Scheduling Interns at Hospitals: Queueing Models and Fluid Approximations

Abstract:

We build a queueing model to compare two policies for scheduling a hospital department's medical interns: one where interns work long shifts on alternating days, and another where interns work short shifts daily. New regulations restricting intern work hours, prompted by concerns over patient safety due to intern fatigue, will force hospitals to find feasible alternatives to long shift schedules. We use tools from queueing theory to show that a daily admitting schedule has a greater capacity to admit patients. However, short shifts can increase patient reassignments to new doctors, reducing the quality of care. Under an asymptotic rescaling, we show that in the fluid limit there is a threshold on the patient arrival rate determining whether long shifts or daily admitting causes fewer reassignments. Interestingly, the fluid limit of our queueing model is a dynamical system with non-trivial long run behavior. This work is motivated by an empirical study done with Brigham and Women's Hospital.

Biography:

Ross Anderson is a PhD student in the Operations Research Center working with David Gamarnik. He graduated from Cornell University with a BS in Operations Research and Computer Science.