Task Switching and Productivity in Collaborative Work: A Field Study of Hospitalists

Resource synchronization is critical to processes in which some tasks require the simultaneous collaboration of multiple professionals. To understand the effect of collaboration on productivity, we conduct an empirical study of hospitalists—hospital medicine physicians—at Northwestern Memorial Hospital in Chicago. A hospitalist’s patient-care routine includes visiting patients, making diagnoses and, often, communicating with other care professionals—a collaborative task—to guide the diagnosis and the patient treatment plan. As the hospitalist and her collaborators have other activities to tend to, consultations often require people to preempt their current work and switch to the collaborative activities. Such task switching introduces setup times that reduce hospitalist productivity. We present a model and conduct a rigorous empirical analysis using a dataset assembled from direct observation of physician activity and pager-log data. We show that more switches between documentation and communication causally lead to longer documenting time, and that “task switching to reach out” has a smaller impact than “task switching to respond.” We empirically estimate the setup time and show that 5 minutes or about 50% of the 11 minutes of documenting time is due to task switching. Given that documentation occupies about 40% of a hospitalist total case processing time, task switching results in about 20% productivity loss: if hospitalists did not incur a setup time, they could serve 20% more patients per day.