred from the OU to inpatient wards typically incur additional hospital stay because of the handoff to a new physician, social worker, and nursing team (Epstein et al 2010, Dwyer-Matzky et al 2020. There is also the cost of potential medical errors and internal hospital logistics related to the transfer.…”

“…In such a case, the OU issues a request for a bed in an inpatient ward, and the patient is transferred to the ward once an inpatient ward bed is available. (See Figure 2 in the case accompanying this article for a flowchart of this process, as well as Dwyer-Matzky et al (2020), Lovejoy and Desmond (2011), and Baugh and Bohan (2008) for more details on observation units. )…”