Physician Discretion and Patient Pick-up: How Familiarity Encourages Multitasking in the Emergency Department

Patient demand for emergency medical services has never been greater. In the United States, as fewer people access medical care through a primary care provider, more people access care through the hospital emergency department (ED). Unlike other types of queuing systems, however, the ED allows physicians discretion in whom they serve. That is, ED queues do not operate solely under a policy of “first-come, first-served, by severity.” Therefore, we wanted to know: “What leads physicians to select which patients, and how many patients, they will treat?” We explore how familiarity between peer physicians affects patient selection and the chosen multitasking level, a process more commonly known in the ED as “patient pick-up.” We find greater familiarity leads to an increase in patient pick-up rate, observed multitasking, and shorter patient wait time, with no identifiable negative impact to patient processing time or length of stay. Patient demand for emergency medical services continues to rise from all-time highs. Physicians generally respond to the rising demand by increasing the level of multitasking. What leads emergency department (ED) physicians to select which patients, and how many patients, to treat? Queuing models frequently assume individual servers operate independently of other servers. In contrast, we consider how familiarity between peer physicians affects patient selection and the chosen multitasking level, a process more commonly known in the ED as “patient pick-up.” Using observations from two EDs, we explore whether familiarity alters patient pick-up behavior, we determine the effect of familiarity on multitasking, and we measure the combined impact of familiarity and multitasking on other ED outcomes. Among ED physicians, greater average familiarity leads to an increase in patient pick-up rate, observed multitasking, and shorter patient wait time, with no identifiable negative impact to patient processing time or length of stay. Moreover, the effects intensify at the end of a physician’s shift and for patients in severe condition. Within more familiar groups, physicians appear willing to exert more effort. Our study clarifies how the benefits materialize and illustrates why researchers must consider server familiarity moving forward. History: This paper has been accepted for the Operations Research Special Issue on Behavioral Queueing Science.